"Business, Sauder School of"@en . "Marketing, Division of"@en . "DSpace"@en . "UBCV"@en . "Clark, Ronald Nicholson"@en . "2011-09-23T18:12:03Z"@en . "1964"@en . "Master of Science in Business - MScB"@en . "University of British Columbia"@en . "This thesis presents a general review of sales forecasting literature with particular attention to the preparation of the sales forecast, the pre-planning activities and the review. In addition, forecasts are developed which show the expected sales of domestic softwood plywood to be realized by the plywood industry for the years 1964 and 1968. A procedure is then presented that Crown Zellerbach Company can follow in using the industry forecast to ascertain their share of the expected softwood plywood sales.\r\nSales forecasting is an essential prerequisite to company planning. Therefore, forecasts must be as accurate as possible because many activities within the firm are based on the sales forecasts. With the assistance of sales forecasts, vital marketing, financial and production plans ultimately emerge, together with their supporting schedules.\r\nThe person responsible for the forecasting task must acquire not only a detailed understanding of company activities but also a thorough knowledge of the characteristics of a sound forecasting operation. \r\nThe forecaster must be familiar with the various judgment, survey and statistical techniques available for developing forecasts and he must understand the necessity of carrying out numerous pre-performance and post-performance activities. The pre-performance activities must be dutifully carried out if the most useful forecasting method is to be chosen. Post-performance activities are equally important. A time-table for review and revision when necessary must be drawn up ahead of time if proper control is to be exercised over the forecast. \r\nA simple regression equation and three multiple regression equations are developed with the intention of using one or more of them to forecast industry softwood plywood sales for the years 1964 and 1968. The three multiple linear regression equations are rejected because each of them possesses one or more unacceptable negative constants. The simple linear regression equation has an extremely high coefficient of correlation and a small standard error of estimate. Since this equation contains these desirable features and seems to incorporate no underlying fallacy, this simple regression equation is the one chosen to forecast industry plywood sales. \r\nThe share-of-market approach is used to determine the proportion of the industry sales to be captured by Crown Zellerbach Company. The total projected industry sales figures are multiplied by a percentage which represents the company's present share of the total market. The figures that result represent the anticipated plywood sales to be achieved by Crown Zellerbach Company for the years 1964 and 1968. \r\nCrown Zellerbach should not depend solely on the technique developed in this thesis for forecasting plywood sales. They should continue to use the subjective or judgment technique that they have used for a number of years, but they would follow a better course if they used one or more statistical or survey methods in addition to the present method. A final forecast could be selected after an analysis had been made of the forecasted figures developed by the various methods."@en . "https://circle.library.ubc.ca/rest/handle/2429/37604?expand=metadata"@en . "SALES FORECASTING IN THE PLYWOOD INDUSTRY by RONALD NICHOLSON CLARK B.B.A., University of Washington, 1962 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF BUSINESS ADMINISTRATION i n the Faculty of Commerce and Business Administration Department of Marketing We accept t h i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1964 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of Bri t i sh Columbia, I agree that the Library shall make i t freely available for reference and study * I further agree that per-mission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publi-cation of this thesis for financial gain shall not be allowed without my written permission* Department The University of Bri t i sh Columbia, Vancouver 8, Canada ABSTRACT This t h e s i s presents a general review of sales forecasting l i t e r a t u r e with p a r t i c u l a r attention to the preparation of the sales forecast, the pre-planning a c t i v i t i e s and the review. In addition, forecasts are developed which show the expected sales of domestic softwood plywood to be r e a l i z e d by the plywood industry for the years 1964 and 1968. A procedure i s then presented that Crown Zellerbach Company can follow i n using the industry forecast to ascertain t h e i r share of the expected softwood plywood sales. Sales forecasting i s an e s s e n t i a l prerequisite to company planning. Therefore, forecasts must be as accurate as possible because many a c t i v i t i e s within the firm are based on the sales forecasts. With the assistance of sales forecasts, v i t a l marketing, f i n a n c i a l and production plans ultimately emerge, together with t h e i r supporting schedules. The person responsible for the forecasting task must acquire not only a d e t a i l e d understanding of company a c t i v i t i e s but also a thorough knowledge of the c h a r a c t e r i s t i c s of a sound forecasting operation. The forecaster must be f a m i l i a r with the various judgment, survey and s t a t i s t i c a l techniques avai l a b l e f o r developing forecasts and he must understand the necessity of carrying out numerous pre-performance and post-performance a c t i v i t i e s . The pre-performance a c t i v i t i e s must be d u t i f u l l y c a r r i e d out i f the most useful forecasting method i s to be chosen. Post-performance a c t i v i t i e s are equally important. A time-table f o r review and r e v i s i o n when necessary must be drawn up ahead of time i f proper control i s to be exercised over the forecast. A simple regression equation and three multiple regression equations are developed with the intention of using one or more of them t o forecast industry softwood plywood sales for the years 1964 and 1968. The three multiple l i n e a r regression equations are rejected because each of them possesses one or more unacceptable negative constants. The simple l i n e a r regression equation has an extremely high c o e f f i c i e n t of c o r r e l a t i o n and a small standard error of estimate. Since t h i s equation contains these desirable features and seems to incorporate no underlying f a l l a c y , t h i s simple regression equation i s the one chosen to fore-cast industry plywood sales. The share-of-market approach i s used to determine the proportion of the industry sales to be captured by Crown Zellerbach Company. The t o t a l projected industry sales figures are m u l t i p l i e d by a percentage which represents the company's present share of the t o t a l market. The figures that r e s u l t represent the anticipated plywood sales to be achieved by Crown Zellerbach Company f o r the years 1964 and 1968. Crown Zellerbach should not depend s o l e l y on the technique developed i n t h i s thesis for forecasting plywood sales. They should continue to use the subjective or judgment technique that they have used for a number of years, but they would follow a better course i f they used one or more s t a t i s t i c a l or survey methods i n addition to the present method. A f i n a l forecast could be selected a f t e r an analysis had been made of the forecasted figures developed by the various methods. X ACKNOWLEDGEMENT S The author wishes to extend h i s appreciation to Dr. S.M. Oberg, Department of Marketing, under whose guidance and encouragement t h i s study was i n i t i a t e d and completed. The writer i s also indebted to Mr. E. R. Blaine and Mr. T. D. Heaver fo r t h e i r assistance i n t h i s work. Several other persons gave valuable advice and information and to them I extend my gratitude. These people include Mr. T. Lannamae and Mr. D. Owen of Crown Zellerbach (Canada) Limited, Dr. J . Parkany of Weyerhaeuser Company, Dr. A. Halter of Oregon State University, Mr. A. Jones of the Douglas F i r Plywood Association and Mr. K. Bromley of the Plywood Manufact-urers Association of B r i t i s h Columbia. Recognition i s due to the s t a f f of the U.B.C. Computer Center. Their co-operation enabled the author to complete the cal c u l a t i o n s for t h i s study. The writer i s g r a t e f u l also to Mrs. A.C. Fiene f o r her c a r e f u l typing of the f i n a l manuscript. TABLE OF CONTENTS Chapter Page I. INTRODUCTION TO SALES FORECASTING 1 I. Introduction 1 I I . Purpose of the Study and the Hypothesis.. 2 I I I . Methodology f o r the Essay 3 IV. Sources of Information 5 I I . BASIC FORECASTS 6 I. q Introduction 6 I I . Economy Forecasts 8 Opinion P o l l 10 Indicator Approach 12 H i s t o r i c a l Analogy 13 Econometric Method.. 15 Gross National Product Approach 18 I I I . Element of Judgment 18 IV. R e l i a b i l i t y of Economic Forecasts 21 V. Selection of an Economic Forecast 23 VI. Industry Forecasts 25 Industry Data 28 E f f e c t s of New Materials 29 VII. Company Sales Forecast 29 Comparison of General Business Forecast with Sales Forecast 31 v i Chapter Page Comparison of Industry Forecast with Sales Forecast . 32 VIII. Short-, Medium- and Long-Range Forecast-ing 36 I I I . DEVELOPMENT OF A SALES FORECAST 41 I. Planning the Sales Forecast 42 I I . Q u a l i t i e s of a Useful Sales Forecast.... 47 I I I . Uses of a Sales Forecast 52 IV. Organization of a Sales Forecast 62 V. Responsibility f o r Sales Forecasting.... 71 VI. Performance Stage 75 IV. METHODS OF SALES FORECASTING 77 I. Hazards of Forecasting 78 I I . Sales Forecaster 79 I I I . Techniques Based P r i n c i p a l l y on Personal Judgment 81 Jury of Executive Opinion 83 Sales Force Composite Method 89 Area Sales Manager Composite 97 Use of Persons Outside the Firm 99 IV. Techniques Based on Surveys 101 V. Techniques Based on S t a t i s t i c a l Methods. 110 Simple Correlation 113 Multiple Correlation 118 Advantages of Correlation 122 v i i Chapter Page Disadvantages of Correlation 125 S e r i a l Correlation 130 Time Series Analysis 132 \u00E2\u0080\u00A2**\u00E2\u0080\u00A2 VI. Multiple Method Approach 143 VII. Forecasting From Scratch 145 - VIII.Forecasting the Demand for New Products 147 IX. Breakdown of the Sales Potential to T e r r i t o r i a l Units 151 V. ADDITIONAL FORECASTING STEPS 156 I. Choice of Method 156 I I . The Concept of a F i r s t Approximation-.. Forecast 166 I I I . Post Performance A c t i v i t i e s 168 IV. P r i n c i p l e s of Control Applied to Sales Forecasting 170 V. Review and Revision 173 VI. Control A f t e r the Forecasted Period.... 181 VI. MAKING THE SALES FORECAST 185 I. Product History 185 I I . T y p i c a l Uses of Plywood 186 III Competitive Products 189 IV. History of Crown Zellerbach Building Materials, Limited 193 V. Preparing to Forecast 195 v i i i Chapter Page Gross National Product ... 196 Residential Construction 198 Non-Residential Construction 198 Personal Expenditure on Consumer... Goods and Services 199 Selection of the Estimating Equation 199 Least-Squares Method 205 End-Use Index Method 206 VI. Making The Industry Forecasts. 208 VII. Canadian Economy 216 VIII. Plywood Sales f o r 1964 227 IX. Plywood Sales for 1968 231 Canada's Population 231 Determination of Gross National Product for 1968 i n 1957 D o l l a r s . . . 233 X. Crown Zellerbach Sales 238 VII. SUMMARY AND CONCLUSION 240 i x LIST OF TABLES Table Page I. Domestic Shipments of Rigid Insulating Boards 191 I I . Canadian Semi-Hardboard Production and Shipments 192 I I I . Dependent and Independent Variables 202-203 IV. Variables i n Constant 1957viDollars 204 V. C o e f f i c i e n t s of Correlation for the Dependent and Independent Variables.... 215 VI. Components of GNP i n Constant 1957 Dollars 230 VII. Number of Families and Number of Households i n Canada 233 VIII. Structure and Performance of the Economy.. 237 CHAPTER I INTRODUCTION TO SALES FORECASTING I. INTRODUCTION If a business i s to be run e f f i c i e n t l y management must be cognizant of the present s i t u a t i o n of the enter-p r i s e and must be aware of the d i r e c t i o n i n which the business i s moving. In order to prepare for the future, firms customarily prepare an estimate of anticipated sales volume for the coming six months, year or several years. Actual sales i n any current time period are compared with the anticipated sales f o r the same period. A study of the differences between and the s i m i l a r i t i e s of the two sets of figures provides guidance i n the control of pro-duction schedules, cost budgets and other commitments. An estimate of sales volume f o r a s p e c i f i c future time period i s c a l l e d a sales forecast. , The sales forecast i s not the same as a general business forecast. The l a t t e r forecast i s usually an e f f o r t to estimate trends i n general business on a broad scale, and could be for the nation as a whole, or for a major segment, such as a forecast for a l l durable goods. Such a forecast i s usually made by an economist. A sales forecast i s narrower i n scope, estimating the expectations of only one 2. company or organization. The sales forecast usually i s developed to a s s i s t management i n planning future a c t i v i t i e s and i n evaluating actual sales when they are achieved. General business forecasts often play an important r o l e i n the developing of a sales forecast. The sales forecasting task usually involves several stages. One stage involves a forecast or estimate of the trend of industry volume. A second stage involves an estimate of company sales. A t h i r d stage may involve further subdivision, as when sales within regional t e r r i -t o r i e s are estimated. I I . PURPOSE OP THE STUDY AND THE HYPOTHESIS The f i r s t objective of t h i s thesis i s to assess the a p p l i c a b i l i t y , v a l i d i t y and r e l i a b i l i t y of the various forecasting methods. Some forecasting techniques are easier to use than others. The simpler, ones can be employed by forecasters with l i m i t e d t r a i n i n g and exper-ience. Other techniques require not only considerable experience but extensive t r a i n i n g i n such areas as s t a t i s t i c s or mathematics. The forecaster may have l i t t l e d i f f i c u l t y j u s t i f y i n g to management the use of some tech-niques, but he may have great d i f f i c u l t y j u s t i f y i n g the use of other methods. 3. The second objective i s to prepare sales forecasts for Crown Zellerbach (Canada) Limited. The intention i s to produce a forecast of the domestic sales of softwood, plywood by Crown Zellerbach Building Materials Limited, for the years 1964 to 1968. This second objective i s based on the hypothesis that one year and f i v e year projections of domestic softwood plywood sales can be developed for Crown Zellerbach Building Materials Limited by using a l i n e a r multiple regression equation. The l i n e a r multiple regression equation arises from a s t a t i s t i c a l method of determining and measuring the rel a t i o n s h i p between sales and other independent a c t i v i t i e s . By forecasting the trend i n these other a c t i v i t i e s , sales volume for the industry and the company can be forecast. I I I . METHODOLOGY FOR THE ESSAY In the majority of discussions on sales forecasting the subject has not been approached from the broad managerial point-of-view. Sales forecasting has been examined as a question of what technique or method w i l l best forecast sales i n any p a r t i c u l a r case. However, approaching the problem of sales forecasting front t h i s narrow viewpoint w i l l not lead to the best solution because there i s no 4. such thing as a management function which requires only performance. From the management viewpoint there are pre- and post-performance a c t i v i t i e s that are just as important as the technique to be employed, so i t i s import-ant that a proper se t t i n g be provided for the actual devel-opment of the forecast as well as for a discussion of the s p e c i f i c techniques which implement that performance. Since the pre- and post-performance a c t i v i t i e s i n sales forecasting have been l a r g e l y neglected, the theory under-lyi n g t h i s area of forecasting has not reached an advanced stage. This essay endeavors to place suitable emphasis on the a c t i v i t i e s that should take place before and afte r the actual development of the forecast as well as on the plan-ning of the actual forecast i t s e l f . The methodology used to achieve t h i s goal i s outlined i n the following para-graph. Chapter Two provides an introduction to economic, industry and sales forecasting. Planning and organizing f o r sales forecasting as well as the uses of sales fore-casts are discussed i n Chapter Three. The various techniques f o r forecasting are described i n Chapter Four. In Chapter Five sales forecasting problems i n the i n d i v i d -ual firm are presented. In Chapter Six the actual f o r e -casts for Crown Zellerbach are developed. Chapter Seven contains the summary and conclusion. IV. SOURCES OF INFORMATION The information for t h i s essay was obtained from many sources. Various aspects of t h i s subject are discussed i n a large number of books and i n a r t i c l e s i n journals. Crown Zellerbach (Canada) Limited and Weyerhaeuser Company provided information as did the B r i t i s h Columbia Plywood Manufacturers* Association and the Douglas F i r Plywood Association of the United States. Additional information was obtained from the Dominion Bureau of S t a t i s t i c s and the Agriculture Department of Oregon State University. CHAPTER II BASIC FORECASTS I. INTRODUCTION Every businessman forecasts constantly whether he r e a l i z e s he i s doing so or not. On h i s forecasts he bases decisions concerning the quantities of materials and parts he should order, the size of the s t a f f he needs, the p r i c e s he should charge f o r h i s goods, the advertising and s e l l -ing a c t i v i t i e s he must promote and the carrying out of innumerable other functions for which he i s responsible. In the past the businessman claimed that i n t u i t i v e knowledge produced a more s a t i s f a c t o r y forecast than d i d s t a t i s t i c a l and economic methods and there was a good deal of j u s t i f i c a t i o n for. h i s viewpoint. Now better s t a t i s t i c a l techniques have been developed and better s t a t i s t i c a l material i s avai l a b l e i n greater quantity, so an analysis based on s t a t i s t i c a l economic methods i s f a r 1 superior to a businessman's i n t u i t i o n . So much depends upon a good forecast that a businessman should a v a i l himself of the modern methods and economic data now at h i s disposal i n order to obtain the best estimates possible. 1. R.L. E d s e l l , \"How to Forecast Your Sales,\" I n d u s t r i a l Canada, January, 1962, p. 21. 7. A forecast that i s accurate cannot be guaranteed even when modern techniques and relevant material are used, but i t w i l l be f a r more r e l i a b l e than a forecast without these b e n e f i t s . I f a businessman uses a poor forecast he stands less chance of r e a l i z i n g when business i s about to go into a decline, or when recovery i s immin-ent, and he could lose a great deal of money. These ups and down of general business a f f e c t every industry and every company i n the industry. Although not a l l industries and not a l l firms are equally affected by changes i n aggregate business a c t i v i t y , sales and net p r o f i t s do move up and down with s i m i l a r fluctuations i n the t o t a l business output, while the costs of operation display movement s similar to those i n the genex-al p r i c e l e v e l or i n the p r i c e of raw mat-e r i a l s and i n the average hourly earnings of labour. This economic a c t i v i t y i s not the only external factor that influences the progress of a business enterprise. P o l i t i c a l regulatory factors also play a part i n deter-mining the success or f a i l u r e of a firm. These external f a c t o r s must be predicted i f short and long-range plans are to be developed. \"To the extent that forces over which management can exert l i t t l e or no influence a f f e c t 3 . sales, p r o f i t s and the a v a i l a b i l i t y of c r e d i t and c a p i t a l , management must have an appraisal of the future course of aggregate economic a c t i v i t y based on a national ordering of a l l available relevant quantit-2 ative and q u a l i t a t i v e evidence.\" The sales forecaster can predict the movements of these forces by using the economic forecast. The welfare of a company depends not only on external factors but on i n t e r n a l ones as w e l l . Internal factors are those over which management has r e l a t i v e l y close c o n t r o l . They include the quantities of material ordered and processed, the methods of marketing, the personnel arrangements, the product mix, the a l l o c a t i o n of costs to products, the organizational structure and 3 the effectiveness of administration. These factors, too, must be predicted i f the company i s to be i n a favourable p o s i t i o n to carry on the planning and devel-opment of i t s operations. I I . ECONOMIC FORECASTS Three types of forecasts - economic, industry and sales - are considered i n t h i s essay. An economic 2. E.J. Chambers, Economic Fluctuations and Fore-casting, Prentice-Hall, Inc., 1961, p. 329. 3. Chambers, Forecasting, p. 327. 9. forecast r e f e r s to a p r e d i c t i o n concerning the future of business i n general as well as the future of i t s major 4 components and most es s e n t i a l processes. Included i n economic forecasting are predictions of gross national product, consumer spending, business inventory, indus-t r i a l production, employment, wholesale pr i c e s and the l i k e . There are several ways i n which forecasts can d i f f e r . F i r s t , the end r e s u l t s d i f f e r i n t h e i r d e t a i l . Second, the t h e o r e t i c a l foundations of economic fore -casts d i f f e r because forecasters hold d i f f e r e n t opinions concerning the operation of the economy. These d i f f e r -ences i n opinion account i n part for the r e l a t i v e success or f a i l u r e of economic forecasts. The forecasters who can best assess the balance of various forces w i l l produce the best forecasts. The t h i r d difference i s that techniques used i n the construction of economic 5 forecasts can vary. Forecasting techniques are b a s i c a l l y mechanical i n the application of a s t a t i s t i c a l technique 4. A.R. Oxenfeldt, \"The Preparation and Use of the Economic Forecast,\" i n Materials and Methods of Sales Forecasting, American Management Association, no. 27, p. 9. 5 . Oxenfeldt, Sales Forecasting, p. 10. based on a f a i r l y r i g i d model, or they are e n t i r e l y subjective. Forecasters usually combine r i g i d tech-niques with personal judgment and a consideration of q u a l i t a t i v e f a c t o r s . Different methods are sometimes used to predict d i f f e r e n t components of the economy. Economic forecasts can be c l a s s i f i e d as short-term, intermediate and long-term forecasts. The short-term forecast usually covers a period ahead of no more than two years, the intermediate a period of two to f i v e years and the long-term a period of f i v e years and over. Forecasters i n various industries have s l i g h t l y varying opinions concerning the length of time that constitutes a short, intermediate or long-range forecast. Various methods can be used to develop a short-term economic forecast. Some of these methods are discussed i n the following pages. Opinion P o l l One of the techniques used i n the development of a short-term forecast i s the opinion p o l l . I t s use involves asking many businessmen t h e i r opinions con-cerning the course of future business development. From t h i s \"sample of many\" there i s developed a comp-11. o s i t e judgment which i s accepted because of a f e e l i n g 6 of \"safety i n numbers.\" Sometimes these p o l l s are c a r e f u l l y planned and co n t r o l l e d by experts i n govern-ment, business and private research organizations. Often, however, they are not properly planned and con-structed, for i n s u f f i c i e n t attention i s given to the questions asked and to the people interviewed. One of the c r i t i c i s m s l e v e l l e d at t h i s type of forecasting i s that the sample of in d i v i d u a l s interviewed does not constitute a s c i e n t i f i c a l l y drawn sample or at least i s not a proper cross-sectional sample of bus-7 inessmen. However, probably no sample, regardless of the method of selection, would produce consistently s a t i s f a c t o r y r e s u l t s i n the area of economic forecast-ing. This statement must be accepted as correct unless the forecaster can show that the general trend of opinion, f o r some demonstrable reason, a c t u a l l y f o r e -t e l l s future events, and such an occurrence i s u n l i k e l y . Nevertheless, changes i n businessmen's opinions and expectations can often be ascertained through opinion p o l l s . 6. W.E. Hoadley, J r . , \"The Importance and Problems of Business Forecasting,\" i n H. Prochnow, ed., Determining the Business Outlook, New York, Harper and Brothers,1954,p.19. 7. Hoadley, Determining the Business Outlook, p. 19. Indicator Approach Another technique used i n short-term economic fore-casting i s the leading indicator approach. This approach i s often adopted after attempts involving the opinion p o l l i n g method have been unsuccessful. Key factors which appear to lead or coincide with general business a c t i v i t y are often used i n forecasting future economic 3 conditions. That i s , changes i n c e r t a i n p a r t i c u l a r factors or indicators such as income, i n d u s t r i a l pro-duction, government surplus or d e f i c i t , plant and equipment expenditures,, and money supply precede changes i n t o t a l economic a c t i v i t y . Leading indicators are usually chosen aft e r comparisons have been made between the movements of p o t e n t i a l indicators and movements i n previous business a c t i v i t y . Thus, an accurate general business forecast could be made i f one or more indicators could be located and appraised. Considering the large number of key factors that are used obviously no indicator i s considered e n t i r e l y r e l i a b l e . The most successful r e s u l t s are obtained when there i s a d e f i n i t e lead-lag r e l a t i o n s h i p between the leading indicator and general business. 8. Chambers, Forecasting, p. 341. When the i n d i c a t o r does not \"lead\" general business, the future movement of the indicator must f i r s t be predicted. The forecast of the indicator i s accepted because i t i s usually considered easier to forecast the key-factor than to forecast business as a whole. H i s t o r i c a l Analogy A t h i r d method, h i s t o r i c a l analogy, involves a l o g i c a l refinement of the key-factor approach. For purposes of forecasting,the i d e a l s t a t i s t i c a l i n dicator would be one with an unchanging sequence that precedes turns i n business by a f i x e d number of months and with an amplitude of expansion and contraction d i r e c t l y r e l a t e d to the extent of the upswing or downswing about to be experienced i n t o t a l economic a c t i v i t y . An i n d i c a t o r as perfect as t h i s has never been found. However, before World War I I , W. M i t c h e l l and A.Burns picked a set of twenty-one indicators from among the several hundred time series that the National Bureau of Economic Research had analyzed i n i t s studies of business cycles. The twenty-one indicators chosen were r e a d i l y available indicators of economic change and seemed to be the most trustworthy. After World War I I , Geoffry Moore studied several hundred time series and i n 1950 he published a revised l i s t of twenty-one indicators, Moore c l a s s i f i e d h i s business cyc l e indicators into three groups - leading, roughly coincident and lagging. They were c l a s s i f i e d accord-ing to t h e i r tendency to reach c y c l i c a l turns ahead of, at about the same time as, or l a t e r than business cycle peaks and troughs. Moore then employed a d i f f u s i o n index to summar-ize the movements of h i s three groups of s e r i e s . This d i f f u s i o n index provides an easy method fo r e v a l -uating the d i r e c t i o n of change i n a group of indicators The idea i s merely to count the number of items i n any group that are r i s i n g at any given time and to take t h i s as a percentage of the t o t a l number i n the group. This i s the percentage of the t o t a l number of items i n the group that are expanding. The percentage w i l l be above f i f t y i f mqre series i n the group are r i s i n g than f a l l i n g , but the percentage w i l l be below f i f t y i f more are f a l l i n g than r i s i n g . The percentage i s c a l l e d a d i f f u s i o n index because i t shows how widely diff u s e d the expansion movements are i n the sector observed. 9. Geoffrey Moore, ed., Business Cycle Indicators, Vol. 1, National Bureau of Economic Research, Princeton University Press, Princeton, 1961, p. 72. This method has not been used long enough to j u s t i f y complete acceptance, however, any forecaster i s wise to take notice when the NBER indic a t o r s s i g n i f y a change i n the d i r e c t i o n of general economic a c t i v i t y . The basic weakness i n t h i s method, as the NBER s p e c i a l i s t s point out, i s that an indicator that has successfully predicted i n the past w i l l not necessarily be success-f u l i n predicting i n the future. Furthermore, no i n d i c -ator has yet been found that has always been s a t i s f a c t o r y i n the past. We are not j u s t i f i e d i n assuming, therefore, that any indicator w i l l be consistently r e l i a b l e i n the future. \"The h i s t o r i c a l analogy approach nevertheless does provide a convenient means for analyzing the current business s i t u a t i o n with reference to past trends, and thus o f f e r s a basis for appraising differences as well as s i m i l a r i t i e s between any immediate business s i t u a t i o n and h i s t o r i c a l periods which may be deemed i n some manner to be comparable.\" Econometric Method In order to describe and project the main economic factors active i n the economy which are thought to influence the general trends of business, forecasters sometimes use the econometric approach which emphasizes the use of mathematical formulae. The objective i s 11. Hoadley, Business Outlook, p. 21. 16. to develop a number of mathematical equations which e f f e c t i v e l y describe previous changes i n general business and then to use these formulae to predict future events. F i r s t , the p r i n c i p a l determining factors are selected by extensive mathematical analysis of past .relation-ships between seemingly important factors and general business. National income and national product series are often used i n econometric models. When these factors have been selected and past relationships determined, the econometric formula or \"model\" i s created. This i s followed by placing the s t a t i s t i c a l data into the equation and using mathematical cal c u l a t i o n s to develop 12 the forecast. The econometric model method may be considered a mathematical refinement of the h i s t o r i c a l analogy method because t h i s technique i s heavily depend-ent on h i s t o r i c a l r e l a t i o n s h i p s . While the econometric method of forecasting has been used f a i r l y frequently i n recent years, the r e s u l t s have f a i l e d to l i v e up to 13 previous expectations. The single equation model has been widely used i n econometric work but i n many cases i n s u f f i c i e n t care 12. Hoadley, Business Outlook, p. 21. 13. Chambers, Forecasting, p. 342. was taken to assure that t h i s technique was the best one to apply to the problem at hand. Since t h i s model resulted from developments i n mathematical s t a t i s t i c s i t s popularity i s not s u r p r i s i n g . A great deal has been written on econometrics and most of t h i s l i t e r -ature has grown around the single equation l i n e a r model, discussing the assumptions to the analysis of economic data and the steps that can be taken i f one or more of 14 the assumptions i s inappropriate. The o b j e c t i v i t y of the econometric method has been advanced as one of i t s favorable q u a l i t i e s . This opin-ion i s open to question, however, because human judg-ment plays a s i g n i f i c a n t r o l e i n the development of mathematical equations. The important aspect of t h i s method i s the insistence upon a c o n t r o l l e d and regul-ated technique which necessitates the selection of f a c -t o r s that are v i t a l and the achieving of r e s u l t s that are consistent. Despite the l i m i t a t i o n s previously mentioned the method has value because of the demand that the quantitative factors used i n the forecasting process,be managed methodically. 14. J . Johnston, Econometric Methods, McGraw H i l l , New York, 1960, p. 145. Gross National Product Approach The f i n a l technique of short-term economic f o r e -casting, the gross national product or cross-section analysis approach, i s included only a f t e r some s l i g h t h e s i t a t i o n , because gross national product i s the most common measure used f o r estimating tire market value of the nations' output of goods and services. Since gross national product i s the best estimate we have of the contribution of economic a c t i v i t y to our material welfare, the r e s u l t s of an analysis developed by any technique w i l l probably, f o r comparative purposes, be 15 stated i n the context of t h i s p a r t i c u l a r measure. The u t i l i z a t i o n of the gross national product approach requires consideration of the predominant economic forces at work i n the country as well as a detailed sector-by-sector analysis of future expectations. I I I . ELEMENT OP JUDGMENT S t a t i s t i c s for the past and s t a t i s t i c s for the present furnish the foundation upon which future short-term forecasts are made, but another element also enters i n t o the pi c t u r e . Future economic developments 15. Chambers, Forecasting, p. 343. 19. depend upon economic and non-economic forces whose d i r e c t i o n of development cannot necessarily be ascer-tained from past and present occurrences. A decision concerning to what extent these past and present s t a t i s -t i c s w i l l be r e f l e c t e d i n the future depends upon the judgment of the forecaster, so the value of the predic-ti o n depends to a large extent upon the q u a l i t y of the forecaster's j udgment. Innumerable sources provide information and a good deal of misinformation concerning the business s i t u a t i o n so that there i s wide scope f o r the forecaster to use h i s judgment. There i s too much information to enable i t a l l to be handled e f f e c t i v e l y , so the forecaster 16 must decide from which sources he w i l l gather information. Further, he must decide which information i s relevant to the s p e c i f i c problem facing him and which factors w i l l exert more influence than others. He must decide also which statements are f a c t u a l l y correct. Anyone who provides data about business i s influenced by h i s own personal opinions and prejudices and the forecaster must be able to take t h i s into account when he i s using 16. V . I i . Bassie, Economic Forecasting, McGraw-H i l l Co., 1958, p. 6. the information he has gathered, and evaluate the 17 s i g n i f i c a n c e of the various statements. The forecaster himself w i l l probably have h i s personal prejudices and he should be aware of them i n order to minimize t h e i r e f f e c t . He should not have the burden of additional prejudices placed upon him by h i s environment. He should not have to adopt a p o l i t i c a l bias i f he i s employed by the government and he should not have to assume that prices w i l l remain stable or that p r i c e s w i l l r i s e continuously because h i s employer has a conservative or i n f l a t i o n -18 ary viewpoint. He must also t r y to free himself from the prejudices of others i n h i s own profession. The forecaster i s most e f f e c t i v e when he i s aware of h i s prejudices and when he makes every e f f o r t to be as objective and detached as possible. Therefore, good judgment i s an i n t e g r a l part of the forecasting process. 17. Bassie, Economic Forecasting, p. 6. 18. L.A. Livingston \"How Wrong Can Economists Be?\" The Reporter, May 26, 1953, p. 18. 21. IV. RELIABILITY OF ECONOMIC FORECASTS Economic forecasting has changed greatly since the pre-war days. In pre-war days so l i t t l e inform-ation was available that the professional forecaster was handicapped to the point where h i s forecasts were often as inaccurate as the forecasts of amateur fore -casters. Today conditions are quite d i f f e r e n t . A vast array of information i s now obtainable, so the forecaster must be trained to make adequate use of the information. He also must be a s t a t i s t i c i a n and he must be capable of analyzing and interpreting the 19 data. As a r e s u l t , the nature of forecasts has changed f o r they are more det a i l e d than they used to be. Some economists also believe that forecasts are more accurate than they formerly were. J.A. Livingston, an American business writer, c r i t i c i z e d forecasts produced from 1946 to 1952 as being inaccurate and unreliable, but he stated i n 1954 that since 1952, 20 forecasts had become much more accurate. Unfortunately, business forecasts have not been analyzed to the point where a statement can be made concerning t h e i r accuracy 19. Oxenfeldt, Sales Forecasting, p. 10. 20. Qxeh\u00C2\u00A3eldt, 'Sales Forecasting, p. 10. i n recent years. We do not know what matters fore -casters can most accurately predict, nor what they most conspicuously f a i l to foresee. We do not know which minor business changes have been predicted accurately, nor which major changes have eluded pre-d i c t i o n altogether. Present forecasting s k i l l i s judged on the a b i l i t y to forecast accurately during a period of mild i n f l a t i o n because since World War I I , business cycles have been of a minor nature. I f major business cycles had occurred during that period, forecasters might have been nearly unanimous i n t h e i r 21 predictions. So long as our cycles continue to be of a minor nature, forecasters w i l l not l i k e l y agree, therefore we cannot expect to know with ce r t a i n t y what the future of business w i l l be. Consequently, businessmen must be prepared not only to take action based on the fore-cast but must be prepared also to take alternate action i f the forecast proves to be inaccurate. While preparing alternate courses involves additional e f f o r t , t h i s action i s a desirable form of insurance for the company. 21. ' Oxenfeldt, Sales Forecasting, p. 16. V. SELECTION OP AN ECONOMIC FORECAST The statement has been made (p. 8 ) that management must have an economic forecast on which to base the sales forecast. Economic forecasts^can be obtained from sources outside the business or they can be devel-oped by the firm's forecaster before he st a r t s the sales forecast. If the sales forecaster plans to use an economic forecast not h i s own he should look f o r c e r t a i n q u a l i t i e s i n the forecast. The chosen fore-cast should be e x p l i c i t about what i s expected, and the underlying assumptions should be c l e a r l y stated. Preference should'be given to a forecast with good records. There i s a greater safety i n choosing pre-d i c t i o n s of forecasters who have been previously successful, avoiding the predictions of those who are given to sensationalism. Those forecasts should be selected i n which the forecasters define c l e a r l y the degree of confidence that they place i n t h e i r predic-22 t i o n s . One of the most important reasons f o r obtaining or making a general economic forecast i s to have a 22. Oxenfeldt, Sales Forecasting, p. 16. firm foundation on which to predict the company's sales volume. The company's sales forecast must be based on the general economic forecast because nearly every company's sales are affected by the condition of general business. There have been instances i n the past where some firms have increased t h e i r sales during a general economic downswing, while on the other hand some firms' sales have f a l l e n during a general economic upswing. In the f i r s t case, the firms' sales would have been even better i f business i n general had been r i s i n g , while i n the second case, the firms' sales would have f a l l e n s t i l l more i f the general economy had been f a l l i n g . The great majority of industries show the same general fluctuations as the general economy but at the same time they display i n d i v i d u a l d ifferences. Changes i n business generally a f f e c t i n d i v i d u a l industries i n varying degrees and with varying speeds. Therefore management should compare the timing and i n t e n s i t y of changes i n i t s own industry's a c t i v i t y with changes i n general business. This r e l a t i o n s h i p w i l l be found to be r e l a t i v e l y stable over a period 23 of time. 23. Oxenfeldt, Sales Forecasting, p. 19. Because most industries fluctuate i n a manner sim i l a r to the general economy, the economic f o r e -cast i s a required f i r s t step i n preparing a firm's sales forecast. The next step i s usually a forecast of the industry's t o t a l sales. With t h i s forecast, the firm's future sales can then be projected by employing the share-of-market method. VI. INDUSTRY FORECASTS The preparation of the general economic forecast must precede the preparation of the industry or the company forecast. Such basic economic factors as gross national product, personal income, construction a c t i v i t y , and p r i c e and production indexes must be related to the i n d i v i d u a l industry forecast being studied. One of the basic steps i n preparing an industry forecast i s to compare the movements i n the industry with the movements i n the general economy in order to understand the past relationships between the two. The knowledge of t h i s r e l a t i o n s h i p enables the forecaster to make use of the economic forecast when developing the industry forecast. An easy way to determine t h i s past r e l a t i o n s h i p i s to chart on graph paper the data f o r the past f i v e or ten years, 26. so that the relationship, i f present between the industry and various economic factors,can be seen. Relationships can be determined by simple v i s u a l comparison or by 24 \ complex s t a t i s t i c a l techniques. Management has considerable confidence when deal-ing with data on industry sales trends, but finds that forecasts on the general economy are d i f f i c u l t to use and of questionable r e l i a b i l i t y . Therefore i f a general economic forecast can be made that i s f a i r l y r e l i a b l e , t r a n s l a t i n g such an estimate i n t o probable industry sales can be managed much more 25 e a s i l y . When the industry forecast i s being prepared the factors that determine industry volume must be studied. There i s no single correct way to i d e n t i f y and analyze the c r u c i a l f a c t o r s . I f the sale of a product i s influenced l a r g e l y by the amount of disposable consumer income available, then i t i s most important to review the economic forecast of personal disposable income and adjust the expected industry volume f o r the product accordingly. If the forecast f o r new construction 24. Oxenfeldt, Sales Forecasting, p. 24. 25. D.M. Phelps, Sales Management, R.D. Irwin, 1951, p. 212. 27. indicates a decline, then the p o t e n t i a l industry volume of a product employed i n the construction industry must be adjusted downward. The influence of compet-i t i v e products also must be considered when industry volume figures are being determined. Other factors must be recognized when industry sales forecasts are being developed. National and even in t e r n a t i o n a l p o l i t i c a l s i tuations often play a major r o l e i n influencing the outlook of c e r t a i n i n d u s t r i e s . For instance, anticipated changes i n t a r i f f s can influence forecasts i n the f o r e s t r y industry. Technological changes must be considered, for new materials and new machines are constantly being devel-oped, and they can have a profound influence on the 26 market. Prices must, be taken into account i n the preparation of an industry forecast. If the product i s sold to farmers and farm produce prices are expected to drop, the industry w i l l probably be unable 27 to maintain past volume, much less increase the output. 26. G. Beise, \"The Preparation and Use of the Industry Forecast,\" In Materials and Methods of Sales Forecasting, American Management Association, no. 27, p. 24. 27. Beise, Sales Forecasting, p. 24. Even weather forecasts can influence the expectations i n some in d u s t r i e s . Industry Data Industry data for use i n forecasting can be obtained from a v a r i e t y of sources. Government depart-ments such as the Dominion Bureau of S t a t i s t i c s and the Department of Trade and Commerce publish and make available f o r public use a great deal of industry data. Trade Associations c o l l e c t data on t h e i r own industries, and publish some of t h i s information at regular i n t e r -v a l s . Unpublished information can often be obtained from these associations. In some instances trade associations are the only available source of data. This information, when broken down int o d e t a i l e d segments by type or size of product as well as by geographical area, can be of p a r t i c u l a r value. A company must be c r i t i c a l when using trade assoc-28 i a t i o n data. Great care must be taken by the i n d i v -idual firm i f comparisons are to be made between industry and company data i n order to evaluate the r e l a t i v e e f f i c i e n c y of the company. Some members of an industry may not belong to the association so the association 28. Beise, Sales Forecasting, p. 26. should make an e f f o r t to keep i t s e l f informed of the t o t a l sales of the non-members. Only i n t h i s way can the member-firms of the association determine t h e i r share of t o t a l industry sales. Effects of New Materials Long-range industry forecasting must take into account new methods, new materials and changes i n material acceptance and preferences. An industry may .represent a p a r t i c u l a r material end-use or competitive product but the forecaster should not assume that such a s i t u a t i o n w i l l continue unchanged. If an industry sees that competitors are introducing new materials, or that another industry i s using new materials that the f i r s t industry's competitors may use, that industry should make a complete market study, i f necessary, i n order to ascertain the advantages of introducing 29 the new material i t s e l f . VII. COMPANY SALES FORECAST The f i r s t step i n preparing a forecast for the i n d i v i d u a l firm i s to prepare an economic forecast, the second step i s to prepare an industry forecast 29. Beise, Sales Forecasting, p. 27. 30. and the t h i r d step i s to prepare the i n d i v i d u a l company sales forecast. The sales forecast i s a statement of the segment of the p o t e n t i a l t o t a l industry sales that a company can reasonably-. expect to secure. Many companies have c o l l e c t e d industry data on past sales over a period of years. With t h i s information the company r e a l i z e s what proportion of t o t a l industry sales were previously obtained as well as the trend 30 i n t h i s proportion. Estimates can then be developed for a future period. Many factors must be considered i n addition to industry data i f a s a t i s f a c t o r y company sales f o r e -cast i s to be developed. Consideration must be given to the r e l a t i v e q u a l i t y of the product, the promotional budget, the e f f i c i e n c y of the sales force, e x i s t i n g plant capacity and measures necessary to increase : 31 production. Therefore, the businessman must face the question of what can be done to secure a larger proportion of the t o t a l expected sales of the industry, and whether production p o s s i b i l i t i e s are such as to make available the volume of goods needed. 30. Phelps, Sales Managemert , p. 213. 31. Pheljas, Sales Management, p. 213. 31. Sometimes no data on past industry sales are a v a i l -able. Companies i n t h i s p o s i t i o n proceed d i r e c t l y from a general business forecast to a sales forecast. Such companies may make shrewd estimates as to t h e i r r e l a t i v e positions i n the industry and they, may be aware of the influence that a change i n general business could have on the industry as a whole and on t h e i r companies i n p a r t i c u l a r i n which case t h e i r sales 32 forecasts may prove quite s a t i s f a c t o r y f o r t h e i r needs. Comparison of General Business Forecast with Sales Forecast The difference between a general business forecast and a sales forecast should now be c l a r i f i e d . A general business forecast predicts trends i n general business a c t i v i t y . These can be trends f o r the nation as a whole or for a major segment of the economy. A general business forecast i s one of the elements u t i l i z e d i n developing a sales forecast. A sales forecast i s more lim i t e d i n scope than a general economic forecast. A sales forecast deals with the sales expectations of an i n d i v i d u a l company and i s used as a guide when management plans future a c t i v i t i e s as well as when i t evaluates the sales that are a c t u a l l y r e a l i z e d . 32. Phelps, Sales Management, p. 214. 32. The sales forecast i s considered easier to prepare than a general business forecast. Since the influences a f f e c t i n g future sales are more limited i n number than those a f f e c t i n g general business conditions, sales forecasts are l i k e l y to be more accurate than economic forecasts. In some situations, expecially i n the early stages of a sales forecasting program, sales forecasts are f a i r l y inaccurate because of the lack of necessary data or the lack of experience on the part of the fore-caster. However, some kind of sales forecast must and w i l l be made, and experience has shown that \"fore-casts based on an extensive search for, and a study of relevant facts are f a r more accurate than those 33 which lack such a f a c t u a l base.\" Comparison of Industry Forecast with Sales Forecast A noticeable difference i n accuracy often occurs between sales forecasts and industry forecasts, the 34 l a t t e r usually being more accurate. S h i f t s are constantly occurring of the i n d i v i d u a l company's competitive shares within the industry volume f i g u r e . 33. R. Crisp, \"Objectives and Techniques of Sales Forecasting,\" i n Sales Forecasting, Uses, Techniques and Trends, American Management Association, no. 16, p. 21. 34. Crisp,\u00E2\u0080\u00A2\u00E2\u0080\u00A2}Sales Forecasting, Uses, Techniques and Trends, p. 22. A large increase i n sales by one company i n an industry may reduce the volume and competitive shares of a l l other firms i n that industry. Most firms are unable to ;determine i n advance what t h e i r competitors' future strategy w i l l be, although t h i s may not apply where there has been a change i n competitive strength i n the preceding period. A s i g n i f i c a n t price-cut by a competitor or an important new product development by a competitor can lead to a great error i n any f o r e -cast. In f a c t , one of the most frequent reasons f o r errors i n sales forecasts can be underestimation of the competition. Thus we see that a sales forecaster must always work with one large, i n f l u e n t i a l , unknown and uncontrollable variable \u00E2\u0080\u0094 competition; and so long as competition cannot be predicted, sales forecasts w i l l not be r e l i a b l y accurate. Market Share f o r Forecasting We know that the forecaster f i r s t predicts t o t a l industry sales and then the company's share of the industry market. Industry sales are usually beyond the control of a single company, and are la r g e l y the r e s u l t of vast impersonal economic forces over which the comp-any has l i t t l e control, unless, of course, that com-pany enjoys a predominant proportion of the t o t a l 35 industry sales. The company does have some control over the percentage of industry volume that i t achieves. Past marketing a c t i v i t i e s , such as p r i c i n g , advertis-ing, number of salesmen, and number of outlets a f f e c t the present and future percentage, or share-of-market, and variat i o n s i n these a c t i v i t i e s w i l l influence future company share. The company's future share-of-market can be changed to an even greater extent by d e l i b e r a t e l y a l t e r i n g present marketing a c t i v i t i e s . Share-of-market concepts of demand are generally useful f o r forecasting purposes i n mature, w e l l -defined industries whose products are r e l a t i v e l y homogeneous. In such industries, market shares are r e l a t i v e l y stable and i n these cases the projection of market shares may be a very useful forecasting 36 device. Conversely, where market shares i n the industry fluctuate widely and unevenly, the projection of market shares cannot be used with any confidence. If market shares change very slowly, i t means that, strong forces perpetuate the established d i v i s i o n 35. Joel Dean, Managerial Economics, Prentice-H a l l , 1951, p. 155. 36. Dean, Managerial Economics, p. 156. of t o t a l industry business. These strong forces may consist of such factors as close personal t i e s between buyer and s e l l e r , or the possession of supplies and repair parts for one brand only. Other factors may be a b e l i e f that one p a r t i c u l a r brand gives a higher trade-in value than any other, or l o y a l t y to a c e r t a i n brand based on nothing more than f a m i l i a r i t y and a desire f o r security i n using what i s known and t r i e d . When a purchaser changes h i s arrangements to adjust to the c h a r a c t e r i s t i c s of a c e r t a i n brand, he helps 37 to maintain the established d i v i s i o n . When the company's share-of-market i s stable, management should not expect to change t h i s share much or quickly. E f f o r t s to do so are almost c e r t a i n to involve costs that could not be economically j u s t i f i e d . Care must be exercised i n using a share-of-market approach to company sales forecasting. I f a company i s maintaining, or even increasing, i t s share-of-market, i t does not necessarily mean that the company i s i n as enviable a p o s i t i o n as would at f i r s t appear, because the t o t a l industry sales may be f a l l i n g 37. A.R. Oxenfeldt, \"How to Use Market-Share Measurement,\" Harvard Business Review, January-February, 1959, p. 50. 38 r a p i d l y . Therefore, the forecaster must know not only the company's share-of-market but the trend of the industry as w e l l . VIII. SHORT-, MEDIUM- AND LONG-RANGE FORECASTING We know that forecasts can be made on three l e v e l s : the general economic, the industry, and the firm l e v e l . Forecasts can also be c l a s s i f i e d according to time span: the short-term, the intermediate and the long-term. Short, medium and long-range forecasts can be used when developing a general economic, an industry or a company forecast. The short-term forecast, which predicts no more than two years i n t o the future, enables the company to develop plans reasonably well ahead and thus to adjust more e a s i l y to an indicated higher or lower volume of sales. The short-run forecast must take into account such factors as the l e v e l of outstanding consumer c r e d i t or rapid changes i n inventories. This forecast i s not p a r t i c u l a r l y concerned with the stock of physical c a p i t a l or the' rate of population growth, because, although they are important, they are not l i k e l y t o a l t e r s u f f i c i e n t l y over a short 38. Beise, Sales Forecasting, p. 28. 37. 39 period to have much e f f e c t on short-run f l u c t u a t i o n s . One of the d i f f i c u l t i e s i n short-range forecast-ing i s the problem of determining when a change i s about to take place. Indeed, the forecaster sometimes has d i f f i c u l t y knowing when a change has already occurred. The simplest forecast to make i s the one i n which the next time period acts i n much the same way as the 40 previous period. If business improved during the preceding period, the forecaster appears safe i n predicting that business i n the next period w i l l improve, f o r he w i l l be r i g h t more often than wrong. He w i l l be wrong only at a turning point. However, one of the most important reasons for making a short-run forecast i s to determine when a change i s imminent, so a method that does not predict the turning points i s of l i t t l e value. Because t h i s simple type of fore-cast cannot predict the turning points, i t i s not used by professional forecasters. They employ more complex techniques which attempt to predict the turning points. I f the turning points are predicted, then short-run 39. J . Meredith, \"Short, Medium, and Long-Range Forecasting,\" i n Sales Forecasting, Uses, Techniques, and Trends, no. 16, p. 94. -40. ;Mejredith\u00C2\u00BB Sales Forecasting, Uses, Techniques and Trends, p. 95. 38. forecasting provides management with r a t i o n a l l y ordered information and a sounder basis f o r decision male ing. The intermediate-range forecast covers approx-imately two to f i v e years and i s a poorly developed area of p r e d i c t i o n . The forecaster cannot obtain surveys of consumer and business intentions for t h i s period of time. He cannot project r e a l i s t i c a l l y the long-term trends nor i s he i n a favourable p o s i t i o n 41 to evaluate the s i g n i f i c a n c e of q u a l i t a t i v e f a c t o r s . If a r e l a t i v e l y accurate intermediate-range forecast could be made, an appraisal of t h i s type would be e s p e c i a l l y valuable i n formulating a c a p i t a l expenditure program and the related f i n a n c i a l plan for research and product development. Intermediate forecasts are the proper place to consider the problems of c y c l i c a l f l u c -tuations. I f t h i s i s not done, the forecasts are meaningless. Judgment plays a dominant r o l e i n the intermediate-range forecasting. The forecaster must select from a l l available material those elements i n the s i t u a t i o n that w i l l be most s i g n i f i c a n t i n shaping future business during the next f i v e years, and t h i s i s a formidable task. 41. Chambers, Forecasting, p. 334. Long-range forecasts cover a period of f i v e or more years i n t o the future, and are used for the pur-pose of presenting a rough picture of prospects i n the future, a picture that has some empirical foundation. The long-range forecast should point the d i r e c t i o n of the most probable outcome of future business a c t i v i t i e s . Long-range aggregate projections are usually set i n a gross national product or gross national expenditure framework. When the d i r e c t i o n of future economic a c t i v i t y has been determined, consideration must be given to the siz e of future industry sales and then to the size of company sales. The knowledge gained from these forecasts a s s i s t s i n determining the need for product development and d i v e r s i f i c a t i o n , f o r additional channels of d i s t r i b u t i o n and f o r additions to the s t a f f . A long-range forecast may demonstrate the d e s i r a b i l i t y of preparing for heavy investment i n plant and equipment. Long-range projections are made \"1. To provide a basis f o r making a choice be-tween alternate courses of action when t h i s choice i t s e l f w i l l not s i g n i f i c a n t l y a f f e c t the economic projection and the objective i s to adapt i n t e r n a l programs to external economic forces beyond the control of the- decision making unit, as i n decisions of an i n d i v i d u a l firm i n a competitive market, and 2., to provide a diagnosis of possible future economic development as a basis for decisions which w i l l a l t e r , or can be expected to a l t e r , the course of economic events so as to i n v a l i d a t e the o r i g i n a l projection.\" There are a number of d i f f i c u l t i e s encountered i n long-range forecasting because these projections extend past present economic relationships into the future, and over a period of years the changes that take place can assume major proportions. Many factors must be considered i n respect to t h i s time element. As Kuznets pointed out,1 future economic conditions are determined not only by present economic conditions, but by growth i n technology, population and the number of workers, as well as by government changes, i n s t i t -u t i o n a l habits, i n t e r n a t i o n a l c o n f l i c t s and exhaustion 43 of natural resources. Thus long-range forecasting becomes an extremely complex operation. 42. James W. Knowles, \"Relation of Structure and Assumptions to Purpose i n Making Economic Projections,\" Proceedings of the 116th Annual Meeting of the American S t a t i s t i c a l Association, 1956, p. 7. 43. S. Kuznets, \"Concepts and Assumptions i n Long-Term Projections of National Product,\" i n Studies i n Income and Wealth, Vol, XVI, Long-Range.Economic Projections, New York, National Bureau of Economic Research; Princeton University Press, 1954, p. 15. CHAPTER III DEVELOPMENT OP A SALES FORECAST The sales forecast plays an important role when management uses the estimates to plan the future course of the organization. When forecasts are made and are reviewed by executives, there i s a looking into the future, a thinking into the future and a providing for the future. Forecasting, by concentrat-ing attention on the future a s s i s t s i n bringing to organizational planning a singleness of purpose 1 that cannot be attained e a s i l y i n any other way. Forecasting also may reveal areas where there i s a lack of adequate control and where there i s a need for control i n order to ensure the e f f i c i e n t operation of the business. Sales forecasting may help to bring unity and co-ordination into plans and so a s s i s t i n developing t h i s c o n t r o l . A discussion of sales forecasting should not commence with a consideration of the techniques or methods that w i l l be of greatest assistance i n a given s i t u a t i o n . A setting i s needed f o r the actual 1. H. Koontz and C. O'Donnell, P r i n c i p l e s of Management, McGraw-Hill, 1959, p. 488. performance of the sales forecast because there are pre- and post-performance a c t i v i t i e s and decisions that are just as important as the actual developing of the sales forecast. I . PLANNING THE SALES FORECAST The sales forecasting program can be broken down into several stages, the f i r s t of which i s planning. In t h i s planning or pre-performance stage the firm must decide i f a forecast i s needed and i f a for e -cast can be attained. The company also must decide i f the p o t e n t i a l p r o f i t a b i l i t y of the predictions warrants developing a forecast. I f a forecast i s to be developed the company must decide whether the resultant figures w i l l be accepted without a l t e r a t i o n , or whether the figures w i l l be raised or lowered. The sales figures to be used i n the forecast then must be selected and company p o l i c i e s that are to be adhered to must be decided upon. When these steps have been taken the actual developing of the for e -cast can begin. The f i r s t step i n planning i s to decide i f a forecast i s necessary. This decision rests upon the i n t e r n a l needs of the company. For example, c a r e f u l scheduling of operations i s necessary for control of production and estimating of future income and expenditures i s required for budgeting. If these needs can be met by developing a forecast, the decision i s made to go ahead with the other steps of the 2 planning stage. The step that follows the decision to forecast i s that of finding i f the forecast i s possible to at t a i n . Some businesses are unable to estimate future sales because of some important variable such as the development of new products or government regulations that cannot be determined i n advance. Data on past a c t i v i t i e s i n the f i e l d may be lacking, without which sales forecasting cannot be undertaken. Management should make sure that a sales forecasting program has a reasonable chance of success before commencing such an undertaking. Even i f the p o s s i b i l i t y of attainment i s reason-ably assured, management should attempt to determine i f the expense involved j u s t i f i e s continuing with the program. Only when anticipated benefits are expected to outweigh the costs of forecasting 2. CM. Crawford, Sales Forecasting; Methods of Selected Firms, University of I l l i n o i s , 1955, p. 15. should the program be established. In some firms costs would a c t u a l l y exceed benefits, but many firms that do not forecast only think costs would not be j u s t i f i e d ; they have not r e a l l y made a great 3 enough e f f o r t to f i n d out. When the f i r s t three steps have been taken and management has decided that a program of sales f o r e -casting i s advisable, possible and p r o f i t a b l e , the fourth step i s taken. This i s the selection of the degree of expectancy, because the forecast can be stated and used as a guide on three d i f f e r e n t l e v e l s regardless of the type of forecast developed. This term, degree of expectancy, i s not used i n a s t a t i s -t i c a l sense, where a forecast may be stated as $1,000,000 plus or minus $100,000, but refer s to a statement of the deliberate d i s t o r t i o n of the predicted f i g u r e . Management may use the figur e that i s given, or, on occasion, r a i s e or lower the fi g u r e . A raised figure provides a goal which the salesmen are urged to s t r i v e f o r . However, even with greater e f f o r t t h i s goal may not be reached. A lowered figure provides the f i n a n c i a l o f f i c e r s of 3 . Crawford, Sales Forecasting, p. 16. the company with protection i n the form of an estimate of minimum sales. This p r a c t i c e of d i s t o r t i n g the figure i s not common, but some managements f i n d 4 the operation valuable. The f i f t h step i n planning the sales forecast i s to decide,which sales figures to s e l e c t . This i s not a simple matter. There i s an endless va r i e t y of figures that may be used, but s u f f i c i e n t v a r i a t i o n can usually be achieved by using s i x breakdown bases, e s p e c i a l l y where several bases are used concurrently. These six bases are time period, product unit versus d o l l a r s , geographical area, product c h a r a c t e r i s t i c s , 5 type of customer and channels of d i s t r i b u t i o n . A common combination i s a forecast of sales, i n d o l l a r s , by time periods and by geographical areas. This i s done i n order to s a t i s f y the needs of sales and finance. Another combination i s a forecast by time periods and product units. This i s done to s a t i s f y production and a l l i e d needs. Usually both of these separate sets of estimates are broken down on the basis of d i f f e r e n t products. With so many possible combinations, determining which combination w i l l be most b e n e f i c i a l 4. Crawford, Sales Forecasting, p. 16\u00E2\u0080\u00A2 5 . Crawford,\"Sales Forecasting, p. 17. to the company i s a complex problem. The decision i s often made by comparing the needs of the various in d i v i d u a l s i n the company for the information with the cost of the forecasting. Some sales forecasters probably base t h e i r decisions on guesswork more often than on knowledge, f o r measuring the cost or p r o f i t -a b i l i t y of making sales estimates i s a d i f f i c u l t thing to do. We now come to the sixth step i n the planning stage, a step that i s too often ignored. An e f f o r t should be made to formulate broad company p o l i c i e s that are conducive to successful sales forecasting. The f i r s t thing needed i s the planning of a general p o l i c y of operations. Some companies do not plan such a pol i c y , they merely repeat the same procedure again and again because the t r a d i t i o n has been estab-l i s h e d . Next, management should provide the sales forecaster with an organizational environment that encourages the interdepartmental flow of information and co-operation. The forecaster should engage a c t i v e l y i n trade association programs that provide for an exchange of information. Furthermore the sales forecaster should be informed of the r e s t r i c t i o n s placed upon him; that i s , he should be t o l d whether he i s a c o l l e c t o r of data, an interpreter, or one who 47. i s completely i n charge of the preparation of the 6 forecast. Marketing plans should also be stated before the forecast preparation period commences. The p o l i c i e s just mentioned do not necessarily guarantee accurate forecasts but they w i l l reduce and perhaps eliminate unnecessary mistakes and expenses. I I . QUALITIES OP A USEFUL SALES FORECAST A sales forecast that i s to be useful to top management should possess c e r t a i n d e f i n i t e q u a l i t i e s . Some of these q u a l i t i e s are discussed i n the follow-ing paragraphs. 1. The industry or market forecast i s an e s s e n t i a l part of the t o t a l forecast. What i s commonly c a l l e d a sales forecast i s r e a l l y a market forecast plus a sales goal i n the expected market that depends upon carrying out successfully a sales plan f o r acquiring a desirable and f e a s i b l e segment of the market. 2. The underlying assumptions of the forecast should be stated b r i e f l y and p r e c i s e l y . The assumptions 6. Crawford, Sales Forecasting, p. 17. should be concerned with s p e c i f i c future events or developments which cannot or should not, as a matter 7 of policy,be forecast. Assumptions should also deal with developments that would a f f e c t the forecast greatly and i n a ce r t a i n s p e c i f i c way i f they occurred. These could be a s t r i k e , the outbreak of war or a major p o l i t i c a l development. Although we may not recognize the fact, every forecast contains many assumptions. They should be written down, f o r doing so often helps to c l a r i f y where assumptions end and forecasting begins. Committing the assumptions to paper helps to c l a r i f y thinking because the statements must be cl e a r and concise. Management can then check the basic concepts behind the forecast and change an assumption that i s unacceptable and the forecaster may a l t e r h i s forecast accordingly. Another important advantage r e s u l t s from writing down the assumptions. Recalling opinions held i n the past i s d i f f i c u l t . The only safe way i s to write them down. At a l a t e r time, differences between assumptions and r e a l i t i e s can be studied and an e f f o r t made to learn the extent to which any differences between assumptions and 7. B. Estes, \"What Management Expects of Fore-Casting\" i n Sales Forecasting \u00E2\u0080\u0094 Uses, Techniques and Trends, American Management Association, Special Report no. 16, 1956, p. 14. a c t u a l i t i e s were responsible for differences between forecast and actual r e s u l t s . 3. When the assumptions have been stated, the 8 forecast should be put i n authoritative form. A f o r e -cast couched i n terms of uncertainty i s of l i t t l e help to management. The forecaster may believe that he i s safeguarding himself i f he avoids making d e f i n i t e statements, but since management i s f u l l y aware that i n the f i n a l analysis the forecast i s based on judg-ment, the forecaster should be decisive, promise only what can r e a l l y be accomplished and accept responsib-i l i t y for whatever forecast he makes. After a l l , management i s interested i n only one basic f a c t . Is business going to change, and i f so, i n what d i r e c -t i o n and how much? Management does not want the answer clouded with innumerable q u a l i f i c a t i o n s . 4. A forecast should not be changed too frequently. At the same time, the forecaster should be ready to revise h i s forecast when there i s a change i n basic conditions. Sometimes feel i n g s of optimism or pessimism permeate a business or even a community. A forecast can often be used to counterbalance these emotional 8. Estes, Sales Forecasting - Uses, Techniques and Trends, p. 15. reactions. To be usable, a forecast must be based on fundamental factors which are seldom changed suddenly or frequently, and the forecast should be adhered to so long as those fundamental factors 9 e x i s t . A forecast that i s adjusted too frequently i s not worthy of the name of forecast, i t i s just a r e f l e c t i o n of the current mental outlook of the business community. I f , however, fundamental factors a c t u a l l y do change the forecast must be adjusted accordingly because i n fairness to the company the forecaster must use the best and most uprrto-date information a v a i l a b l e . Use of such information means that the forecast w i l l be u t i l i z e d to the best advan-tage. 5. A statement should be given of the precise period covered by the forecast. In a short-range for e -cast the period covered may be obvious, but i n a long-range forecast t h i s i s not always true. Not only should a statement be made concerning the year or years to which the forecast r e f e r s , but the expected p o s i t i o n i n the business cycle should be assumed and stated. The forecaster must determine whether he i s 9. Estes,, Sales Forecasting - Uses, Techniques and Trends, p. 15. going to predict on the basis of a normal l e v e l , a peak l e v e l or a base l e v e l . 6 . Most managements prefer that d e t a i l s and 10 techniques be omitted from the forecast. The f o r e -casting organization must use a l l available v a l i d techniques, gather as many fa c t s as possible, go into great d e t a i l and use every care, but should consider the techniques as t o o l s to be kept a v a i l -able i f management requires them, not as d e t a i l s that must be entered into either the o r a l or written presentation of the forecast. Should management require information on the d e t a i l s or techniques, a s p e c i f i c request can be made. 7. Forecasts should always be checked against what a c t u a l l y occurred. A good forecaster w i l l always i n s i s t on determining the main reasons for s i g n i f i c a n t differences between h i s forecast and the actual 11 r e s u l t s . This i s the only way by which he can make each forecast better than the l a s t . In addition, checking of t h i s kind enables the forecaster to gain some insight into the probable r e l i a b i l i t y of the 10. Estes, Sales Forecasting - Uses Techniques and Trends, p. 16. 11. Estes, Sales Forecasting - Uses,Techniques and Trends, p. 16. forecasts he i s providing. I I I . USES OP A SALES FORECAST The basic management function of any business organization is to co-ordinate sales and production i n order to achieve c e r t a i n p r o f i t objectives. These p r o f i t objectives are usually a l l i e d with sales volume. This makes the sales forecast the basic t o o l f o r ind i c a t i n g the future trend of business. If a l l functions of an enterprise are to co-operate i n an e f f o r t to at t a i n the decided goal, then a l l departments concerned should be completely f a m i l i a r with the future prospects. This can best be accom-plished by use of the sales forecast. The sales forecast can be useful i n many ways. The forecast estimates the goods that w i l l be sold during a s p e c i f i c time period. The period may be a short-term or a long-term one. This discussion covers both types unless s p e c i f i c mention i s made otherwise. Furthermore, the sales forecast i s assumed to be s u f f i c i e n t l y accurate to be useful to top management i n predicting future operations. Sales forecasts can be used i n the planning a c t i v i t i e s of the produc-t i o n department when plant operations are scheduled, expansion requirements considered and t r a f f i c manage-ment de t a i l e d . They are useful to the personnel department and to the purchasing department when the l a t t e r considers requirements of raw materials and other supplies. The finance department can use the sales forecasts when c a l c u l a t i n g the cash inflow, the p r o f i t p o s i t i o n and c a p i t a l requirements, and the sales department can u t i l i z e sales forecasts to adjust to changing trends, to emphasize p r o f i t items and to plan advertising and promotion campaigns. The success of a manufacturing business depends lar g e l y upon the co-ordination of two important functions \u00E2\u0080\u0094 production and sales. The production department, therefore, should f i n d a wealth of useful information i n the sales forecast, i n fact, t h i s department should be able to use the sales forecast as a production schedule. The extent to which the production department uses the sales forecast depends l a r g e l y upon the type of business involved. In a job-order shop, an order i s necessary f o r everything produced and the l e v e l of plant operations depends upon the backlog of orders, so a sales forecast i s not important. In those larger businesses where demand i s c y c l i c a l , the firm must know i f business w i l l increase or decrease as well as the time and duration of the expected change, so i n t h i s case a forecast i s e s s e n t i a l . Information regarding the depletion of warehouse stock can be obtained from the sales forecast and the necessary steps can be taken to bring t h i s stock up to a proper l e v e l . When a firm deals i n seasonal goods which have d i f f e r e n t inventory l e v e l s at d i f f e r e n t seasons and which have d i f f e r e n t production runs at d i f f e r e n t seasons, the employment of the sales forecast becomes e s s e n t i a l . I f the firm has several warehouses throughout the country and several producing plants the complexity of operations emphasizes s t i l l further the i n d i s p e n s a b i l i t y of the sales forecast. The co-ordinating of production with sales i s done on a short-term basis but long-term estimates are necessary for other operations. The scheduling of production requires detailed accuracy i n a sales forecast that the long-term forecast i s usually unable to provide. However, when production capacity i s measured against sales demands i n determining any required expansion of f a c i l i t i e s and i n projecting the competitive s i t u a t i o n , the long-term sales forecast i s v i t a l . This forecast may indicate that production economies or product improvements are mandatory i f the firm's p o s i t i o n i n the industry i s to be maintained. Another use the production department can make of a sales forecast i s i n the area of t r a f f i c manage-ment. Today, transportation costs form a large part of the cost of d i s t r i b u t i n g products. A company that moves a large volume of product fin d s t r a f f i c co-ordin-ation a big problem. If the sales forecast i s used by the t r a f f i c department as a means of keeping informed concerning sales expectations, the transport-ation function can be co-ordinated more e f f i c i e n t l y 12 with production and sales. These expectations include not only changes i n volume but changes i n the geographical areas of the market which include consideration of a d i f f e r e n t type of transportation, as well as changes i n destination and other related problems. The most useful sales forecasts are usually those that are broken down into geographical areas. As l e v e l s of operations increase or decrease, so 12. J . Dodge, \"The Uses of Sales Forecasts by Other Departments, \"\u00E2\u0080\u00A2 i n Sales Forecasting - Uses-, Techniques and Trends, American Management Association, Special Report No. 16, 1956, p. 83. do personnel requirements and personnel managers are better prepared to meet s h i f t s i n demand where they have the use of accurate sales forecasts. A long-term forecast can show the need for additional personnel and indicate the si z e of the h i r i n g and t r a i n i n g programs that may be required i n order to reach anticipated objectives. I f substantial expan-sion i n the size of a business i s planned, the long-term sales forecast i s obligatory. Purchasing departments can make extensive use of sales forecasts. Armed with figures concerning probable sales and production, the purchasing department i s better able to maintain stocks of raw materials and supplies adequate for insuring uninter-rupted production. Overstocking, with possible loss due to declining prices, deterioration and obsolescence can be minimized and warehousing and carrying costs can be kept under c o n t r o l . A long-range forecast of requirements enable s the purchasing department to plan f a r enough ahead to take advantage of favourable pr i c e s , at the same time lessening the danger of 13 over- or under-stocking. These advantages are 13. C.G. Thompson, Forecasting Sales, National I n d u s t r i a l Conference Board, Studies i n Business P o l i c y , No. 25, 1947, p. 38. increased when markets f o r raw materials and supplies are unstable. The purchasing department should a s s i s t the forecaster by keeping him informed of changes i n the p r i c e and supply s i t u a t i o n of important inputs because such changes would a f f e c t the p r i c e 14 and thus the demand for the product. The f i n a n c i a l department which supervises the disbursement of money can make use of the sales f o r e -cast to determine the expected cash inflow. Every business must know how much money i s being received or i s expected i n order to estimate i t s operations and t h i s cash flow i s fundamentally dependent upon 15 sales. This department also uses the sales forecast to estimate cash requirements and to plan short and long-term financing. The assumed l e v e l of production and sales serves as a basis for the development of standard costs and for the preparation of operating-budgets. Many companies report that without the assistance of reasonably accurate sales forecasts, t h e i r finance departments could not serve the company adequately. 14. Thompson, Forecasting Sales, p. 38. 15. Dodge, Sales Forecasting - Uses, Techniques and Trends, p. 48. Another r e s p o n s i b i l i t y of the finance department i s to determine c a p i t a l requirements and again the department turns to the sales forecast. This inform-ation i s e s p e c i a l l y necessary at a time of rapid expansion when long-term plans must be developed i n 16 some d e t a i l . Cash inflow and expenditures are compared to determine the r e l a t i v e area of p r o f i t a b i l i t y , and p o l i c y decisions to pay dividends, or to increase or decrease them depend on the p r o f i t s . A l l of t h i s information i s a l l i e d to the sales forecast. The sales forecasts influence the p o l i c y planning of a l l departments d i r e c t l y concerned with sales. These departments use forecasts i n various ways. P o l i c y must be planned i n regard to changes i n trends when 1hose changes are of major proportions. Sales depart-ments must use long-range forecasts that indicate major changes i n trends i n order to persuade the company to d i v e r s i f y i t s products. When the sales forecast indicates the probable siz e of the market i n the period ahead, the projection provides a goal for the sales force. Frequently, the forecast i s broken down into quotas for products, for regions and often for i n d i v i d u a l salesmen. Salesmen and sales managers w i l l exert themselves to meet 16. Thompson. Forecasting Sales, p. 38. these quotas i f they are v a l i d and r e a l i s t i c , which i s unlike t h e i r attitude towards the \" f i f t e e n percent more than l a s t year\" method which often discourages salesmen when they see no chance of achievement. Sales quotas based on the sales forecast are often the foundation of a sales compensation plan. When t h i s approach i s used, the past accomplishments of salesmen are ignored and the true p o t e n t i a l of each 17 area i s measured. Salesmen can then compete fo r recognition and large incomes upon an equitable b a s i s . Adjustment must be made i n regard to those factors which are beyond the salesman's control but which may a f f e c t greatly h i s performance. Sales forecasts have been useful i n i n d i c a t i n g i f sales t e r r i t o r i e s are properly established. A study of the sales forecasts reveals which t e r r i t o r i e s are too large to be handled adequately by the number of salesmen assigned to them and which cannot provide s u f f i c i e n t volume to guarantee a f a i r return. Sales forecasts are h e l p f u l when they can be used as a basis f o r d i r e c t i n g the e f f o r t s of salesmen. Areas where the company i s performing e f f i c i e n t l y i n 17. Thompson, Forecasting Sales, p. 38. r e l a t i o n to competition and areas where performance i s not s a t i s f a c t o r y are revealed by the sales f o r e -casts. With t h i s information, sales executives can d i r e c t increased e f f o r t to the areas where i t i s needed. There i s no need to increase promotional or sales costs i n areas where the company's p o s i t i o n i s so strong that improvement i s u n l i k e l y and sometimes promotional and sales costs can be lowered, a l l of vhich means increased p r o f i t s f o r the company. Prices are l a r g e l y dependent upon costs, costs are influenced by volume and volume i s affected by p r i c e . A company that wishes to establish s e l l i n g p rices and at the same time make a reasonable p r o f i t should have an idea of p o t e n t i a l sales volume. The p r i c e s can be based upon the finance department's standard costs and they i n turn can be calculated on an assumed rate of production and t h i s rate may be based on the sales forecast. By following t h i s procedure prices can be set at a p r o f i t a b l e l e v e l . There are some products where s l i g h t p r i c e f l u c -tuations w i l l not have a net iceable e f f e c t upon volume and other products where lim i t e d changes i n the rate of output w i l l have l i t t l e e f f e c t on costs. But there are other cases where volume i s highly sen s i t i v e to p r i c e changes and where costs are c l o s e l y a l l i e d to volume. Here a r e l i a b l e sales forecast can be u t i l i z e d i n the setting of p r i c e s that w i l l assux-e a high volume and a reasonable p r o f i t . Sales forecasts are indispensable i n the adver-18 t i s i n g and sales promotion f i e l d s . The sales forecast i s usually prepared on the basis of the t o t a l market and i s broken down by employing the share-of-market technique. In most companies, the amount of money spent on sales promotion has a s i g n i f i c a n t e f f e c t on t o t a l sales r e a l i z e d . For t h i s reason, the advertising and sales promotion departments must be thoroughly f a m i l i a r with a l l d e t a i l s of the sales forecast i f they are to co-ordin-ate t h e i r e f f o r t s with the a c t i v i t i e s of the sales department. F a i l u r e to do t h i s may r e s u l t i n f a i l u r e to achieve anticipated sales. Forecasts showing p o t e n t i a l markets f o r new products have been found to be e f f e c t i v e tools f o r aiding management i n d i r e c t i n g the a c t i v i t i e s of 18. Dodge, Sales Forecasting \u00E2\u0080\u0094 Uses, Techniques and Trends, p. 85. i n d u s t r i a l research laboratories. Before a company undertakes the expensive processes of creating, developing and d i s t r i b u t i n g a new product, studies are made of estimates of the p o t e n t i a l p r o f i t s to discover i f production i s advisable. In t h i s way, cos t l y errors can be kept to a minimum. The sales forecast as i n i t i a l l y produced by the sales forecaster can be used only by the sales and production departments. I f i t i s to be used by other departments, something must be added i n broad terms to make the forecast of intere s t to other departments, but each department must make detailed interpretations f o r i t s e l f . This r e s u l t s i n the most e f f i c i e n t co-ordination around an accepted projection of future operations. IV. ORGANIZATION OF A SALES FORECAST In the planning stage of the sales forecast, many and varied decisions were made. The need for a forecast was decided, the a t t a i n a b i l i t y of the forecast was considered, the p o t e n t i a l p r o f i t a b i l i t y was calculated, the degree of expectancy was se t t l e d upon, the figures to use were selected and the company p o l i c i e s were defined. When t h i s task i s completed, management has to consider the second stage of the program. Here, management must specify the type of organization that w i l l be of most use to the forecaster. In the past l i t t l e attention was given to organizing f o r sales forecasting, but some companies recognize that consideration must be given to t h i s phase of the task. There are three aspects to the organizational problem that are att r a c t i n g the attention of business-men today. Business executives are considering the a d v i s a b i l i t y of a separate department for sales fore-casting and a l l i e d a c t i v i t i e s . They wonder who should be made responsible for forecasting, and to what extent he should be held responsible, and they speculate as to where the forecasting a c t i v i t y should' be formally placed i n the organizational structure. In the past, sales forecasting was done most frequently i n the sales department, i n the accounting department or i n the finance department, but today some firms are placing the function i n a separate department. Formerly, a person primarily responsible for other duties was given sales forecasting as a secondary a c t i v i t y . Since the Second World War a l l managerial aspects of business have been undergoing intensive reconsideration and sales forecasting has received some thought. Most executives i n both large and small companies who have studied the si t u a t i o n c a r e f u l l y and weighed the advantages and disadvantages believe that forecasting should be placed i n a 19 separate department whenever possible. The factors favouring a separate department f o r sales forecasting can be expressed i n terms of general management p r i n c i p l e s . F i r s t , the organization of a company should provide f o f functional s p e c i a l i z a t i o n Sales, production and finance a c t i v i t i e s are becoming so complex that separate departments are now common for such services as personnel, t r a f f i c and sales promotion. This d i v i s i o n of labour makes best use of the a b i l i t i e s of each i n d i v i d u a l , r e s u l t i n g i n greater e f f i c i e n c y , and enables each person to develop s k i l l that increases h i s effectiveness by allowing 20 him to concentrate i n a lim i t e d f i e l d . In addition, persons doing the same kind of work, or a l l persons whose work requires s i m i l a r a b i l i t i e s may be grouped 19. Crawford, Sales Forecasting, p. 50. 20. W.E. Newman, Administrative Action, New York, Prentice-Hall, 1951, p. 132. together into a single administrative u n i t . In such cases, the work of the s t a f f members as well as that of the executive i s s p e c i a l i z e d . Any a c t i v i t y that i s designed to serve as a check on another a c t i v i t y should be under the control of a separate executive. Since sales forecasts provide a means of determining the e f f i c i e n c y of many a c t i v i t i e s within the firm, forecasting should 21 occupy an independent p o s i t i o n . Top management na t u r a l l y compares actual r e s u l t s with estimated r e s u l t s at the end of the forecasted period. Dis-crepancies c a l l f o r an explanation. The closer the actual r e s u l t s are to the forecast, the less l i k e l y w i l l a l i n e executive be c a l l e d on to explain the discrepancy. If the forecaster i s not independent from l i n e executives he may f e e l compelled to a l t e r h i s forecast so that the estimates w i l l be more i n l i n e with the actual r e s u l t s achieved by the l i n e executives. To avoid t h i s s i t u a t i o n , forecasting should be placed i n an independent department. If a c e r t a i n a c t i v i t y i s p a r t i c u l a r l y important to the success of a company, that a c t i v i t y deserves 21. Newman, Administrative Action, p. 134. s p e c i a l recognition. Businessmen are r e a l i z i n g to an ever increasing degree that c a r e f u l organizational planning i s absolutely necessary, and because sales forecasts are an i n t e g r a l part of t h i s planning, forecasting i s becoming an indispensable operation. For t h i s reason the status of forecasting should be considered. When a forecaster occupies a po s i t i o n of low status i n the organization he w i l l be unable to get the co-operation of executives i n higher p o s i t i o n s . The greater the prestige the company wishes forecasting to have, the higher i n the formal 22 structure should the a c t i v i t y be placed. Only i n t h i s way can the forecaster obtain the co-oper-ation the company f e e l s he should receive. Forecasting i s so important to the success of a company that adequate attention should be given to the function. I f a highly placed executive i s responsible f o r forecasting i n addition to h i s other duties he cannot give top p r i o r i t y to a l l h i s tasks and he may even neglect one or more of them. Fore-casting may be one of the duties neglected. If the a c t i v i t y i s the only or primary r e s p o n s i b i l i t y of 22. Newman, Administrative Action, p. 137. an executive, that person at least w i l l bestow upon forecasting the necessary attention. Because a forecaster i s concerned with the t o t a l operation of h i s company he should have a general knowledge of the a c t i v i t i e s of the company rather than of one s p e c i a l department, and he should be aware of opinions concerning the company and industry both inside and outside the firm. Different departments within the organization have d i f f e r e n t goals, d i f f e r -ent ideas and d i f f e r e n t tasks as well as varying degrees of optimism concerning the industry's outlook. Persons outside the firm also express diverse opinions regarding the future business outlook. Prom an appraisal of these diverse opinions, the forcaster must evolve a set of estimates that w i l l serve as a basis for a l l departments, and for t h i s task he i s best equipped i f he i s not involved with the a f f a i r s of one s p e c i a l department. One of the axioms of present day management i s that i n t h i s day of highly spe c i a l i z e d functions one i n d i v i d u a l should be responsible and have authority fo r one function. Where there i s r e s p o n s i b i l i t y there must be authority and for every task that i s 68 . undertaken authority must rest on someone. Authority gives an executive power to undertake assigned duties and r e s p o n s i b i l i t y demands that he use t h i s 23 authority to accomplish them. The forecaster, unfortunately, finds that t h i s concept does not always apply to him. Some firms define the r e s p o n s i b i l i t i e s and authority of the forecaster p r e c i s e l y , stating what should be done and by whom, but many firms include so many persons i n the forecasting a c t i v i t y that r e s p o n s i b i l i t y and authority cannot be centered i n any one i n d i v i d u a l . The firms that have stated c l e a r l y what sales forecasting should include and have a l l o t t e d the r e s p o n s i b i l i t i e s and authorities c a r e f u l l y are generally those firms which place the function i n a separate department. Despite the advantages of placing forecasting i n a separate department, there are circumstances where t h i s may not be desirable. In the f i r s t place, there must be assurance of f u l l time work for the fore-24 caster. In some companies forecasting does not require f u l l time application so the function i s added to a person such as marketing research d i r e c t o r or 23. Koontz and O'Donnell, P r i n c i p l e s of Management, p. 95. 24. Newman, Administrative Action, p. 141. sales manager. In a few firms the marketing research department has been raised to s t a f f capacity at the top management l e v e l and has been made responsible f o r several areas of business research including sales forecasting. Where there i s such an arrangement, forecasting i s a l o g i c a l function of such a department. Expense must be considered when there i s a 25 separate department for sales forecasting. The creation and maintenance of a separate department costs money. More o f f i c e space i s usually required, more executives and s e c r e t a r i a l help are needed and addit i o n a l services usually must be provided. When a choice i s to be made between departmentalizing and not departmentalizing one of the factors to consider i s the number of executives and s t a f f personnel required and t h e i r respective s a l a r i e s . I f the less expensive arrangement i s also the less e f f e c t i v e one i t may not be the better choice, but i f the more c o s t l y arrangement i s chosen the additional benefits should very c l e a r l y exceed the additional expense. The expense involved may provide the major reason why a separate department i s not established for 25. Newman, Administrative Action, p. 142. forecasting. In setting up a separate sales f o r e -casting department management may be able to determine, with some degree of accuracy, the f i n a n c i a l benefits to be r e a l i z e d . Unfortunately, the burden of addition-a l red tape and i n f l e x i b i l i t y which should be included as part of the expense i s almost impossible to measure i n f i n a n c i a l terms. One of the most d i f f i c u l t decisions of management i s to determine to what extent the enterprise i s j u s t i f i e d i n setting up a more elaborate and s p e c i a l i z e d form of organization. Departmentalization i s known to increase the complexity of a firm's organization, which i n turn adds to the d i f f i c u l t y of maintaining clear-cut 26 l i n e s of authority and r e s p o n s i b i l i t y . The whole problem of organizational complexity i s a d i f f i c u l t one to deal with, but the problem i s there with sales forecasting as with other functions. F i n a l l y , we must note that basic organizational p r i n c i p l e s cannot be applied indiscriminately to a l l firms. Some companies have proved that a separate sales forecasting department i s successful. Others 26. Koontz and o'Donnell, P r i n c i p l e s of Manage-ment, p. 81. may not have the necessary talent at t h e i r disposal to handle such a department and buying talent for a high executive p o s i t i o n may cause widespread d i s s a t -i s f a c t i o n throughout the firm. On the other hand, some men already i n the firm may have the necessary talent to supervise sales forecasting i n addition to some or a l l of t h e i r other duties. Generalizations assume away the p e c u l i a r i t i e s of company personnel, but the operating executive must take them into account. V. RESPONSIBILITY FOR SALES FORECASTING. Whether forecasting i s done i n a separate department or not the r e s p o n s i b i l i t i e s entailed should be c a r e f u l l y stated and assigned. In forecast-ing, as i n most jobs, only one person should have o v e r - a l l r e s p o n s i b i l i t y for the work, and the extent to which he i s responsible should be c l e a r l y stated. There are several reasons why t h i s should be done but the main reason i s that someone must be held account-able, not only because the job of forecasting must be done, but because i t must be done on schedule and also because the task involves co-ordinating the a c t i v i t i e s of various members of the firm. When res-p o n s i b i l i t y i s not fixed, the burden of co-ordination f a l l s on a busy top executive. 72. When r e s p o n s i b i l i t y for forecasting i s to be placed i n an i n d i v i d u a l , there i s general agreement among forecasters concerning which executive o f f i c e r s would be suitable, but there i s no such agreement concerning the placing of r e s p o n s i b i l i t y for the f i n a l estimates. The o f f i c e r s usually considered for the position are the sales manager, the c o n t r o l l e r , the marketing research d i r e c t o r and the head of a separate sales forecasting department or top-level research u n i t . Most forecasters agree that any of these i n d i v i d u a l s could do the job i f h i s r e s p o n s i b i l -i t i e s were c l e a r l y defined and i f top management 27 valued the function. There i s l i t t l e agreement, however, on exactly what r e s p o n s i b i l i t i e s should be included i n sales forecasting. The following tasks are generally con-sidered necessary: the forecasts should be prepared, they should be d i s t r i b u t e d to the people concerned, they should be studied c a r e f u l l y to insure the highest degree of accuracy possible and they should be 28 altered i f and when necessary. There i s agreement on the fact that someone should be responsible for 27. Crawford, Sales Forecasting, p. 5 3 . 28. Crawford,:Sales Forecasting, p . 54. continuously studying a l l sales forecasting arrangements with a view to improving them i f possible, as well as agreement on the fact that some conclusion must be reached concerning the extent to which the f o r e -caster w i l l be blamed fo r errors i n the f i n a l estimates. However, i d e n t i f y i n g the causes of errors i s not e a s i l y done. Errors may be due to a lack of data, to uncontrol-labl e variables or to poor judgment, and as a r e s u l t management may be unable to pin-point r e s p o n s i b i l i t y . To make the s i t u a t i o n more d i f f i c u l t the forecaster usually has to engage the assistance of other executives to help prepare the forecasts or to approve them, and some of the errors may be p a r t l y due to them. Forecasters generally agree that they should be held accountable for errors due to t h e i r own poor judgment, but agreement probably never w i l l be reached on the question of how much r e s p o n s i b i l i t y should be attached to the senior executive who places h i s stamp of approval on the forecast prepared by a subordinate. Once the r e s p o n s i b i l i t i e s for sales forecasting have been determined, management can proceed to study the problem of where to place the a c t i v i t y on the organization chart. Several p o s s i b i l i t i e s are a v a i l -able. The f i r s t of these i s found i n many medium sized companies where an i n d i v i d u a l or a department i s responsible for marketing research a c t i v i t i e s 29 including the development of periodic sales forecasts. In many small firms the sales manager or a senior executive i s held accountable for sales forecasting. The second method i s to set up a business research department d i r e c t l y accountable to the executive vice-president and the sales forecasting would then be one of the r e s p o n s i b i l i t i e s of t h i s department. A t h i r d a l t e r n a t i v e i s a complex development of the second. Production planning and sales planning departments are set up and t h e i r heads report to the executive i n charge of operations planning and r e -search. This executive has an advisory s t a f f r e l a t i o n ship to the executive vice-president. This arrange-ment i s desirable when the planning a c t i v i t i e s of the executive vice-president are so important to the success of the firm that adjustments are made i n the organizational structure to enable him to receive every assistance with h i s co-ordinating and 29. H. Bund and J . C a r r o l l , \"The Changing Role of the Marketing Function,\" i n The Journal of Market-ing, January, 1957, p. 307. 30 organizational a c t i v i t i e s . The production and sales planning departments provide the information necessary for i n t e l l i g e n t planning and e f f e c t i v e co-ordinating by the executive vice-president as wel l as by other executives. The sales forecast i s an example of the type of information that can be provided by the sales planning department. VI. PERFORMANCE STAGE With the completion of the f i r s t two stages i n the forecasting program, planning and organizing, the t h i r d stage i s embarked upon. This stage, which we s h a l l c a l l the performance stage, can be broken down int o a series of steps that d i f f e r but l i t t l e from steps followed i n performing other management a c t i v i t i e s . The f i r s t step i n the performance stage i s choo ing the i n d i v i d u a l who w i l l d i r e c t the forecasting procedure. He should possess c e r t a i n general q u a l i f i c a t i o n s . One of these i s sound judgment, for he must be capable of evaluating the s i g n i f i c -ance of changes; another i s a knowledge of general business conditions and another i s the capacity to 30. Crawford, Sales Forecasting, p. 57. command the respect of the senior executives. He should also be f a m i l i a r with basic modern business s t a t i s t i c s . Management does not know p r e c i s e l y which q u a l i t i e s and how much knowledge the forecaster needs to be successful, so i t i s not surprising that various management have widely d i f f e r e n t opinions regarding the q u a l i f i c a t i o n s that they consider desirable i n a forecaster. Making the forecast i s the second performance step. This includes such a c t i v i t i e s as delegating duties, scheduling a c t i v i t i e s , c o l l e c t i n g and analyzing the data and then stating the f i n a l forecast. The completed forecast i s usually forwarded to senior executives f o r appraisal and acceptance or r e j e c t i o n . The techniques that can be used to develop the fore-cast are described i n the following chapter. CHAPTER IV METHODS OF SALES FORECASTING The most d i f f i c u l t and most important aspect of sales forecasting i s deciding which technique to use. There are innumerable techniques to choose from and forecasters have d i f f e r e n t opinions concerning the merits of the various methods. Even today new techniqu and adaptations of old techniques are constantly being developed. The methods employed by forecasters are d i f f i c u l t to describe i n general terms because forecasters adjust the techniques to s u i t the p a r t i c u l a r features of the situations confronting them. Techniques can, however, be c l a s s i f i e d a r b i t r a r i l y as follows: 1. those based mainly on personal judgment, 2. those based on surveys and 3. those based on s t a t i s t i c a l methods. No technique has yet been devised that does not require the use of some degree of personal judgment and there i s l i t t l e l i k e l i h o o d that a technique i n which judgment plays no part w i l l ever be found. Although a l l forecasting commences on a judgment basis the forecaster uses various techniques i n an e f f o r t to reduce h i s dependence on a r b i t r a r y judgment, and i f t h i s i s not possible, at least to enhance the qua l i t y of h i s judgment i n order to increase the accuracy of h i s forecast. I. HAZARDS OF FORECASTING The degree of accuracy attained and the types of problems encountered i n forecasting vary from product to product, from company to company and from industry to industry, but there probably i s no area where problems are so insurmountable or accuracy i s so unattainable that forecasting of sales i s absolutely impossible. Problems that contribute to the d i f f i c -u l t i e s of accurate forecasting a r i s e i n various ways. The extent to which accuracy can vary i s usually 1 related to the per unit cost of the product. I f a single unit costs several hundred thousand d o l l a r s the accuracy of the forecast can depend on obtaining or losing one order. On the other hand the sale of a mass-produced product costing only a few d o l l a r s can vary by a few thousand units and the forecast would not be unduly affected. 1. C.G. Thompson, Forecasting Sales, National I n d u s t r i a l Conference Board, Studies i n Business Policy, No. 25, 1947, p. 2. The forecasting of company sales i s easiest and most accurate when consumption i s rapid and purchasing i s regular. I f the purchase of a product can be deferred the forecasting task becomes d i f f i c u l t , and the longer the purchase can be deferred the more d i f f i c u l t the estimating becomes. When there i s a p o s s i b i l i t y that consumers w i l l turn to substitute products the same problem a r i s e s . The very factors that increase the hazards and l i m i t the accuracy of forecasting are the same factors that make forecasting e s s e n t i a l . The presence of a noticeable time i n t e r v a l between the purchase of the raw material and the sale of the f i n i s h e d goods to the consumer increases the d i f f i c u l t i e s of fore -casting but also increases the value of forecasting to manufacturers, wholesalers and r e t a i l e r s who produce, stock and s e l l products i n an t i c i p a t i o n of 2 consumer demands. I I . SALES FORECASTER When one i n d i v i d u a l i s placed i n charge of sales forecasting and i s given some independence i n h i s posi t i o n , he finds that the p o s i t i o n c a r r i e s with i t 2. Thompson, Forecasting Sales, p. 2. c e r t a i n advantages and disadvantages. He w i l l have a comprehensive view of the company as a whole so he w i l l not make the mistake of emphasizing the problems 3 i n one area and ignoring the problems i n another. He may f i n d i t easier to get information from the various departments than do other executives, and through experience he may do a better job of analyzing and interpreting data than would an executive who had other duties i n addition to forecasting. Having one person i n charge of forecasting i s inexpensive, usually f a s t , and avoids interrupting the work, of other Weaknesses i n the forecaster's position are appar-ent i f the forecaster a l t e r s h i s judgment to moire nearly conform with the sales objectives that he knows othershold. His p o s i t i o n i s a weak one i f he lacks status i n the company or i f he lacks the exper-ience necessary to evaluate the effectiveness of h i s company's planned operations. The forecaster should be capable not only of making good forecasts but of e n l i s t i n g the confidence and co-operation of the company's executives 3, E.W. Grunow, \"The Role of Consulting Services i n Forecasting\" i n Materials and Methods of Sales Fore-casting, American Management Association, no. 27, p. 58 because the information, advice and judgment they contribute to the forecast may have much to do with 4 the successful conclusion of the program. When a sales forecaster i s a s t a f f s p e c i a l i s t he usually reports to the senior sales executive or to the executive vice-president, and i s thus given an excellent opportunity to add the judgment of t h i s executive to h i s own. Other senior executives can be c a l l e d upon to evaluate the estimates of the fore-caster, or a committee may be formed of as many as ten executives who meet to evaluate the forecaster's estimates. When estimates are being made general economic expectations or anticipated behaviour i n major segments of the economy may also be taken into account. The firm's own sales executives are often consulted concerning the effectiveness of planned marketing a c t i v i t i e s . I I I . TECHNIQUES BASED PRINCIPALLY ON PERSONAL JUDGMENT Sales forecasting techniques cannot be expressed i n general terms because they are so numerous and because they are adapted to varying s i t u a t i o n s . An 4 . Thompson, Forecasting Sales, p. 3. 8 2 . a r b i t r a r y c l a s s i f i c a t i o n can be made because they do have c e r t a i n features i n common. The f i r s t such c l a s s i f i c a t i o n covers techniques based p r i n c i p a l l y on personal judgment. There are several ways to define \"judgment\", but here the word i s used to designate a mental procedure i n which values and p r o b a b i l i t i e s that are known are welded into a reasonable conclusion 5 about the unknown. Judgment, therefore, is, more than a guess and more than an opinion. Not a l l judgment, however, i s good judgment. Various i n d i v i d u a l s both inside and outside the company may be asked to give the sales forecasting program the benefit oftheir consideration. The individu a l s i n the firm that the forecaster i s most l i k e l y to c a l l upon for opinions are senior executives, usually the president, the executive vice-president, and the top sales, production and finance executives, as well as executives i n the advertising and promotion, c r e d i t , budgeting, product development, and marketing research departments, depending upon the type of information he i s seeking. Another i n t e r n a l source b. CM. Crawford, Sales Forecasting; Methods of Selected girms, University of I l l i n o i s , 1955, p. 21. of judgment i s the sales organization which includes both the salesmen i n the f i e l d and t h e i r immediate supervisors. Opinions can also be obtained from sources outside the firms such as consultants, suppliers, personal friends and even competitors. The i n d i v i d u a l s whose opinions are sought may-be asked to estimate industry sales, company sales, or sales of a l i n e or sales of a single product. Usually opinions are sought f i r s t on matters at the industry l e v e l , often on a product-by-product basis, and a f t e r that at a company l e v e l , on a share-of-6 market ba s i s . Jury of Executive Opinion. A group of top executives may be formed into a committee either to pass judgment on the forecaster's predictions or to prepare forecasts themselves. Such a committee i s said to constitute a \"jury of executive opinion.\" This p o l l i n g of the opinions of executives i s one of the oldest and simplest means of forecasting industry and company sales and i s based on a b e l i e f i n \"safety-in-numbers\". A forecast based on the combined judgment of several executives 6. Crawford, Sales Forecasting, p. 21. i s assumed to be superior to a forecast based on the judgment of only one. Some companies ask the members of t h e i r executive committees to prepare i n d i v i d u a l estimates of sales and hand them to the president, who makes a f i n a l forecast based on the opinions expressed or on a 7 s t a t i s t i c a l average of the estimates. The forecast i s sometimes more accurate when the president evaluates the estimates than when he averages them because he learns by experience which men are usually more accur-ate than others and he can give the estimates of the more accurate ones greater consideration. In some firms the sales and marketing research departments prepare independent sales forecasts. These predictions are then considered by a jury consisting of the pres-ident, comptroller, the sales manager and the advert-i s i n g and market research d i r e c t o r s . Discussions are held f o r days or even weeks u n t i l a figure i s reached upon which a l l agree. The executive p o l l i n g approach has been reason-ably successful i n the companies where projections 7. M. Spencer, C. Clark, P. Hoguet, Business and Economic Forecasting, Homewood, I l l i n o i s , R.D. Irwin, 1961, p. 17. develop from a study and analysis of market reports, sales reports and business expectations. Many companies now a s s i s t t h e i r j u r i e s by supplying them with a quantity of f a c i a l background material which helps the members to understand past events and to convert t h e i r opinions about future trends into pre-di c t i o n s of future sales. Use of t h i s f a c t u a l material means that forecasts are based on more than judgment, so a more accurate evaluation can be obtained concerning the factors that a f f e c t sales. Although d e t a i l s may vary, the development of a sales forecast usually proceeds along s i m i l a r l i n e s . F i r s t the forecaster, a committee, or an outside consultant decides upon the d i r e c t i o n the economy i s expected to take during the forecast period. Whether the expected conditions are presented i n broad outline or i n det a i l e d statements they are accepted as the firm's assumptions. Past and present conditions i n the industry and the company are then studied and important trends are emphasized. Taking i n t o consideration a l l the fa c t s available and judg-ing t h e i r respective importance, the forecaster then projects these trends into the future. 8. Thompson, Forecasting Sales, p. 4. 86. At t h i s point the jury of executiv e opinion i s brought into the p i c t u r e . The jury can place i t s stamp of approval on the estimates, can question the thinking behind the forecast or can add specialized knowledge that has been gained by experience i n the f irmv There are several advantages to be gained when the forecaster u t i l i z e s the judgment of other execu-t i v e s i n the firm. The most important i s the pool-ing of the knowledge of s p e c i a l i s t s , because one man cannot be an expert i n a l l phases of the past, present, 9 and future operations of the company. The s p e c i a l -ized information i n i t s e l f i s not as s i g n i f i c a n t as the judgment of the i n d i v i d u a l s with the specialized knowledge. For example, the sales manager can a s s i s t the forecaster by informing him that channels of d i s t r i b u t i o n are undergoing rapid changes, and he can be of ev\u00C2\u00A7n greater help i f he can evaluate the e f f e c t s of the changes i n d i s t r i b u t i o n channels on industry sales and on the company's share of these sales. This method has the further advantage of 9. R.D. Crisp, \"Objectives and Techniques of Sales Forecasting,\" i n Sales Forecasting - Uses, Techniques and Trends, American Management Association, Special Report no. 16, 1956, p. 24. bringing diverse opinions together for consideration, which encourages consideration of a l l facets of the problem. Because senior executives always think i n terms of the future they apply themselves to i t s problems i n a more r e a l i s t i c and l o g i c a l manner than do salesmen, customers and others. Another advantage of t h i s technique l i e s i n the ease and speed with which i t can be employed. One reason for t h i s i s that the forecaster does not have t o be an expert i n the use of surveys or s t a t -i s t i c a l procedures, and another reason i s that d i s -cussing and judging are normal human a c t i v i t i e s that can be done e a s i l y and quickly i n the forecast-ing a c t i v i t y i f the executives are available, i f they are w i l l i n g to co-operate, and i f only a l i m i t e d number of sales figures are being forecast. The method i s also conducive to high morale because the executives p a r t i c i p a t e , or at least are consulted, i n the development of the forecast. This p a r t i c i p a t i o n also helps to secure t h e i r co-10 operation i n following through on the plans. 10. Crisp, Sales Forecasting - Uses, Techniques and Trends, p. 25. Although the advantages of the executive p o l l i n g method are numerous and important some weaknesses are apparent. One objection i s that personal opinion plays too strong a r o l e . In fa c t , the method can degenerate sometimes into group guessing. Further-more, one strong personality may dominate the proceed-ings. These disadvantages can be minimized i f there i s a good supply of fac t s , executives with experience and a forecaster who has the a b i l i t y and authority 11 to r e j e c t c e r t a i n opinions;; Another disadvantage i s the expense to the company i f executives spend a good deal of time at meetings, and even i f they do not give the matter excessive time, work schedules are disrupted to a c e r t a i n degree. A further disadvantage of the method i s that once the f i n a l forecast has been developed, the exec-utives must proceed to break down the estimate into seasonal and i n d i v i d u a l product forecasts so that production, purchasing, sales and finance plans can be made. One solution to t h i s major stumbling block i s to use past records as a basis f o r breaking down the forecast. But t h i s i s not very s a t i s f a c t o r y unless conditions remain unchanged and t h i s i s most 11. Thompson, Forecasting Sales, p. 4. u n l i k e l y . However, to produce a forecast f o r each i n d i v i d u a l item by the jury method would be an end-les s task i n companies with a number of l i n e s . As a r e s u l t , companies often use past experience as a basis for breaking down the forecast into seasonal and i n d i v i d u a l item figures and accept the f a c t that there w i l l be a number of sizeable errors. Sales Force Composite Method Many firms go beyond the executive l e v e l and obtain the assistance of t h e i r salesmen i n planning probable future sales. This i s c a l l e d the \"sales force composite method.\" There are widely d i f f e r i n g opinions concerning the e f f i c a c y of t h i s method, but since no published study of the system i s available an accurate appraisal cannot be made. In the sales force composite method each sales-man i s informed of the estimates required from him. He i s t o l d the period the forecast w i l l cover, the products that must be included, the extent to which customers' opinions are to be used, the method to be used i n c o l l e c t i n g the information and the time period during which the information i s to be gathered. This process of c o l l e c t i n g estimates from each salesman on the probable future sales for h i s t e r r i -t ory i s known as the \"grass roots\" approach. Some-times the salesman makes estimates on h i s own on forms provided for the purpose, and sometimes he makes them afte r consulting with the branch or regional manager. The l a t t e r method i s preferable because the sales manager has an opportunity to become f a m i l i a r with the salesman's reasons f o r h i s estimates, so the manager can modify whatever he believes i s out of l i n e . This approach i s , unfor-tunately, a heavy drain on the manager's time. Companies have found that they receive more accurate reports when salesmen have been supplied 12 with a record of past sales. These past sales figures may be supplied i n various forms such as special charts or s t a t i s t i c a l analyses of the variati o n s of past sales. The salesmen then have basic figures which they can increase or decrease, so t h e i r e f f o r t s are concentrated on these changes. When the salesmen have handed i n t h e i r estim-ates the r e s u l t s are accumulated f o r the d i s t r i c t 12. Crisp, Sales Forecasting - Uses, Techniques and Trends, p. 32. or region and sent to the c e n t r a l o f f i c e where they are studied and analyzed, and are then incorp-orated into a single a l l - i n c l u s i v e forecast which usually provides estimates of demand by t e r r i t o r i e s , 13 regions and products. As the forecast progresses through the organiz-ation from salesman to the top sales management group, the estimates are c a r e f u l l y inspected. Each d i s t r i c t manager examines h i s salesmen's estimates and also compares the t o t a l for the d i s t r i c t with past performances and with h i s own estimates of future sales. D i v i s i o n a l managers and senior sales managers also s c r u t i n i z e the forecast and make what-ever changes t h e i r judgment demands. Frequently a head o f f i c e group, usually the market or economic department or the treasurer's o f f i c e , makes an independent forecast which i s based on figures not normally available to the salesmen. These estimates act as a cross-check on the composite forecast formed from the salesmen's estimates. The causes of major differences between the two forecasts are c a r e f u l l y 13. H. Holmes, \"The Role of the F i e l d Sales Force\", i n Materials and Methods of Sales Forecasting, American Management Association, no. 27, p. 206. investigated before a f i n a l forecast i s made. These differences are often a t t r i b u t a b l e to the f a c t that the head o f f i c e can influence eventual perform-ance greatly by c o n t r o l l i n g the amount spent on advertising and promotion, by regulating production 14 and by endeavoring to enter or abandon certa i n markets. Because a great deal of time can be spent devel-oping a forecast by the sales force composite method, the forecast i s usually done on an annual ba s i s . There have been occasions, however, when the estimates have been developed semi-annually, quarterly and even weekly. One of the main reasons f o r using t h i s method in forecasting i s to p r o f i t from the knowledge the salesmen possess of l o c a l conditions. Unlike the senior executives who may be a thousand or more miles away, the salesmen have ample opportunity to become fa m i l i a r with the economic features of t h e i r t e r r i t -o r i e s . Top executives are provided with a l l the current relevant data that can be obtained, but the sales force composite method gives them the advan-tage of judgment developed at the customer l e v e l . 14. Spencer, Business and Economic Forecasting, p. 18. The l o c a l i z e d knowledge of salesmen may not always be a great asset, however. A great deal depends on the a b i l i t y of the salesman to acquire and interpret l o c a l f a c t s , as well as on the s i g n i f -icance of the l o c a l information to the company's sales. Also the type of information required by the company may be obtained from sources other than the customer, f o r published a r t i c l e s now provide a great deal of authoritative data on small areas. The popularity of the sales force composite method i s due la r g e l y to the f a c t that most of the r e s p o n s i b i l i t y for forecasting appears to rest upon 15 those who must meet the accepted goals. This method develops greater confidence i n the salesmen and sales executives than any other. When these men know how the sales estimates are calculated they are l i k e l y to accept them i n an attitude of co-oper-ation. When the estimates are developed i n a manner they do not understand the men regard them d i s t r u s t -f u l l y , e s p e c i a l l y i f they believe the quotas are too high. Attention should be drawn to the fa c t that the salesmen's estimates may undergo considerable 15. Holmes, Materials and Methods of Sales Fore-casting, p. 205. 94. r e v i s i o n , so the r e s p o n s i b i l i t y i s r e a l l y not a l l t h e i r s . Furthermore, the implication that the salesman i s the one responsible f o r sales seems open to question. His e f f o r t s i n h i s t e r r i t o r y are only a part of the e f f o r t s of the company as a whole. The sales force composite method i s said to give geographical and product breakdowns that are more accurate than those calculated i n the home o f f i c e . No evidence seems to be available to j u s t i f y t h i s assertion, for records show that both types of breakdown have been developed without consulting 16 salesmen. A f i n a l advantage claimed for t h i s technique i s the s i m p l i c i t y of operation and the fac t that detailed 17 forecasts can be developed e a s i l y . I f forecasts are required for establishing quotas and develop-ing production schedules t h i s method s i m p l i f i e s the procedure, for as the company grows subtotals are obtained f o r salesmen, t e r r i t o r i e s , regions and products. Other methods require considerably more work to break down t o t a l company estimates into 16. Crawford, Sales Forecasting, p. 26. 17. Thompson, Forecasting Sales, p. 8. subtotals for operating purposes. There are a number of disadvantages associated with the sales force composite method that should be considered. F i r s t , the method i s slow. Several weeks are required f o r making the estimates, having them checked and sending the composite forecast through the various l e v e l s f o r examination. Further-more, the amount of time spent at each l e v e l makes the technique expensive. A more v i t a l problem i s the a b i l i t y of the salesmen to forecast accurately, and not every salesman has t h i s a b i l i t y . Sometimes they err 18 badly. If a salesman does not have enough time to gather s u f f i c i e n t information he may resort to guessing. Again a salesman may have the best of intentions but be unable to forecast. The salesman who i s to make forecasts should be more than a taker of orders, he should be capable of u t i l i z i n g knowledge that the home o f f i c e cannot obtain, and he should know which information i s relevant to the developing of accurate sales forecasts. Further-more, he should have a c l e a r understanding of the 18. Spencer, Business and Economic Forecasting, p. 17. market forces at work i n h i s t e r r i t o r y , and have frequent contact with people whose positions enable them t o have an understanding of the economic forces that are shaping the future. Not every salesman has these a t t r i b u t e s . The high cost of the sales force composite method 19 i s t h i s technique's greatest weakness. Since forecasters are always comparing the expense as well as the accuracy of forecasting techniques, an examination of t h i s method w i l l encourage forecasters to consider other approaches. They w i l l probably select t h i s one only f o r some outstanding reason. There are two facts that must be borne i n mind by companies using t h i s method. F i r s t , the home o f f i c e should make sure that someone i s responsible f o r keeping a record of a l l forecasts made, with s p e c i a l reference to t h e i r accuracy. The sales force composite method should be compared with other methods and used only when sup e r i o r i t y i s assured. Second, the forecasts produced by t h i s technique must not be used as o v e r - a l l sales \"goals\". I f they are, appraisals are meaningless. 19. Crawford, Sales Forecasting, p. 28. 97 . Area Sales Manager Composite E f f o r t s have been made to overcome the disad-vantages of the sales force composite method while reta i n i n g the advantages. The method used with t h i s i n mind i s the area sales manager composite, and gives the r e s p o n s i b i l i t y of making estimates to area sales managers instead of to salesmen. This approach works well i n a firm where the branch managers t r a v e l extensively with the salesmen, for the managers become f a m i l i a r with the conditions i n the t e r r i t o r i e s and with the needs of the customers. Sales managers usually forecast competently, for they are not as emotionally biased as the salesmen when future sales are anticipated. Salesmen are personally involved i n the gain or loss of orders and tend to become optimistic or pessimistic according to t h e i r recent personal successes or f a i l u r e s . Sales managers can view the whole t e r r -i t o r y under them and can take a more detached view of the fluctuations i n sales of i n d i v i d u a l salesmen and therefore aire able to discern the trend of sales with greater accuracy. Sales managers are faster i n making up the estimates and forward-ing them to the head o f f i c e , p a r t l y because they have a stronger f e e l i n g of r e s p o n s i b i l i t y to head o f f i c e and p a r t l y because they r e a l i z e the import-20 ance of having a forecast that i s accurate. When a sales manager i s appointed to h i s p o s i t i o n one of the factors taken i n t o consideration i s h i s judgment. As a r e s u l t he should be more capable than a salesman of interpreting the si g n i f i c a n c e of l o c a l sales conditions i n the area. One aspect of the area sales manager composite method has a d e f i n i t e advantage over the sales force composite method. The task of i n s t r u c t i n g salesmen about the d e t a i l e d plans of management for the coming period i s a d i f f i c u l t one, but i f the sales-men do not know what those plans are they cannot provide a s a t i s f a c t o r y forecast. The area sales manager has knowledge both of management's plans and of l o c a l conditions and, therefore, he i s i n a good p o s i t i o n to weigh factors i n both spheres. One f i n a l advantage of t h i s method i s the cost. The forecaster can maintain contact with a smaller number of managers less expensively than he can with a greater number of salesmen. 20. Crawford, Sales Forecasting, p. 40. Use of Persons Outside the Firm The use of persons outside the firm cannot be c a l l e d a technique i n i t s e l f because these people are seldom able to consider the whole program planned by the company and are unable to derive a f i n a l sales projection. However, consultants are often brought into a firm to a s s i s t i n various ways, of which three seem most common. These con-sultants make projections of general business con-di t i o n s , they make estimates of industry sales and they a s s i s t i n solving p a r t i c u l a r l y d i f f i c u l t prob-lems such as the expected sales f o r a new product or a major change i n d i s t r i b u t i o n channels. Only occasionally does a consultant assume almost complete r e s p o n s i b i l i t y for the sales forecast of an i n d i v i d -ual firm. When a firm wishes to have a general business forecast developed extensive benefits can be derived by h i r i n g a consultant. The economic forecast can be presented when i t i s needed, the forecast w i l l not be biased, the procedure used by the consultant can be checked, the basic data used i n making the forecast can be evaluated, and the time period and area of in t e r e s t i n the economic estimates can be de t a i l e d to s a t i s f y s p e c i a l needs. Although using consultants for developing econ-omic forecasts has several advantages, using consult-ants f o r making industry or company forecasts i s not so desirable. A better job i s usually done i f a f u l l - t i m e s t a f f i s maintained to develop industry and company projections because speci a l i z e d know-21 ledge of the industry or company i s usually needed. On the other hand there are many situations where only a consultant can acquire and analyze the nec-essary data. A consultant, however, i s expensive, so a firm should not h i r e one unless h i s service i s necessary. Competitors sometimes prove useful i n a limited way. Sometimes a forecaster can make use of the judgment of a f r i e n d i n a competing firm, but t h i s i s not common p r a c t i c e . More information can be gathered through a trade association. Here the information i s usually l i m i t e d to past sales data and general information. A few trade associations have been able to persuade members to exchange forecasts of industry sales. Of the \"outside\" group, suppliers are least h e l p f u l and are of aid large l y to wholesalers and r e t a i l e r s . 2. Grunow, Materials and Methods of Sales Forecasting, p. 60. IV. TECHNIQUES BASED ON SURVEYS 101. The second c l a s s i f i c a t i o n covers techniques based on surveys. This device i s used i n an endeavor to acquire accurate forecasts by taking surveys of customer opinion by personal interviews. When t h i s method i s used customers are asked what they plan to buy i n the near future. I f a l l of the customers cannot be consulted the information may be obtained by sampling or by turning to the few who provide the major portion of the market. These surveys may 22 be taken f o r the industry or f o r the i n d i v i d u a l f irm. The i n d i v i d u a l company may take a survey of actual or p o t e n t i a l customers on a sample basis or by t r y i n g to reach every customer, but the r e s u l t s cannot be r e l i e d on. The information may be inaccur-ate f o r various reasons. If the customers purchase t h e i r products from more than one source they may t e l l the poll-taker what they expect t h e i r t o t a l requirements to be rather than what portion of t h e i r needs they expect to purchase from the company making the survey. The condition of the market w i l l also influence the customer. If a product i s 22. J . Dean, Managerial Economics, Prentice-H a l l , New York, 1951, p. 167. expected to be i n short supply the customer may magnify h i s requirements to the poll-taker, b e l i e v -ing that rationing on the part of the supplier w i l l 23 give him the quantity he r e a l l y wants. Surveys on the industry l e v e l have been conducted by government organizations, trade associations and publishing companies but the information has been of li m i t e d use to sales forecasters. Some trade associations have taken annual surveys and used them as a basis f o r issuing annual forecasts. The Survey Research Centre of the University of Michigan has done a good deal of survey work concerning consumer financing with sp e c i a l emphasis on durables and 24 housing. When the customers are i n d u s t r i a l producers the survey method puts the forecasting burden on the customers and the s e l l e r s ' chore i s to persuade h i s buyers to confide i n him t h e i r purchase or production plans and then to know what allowances to make for errors i n t h e i r r e p l i e s . If a buyer gives an honest estimate of h i s needs he a s s i s t s the s e l l e r i n making more accurate predictions provided he has 23. E. Neramers, Managerial Economics, J . Wiley and Sons, New York, 1962, p. 7. 24. Nemmers, Managerial Economics, p. 8. 103. some basis for estimating h i s own requirements. A buyer may base h i s estimates of future purchases on the present p r i c e structure and h i s present l e v e l of inventory. However, i f a change should occur i n the p r i c e structure of the goods he purchases, the buyer may decide to deviate from h i s planned inventory p o s i t i o n . As a r e s u l t , the s e l l e r to i n d u s t r i a l producers must modify the buyer's estimates i n keep-ing with short-run p r i c e movements. Data based on consumer opinions can be h e l p f u l i n p r e d i c t i n g sales. However, use of these data should be tempered with recognition of t h e i r l i m i t -ations. The p r e d i c t i v e record of consumer opinion data has been one of both successes and f a i l u r e s . Successes have been achieved i n predicting the general d i r e c t i o n of t o t a l consumer durable goods sales and in p r edicting the d i r e c t i o n of sales of large durable 25 purchases. Past f a i l u r e s of these data have been: f a i l u r e to indicate most of the r i s e s and f a l l s i n consumer durable goods sales, l i m i t e d success i n predicting accurately the d i r e c t i o n of appliance 25. S. Paranka, \"Marketing Predictions from Consumer A t t i t u d i n a l Data\" i n Journal of Marketing, American Marketing Association, Chicago, July, 1960, p. 50. sales, and a general lack of success i n predicting the volume of durable goods sales. Analysis of basic economic factors i n combination with consumer opinion data has been shown to improve s i g n i f i c a n t l y the accuracy of the forecasting record. Adoption of the survey technique i n sales f o r e -casting has not been widespread. The advantages and disadvantages to be discussed from a p r a c t i c a l point of view should shed some l i g h t on t h i s lack of acceptance, but l e t us f i r s t examine the method from a t h e o r e t i c a l viewpoint. Theoretically, the surveys assume a good deal. They assume that customers can foresee the future, know what conditions w i l l be for them and how they w i l l react to those conditions. They further assume that customers plan t h e i r purchases well i n advance, even determin-ing the s p e c i f i c product they w i l l select and the pr i c e they w i l l pay. They also assume that what a customer desires to buy he can afford to buy. In addition they assume that consumers w i l l report t r u t h f u l l y and thoroughly to the interviewer, and also that i f the consumer expects to maintain h i s economic 26 status he w i l l do so. Because these assumptions 26. Crawford, Sales Forecasting, p. 29. w i l l r a r e l y be met the t h e o r e t i c a l basis for f o r e -casts i s weak. From the p r a c t i c a l standpoint, however, there are advantages. The idea of considering the intentions of buyers i s l o g i c a l enough. Many i n d u s t r i a l products f o r instance, are bought i n large amounts at i n f r e -quent i n t e r v a l s so buyers make estimates during long planning periods. These buyers can supply f a i r l y accurate information f o r the survey interviewer. Sometimes information can be quite useful even when i t does not lead d i r e c t l y to a sales forecast, such as the information that a customer plans to obtain 27 h i s goods from a d i f f e r e n t source. I f the surveys are being made for other purposes, questions on buying intentions can be inserted at l i t t l e extra cost. There are various disadvantages to be found i n t h i s method. When a survey i s being prepared for householders every e f f o r t i s made to word the questions so that the customers being queried can be exact i n t h e i r r e p l i e s . The questions are con-fined to c e r t a i n products and are stated i n precise terras. Unfortunately many customers are not capable 27. Crawford, Sales Forecasting, p. 29. of estimating to what extent future happenings w i l l have a bearing on t h e i r purchasing and they do not always give answers that are complete. When a comparison i s made at the end of the year between the buying plans that were indicated i n the surveys and the buying a c t u a l l y done discrepancies are great enough to prevent a great deal of f a i t h i n t h i s kind 29 of forecasting. When planned expenditures are compared with actual expenditures f o r various years the d i r e c t i o n of the year-to-year changes do not correspond always with actual changes for durable goods and houses. One reason for f a i l u r e to receive more correct information by survey i s that a year i s a long time for a consumer to look into the future and most consumers adapt buying plans r e a d i l y to changes i n t h e i r economic condition or to other changes i n the home. Even a reaction to a change i n p r i c e s may not be foreseen e a s i l y . Sometimes consumers w i l l r e f r a i n from buying for a short time, thus i n s t i t u t i n g a buyer's s t r i k e , and sometimes they are so anxious to acquire goods they w i l l ignore the p r i c e change and purchase as planned. Another disadvantage that i s always present i n 28. Paranka, Marketing Predictions, July, 1960, p. 51. survey forecasting i s the d i f f i c u l t y of d i s t i n g -uishing between actual plans and wishful thinking. Afte r a l l , wishful thinking enters into a good many plans of everyone. Then, too, an interviewer presents the customer with two choices, to buy or not to buy, but l a t e r when the customer i s ready to make the purchase he finds there are many a l t e r -native ways of spending h i s money. When sampling i s c a r r i e d out by personal i n t e r -view i t i s one of the most expensive methods of 29 forecasting the sale of consumer goods. Some people believe that the cost of sampling i s t r i v i a l compared with the p r o f i t s that may be r e a l i z e d but few small consumer goods companies w i l l go to the expense of t h i s method when there are other methods that are f a r less c o s t l y and when r e s u l t s from the survey method are not necessarily s a t i s f a c t o r y . There are various factors which tend to decrease \u00E2\u0080\u00A2i. or increase the time required to complete a forecast based on the survey method. The required time i s decreased because p r i o r information i s seldom 29. R. Ferber, \"Sales Forecasting by Sample Surveys,\" Journal of Marketing, July, 1955, p. 2. required f o r the forecast, whereas when c o r r e l a t i o n techniques or methods using national income data are used time must be spent estimating the independ-ent variables or the i n d i v i d u a l components of income. At the same time, the survey method i s slowed down because of the extended period required to c o l l e c t the data, e s p e c i a l l y i f the c o l l e c t i n g i s done by 30 m a i l . The survey approach can be very successful i f an i n d u s t r i a l product such as heavy machinery i s sold by the manufacturer d i r e c t l y to the user. Here the salesmen can interview the customers when they make t h e i r c a l l s so the sampling i s r e l a t i v e l y inexpensive. This advantage does not apply when the i n d u s t r i a l products are not sold d i r e c t l y to the ultimate users. A sample survey may stand the best chance of succeeding when i t forecasts the sale of items i n two categories. The f i r s t type i s a good that i s bought after the purchase has been c a r e f u l l y consid-ered and planned f o r , so the interviewer learns when the purchase i s l i k e l y t o be made and the character-30. Perber, Journal of Marketing, July, 1955, p. 2. i s t i c s desired i n the product to be bought. The second type i s a good purchased often and with l i t t l e thought, the sales of which are known to be related to such factors as income or the number of f a m i l i e s . However, few products show a co n s i s t -ent or dependable r e l a t i o n to either of these f a c t o r s . The sampling method may increase i n accuracy with the size of the expenditure involved and 31 decrease with the frequency of the purchase. The l a t t e r indicates the r e l i a b i l i t y with which information can be obtained regarding purchases made i n the past, but i t s v a l i d i t y f o r predicting behavior i n the future has not been established. The shorter the i n t e r v a l before the forecast i s desired, the more accurate i s a sales forecast based on sampling l i k e l y to be. Short i n t e r v a l s provide less time for new developments which might change the plans people previously made or a l t e r movements of business a c t i v i t y . The accuracy of a sample-based sales forecast i s not necessarily inversely related to the length of time covered by the fore-cast. This i s e s p e c i a l l y true where the d i r e c t 31. Ferber, Journal of Marketing. July, 1955, p. 2. approach i s used. Up to a c e r t a i n point the increase i n the period covered by the forecast may allow time for previously stated plans to become e f f e c t i v e . Beyond that period, however, the i n -creasing number of purchases -represented by plans and developments made after the sample date w i l l more than n u l l i f y advantages gained from extending the period. Surveys of buying intentions rest upon the assumption that future conduct depends p a r t l y on the people's present b e l i e f s and plans concern-ing the future. To the extent that present plans are only a ' p a r t i a l basis f o r future conduct, and t o the extent that expected conditions which form the groundwork for those plans f a i l to materialize and so change future conduct, the survey w i l l be 32 inaccurate as a forecasting t o o l . V. TECHNIQUES BASED ON STATISTICAL METHODS The most popular means of supplementing personal judgment i n order to increase the accuracy of sales forecasting i s to use methods of s t a t i s t i c a l 32. J . Schweiger, \"Forecasting Short-term Consumer Demand\" Journal of Business, University of Chicago Press, January, 1956, p. 99. c o r r e l a t i o n . This i s done by noting the r e l a t i o n -ships and movements i n a company's or industry's sales and then measuring them s t a t i s t i c a l l y and projecting them. The purpose of the c o r r e l a t i o n approach i s to f i n d a mathematical equation c a l l e d an \"estimating equation\", \"predicting equation\" or \"regression equation\" Which best shows the r e l a t i o n -ship between a dependent variable and one or more independent varia b l e s . I f we take the sale of automobiles as our dependent variable, and income, number of f a m i l i e s , replacement rates and so on as our dependent variables we may predict variations i n the sale of automobiles on the basis of v a r i a t i o n s i n the selected independent va r i a b l e s . Such factors are chosen because, on the basis of knowledge and l o g i c they are considered to be the c o n t r o l l i n g ones. The s t a t i s t i c a l analysis i s c a l l e d simple c o r r e l a t i o n when only one independent variable i s used, but i s known as multiple c o r r e l a t i o n i f two or more independent variables are involved. In c o r r e l a t i o n analysis the regression equation that ultimately arises expresses the change i n one series of data which tends to occur with a given change i n one or more independent series of data. The r e l a t i o n s h i p s i n c o r r e l a t i o n analysis may be i l l u s t r a t e d as follows. I f we l e t Y denote the sales of a product and X denote the p r i c e of the product, variations i n Y w i l l depend on variatio n s i n X, so the r e l a t i o n -ship can be written conceptually as Y = f ( X ) . This i s read as \"Y i s a function of X\", or \"sales are a function of p r i c e . \" What i s being said i s that a dependent or functional r e l a t i o n s h i p e x i s t s between the two variab l e s . The re l a t i o n s h i p i n the equation i s one of simple c o r r e l a t i o n because only one independent variable i s involved. I f further analysis shows that other factors such as income and the number of fam i l i e s have an important influence on sales i n addition to price, the function can be written Y = f(X^, X 2, X 3 ) , where Xj_ s t i l l r e f e r s to p r i c e and X 2 and X^ refer to income and number of fam i l i e s respectively. This equation now reads, \"Y i s a function of X-j_, X 2 and X 3\" or \"sales are dependent on price, income and the number of familes.\" Here the r e l a t i o n s h i p i s known as multiple c o r r e l a t i o n since more than one independ-ent variable i s involved. The purpose of simple or multiple c o r r e l a t i o n i s to a r r i v e at the actual equation of r e l a t i o n s h i p among the variables, instead of the conceptual ones stated above. That i s , the i n d i v i d u a l independent variables must be weighted according to the importance they derive from t h e i r e f f e c t 33 on the dependent v a r i a b l e . A more detailed description of c o r r e l a t i o n analysis w i l l now be undertaken. Instead of p l a c -ing the emphasis on computational techniques we s h a l l focus attention on the fundamental concepts. Simple Correlation I f only two variables are involved, a dependent and an independent, the measurement of the r e l a t i o n -ship which exists between the two series of data i s known as simple c o r r e l a t i o n . The r e l a t i o n s h i p can be written conceptually as Y i s a function of X, or i n the symbolic form Y = f ( X ) . Where there i s simple c o r r e l a t i o n a straight l i n e may best describe the r e l a t i o n s h i p between the dependent and independent varia b l e s . I f the r e l a t i o n s h i p between the two variables, as shown by a scatter diagram, i s l i n e a r , the s t a t i s t i c a l procedure for reaching the equation of r e l a t l o n -3 3 . Spencer, Business and Economic Forecasting, ship i s known as simple l i n e a r c o r r e l a t i o n . When a unit change i n one variable produces a constant change i n the other variable over the ent i r e relevant range of the data, that i s , when the slope 34 i s constant, the r e l a t i o n s h i p i s l i n e a r . The equation to show t h i s i s Y = a + bX, the equation for a straight l i n e , where Y i s the dependent variable and X the independent v a r i a b l e . For t h i s r e l a t i o n s h i p a and b are not known. They are con-stants whose values we wish to determine, and we usually use the least squares technique to estimate t h e i r values. The method of least squares gives the l i n e of best f i t under the assumptions of that method: a normal d i s t r i b u t i o n of the observations around the l i n e and the reduction of the squared residuals to a minimum. The le a s t squares technique produces consistent, unbiased r e s u l t s , i s f a i r l y easy to use and i s the f a m i l i a r procedure learned i n elementary s t a t i s t i c s . An a l t e r n a t i v e method for estimating the constants, the method of maximum li k e l i h o o d , chooses the value which makes the probab i l i t y of occurrence of the estimate a maximum. In most cases the methods of least squares and of 34. M. Ezekiel, Methods of Correlation Analysis John Wiley and Sons, New York, 1947, p. 36. maximum l i k e l i h o o d estimation produce exactly the 35 same r e s u l t s . The value of a i n the equation i s the value of Y at the mathematical o r i g i n of the equation and b represents the average amount of change i n Y which occurs with each unit change i n X. I f we f i n d that a = S and b = 3, the pr e d i c t -ing equation would then be Y =8 + 3X. With the equation Y = 8 3X, a unit increase i n X w i l l cause Y to r i s e i n value by three units, regardless of whether X increases from one to two, or from f i f t y - o n e to f i f t y - t w o . The slope i s constant and equals three. I f we forecast the value of X for a c e r t a i n period i n the future the value of Y fo r that period can be predicted by substituting f o r X i n the equation. When an increase i n the independent series i s accompanied by an increase i n the dependent series, the r e l a t i o n s h i p i s said to be p o s i t i v e . P o s i t i v e l y correlated data on a chart would proceed from the lower l e f t to the upper r i g h t . A negative c o r r e l -ation exists when increases i n one series are accompanied by decreases i n the other s e r i e s . A l i n e showing t h i s r e l a t i o n s h i p would proceed from 35. Spencer, Business and Economic Forecasting, p. 372. 116. the upper l e f t of a chart toward the lower r i g h t . A l i n e drawn on a scatter diagram to depict the r e l a t i o n s h i p between two series of data may be curved. I f the l i n e on the chart shows a c u r v i l i n e a r r e l a t i o n s h i p , the purpose of the s t a t i s t i c a l analysis i s s t i l l to derive an equation which best appears to explain the r e l a t i o n s h i p between the two variab l e s . The s t a t i s t i c a l process i s known as \"simple c u r v i l i n e a r c o r r e l a t i o n \" . Simple c u r v i l i n -ear c o r r e l a t i o n expresses a changing r e l a t i o n s h i p between the two variables instead of the uniform l i n e a r r e l a t i o n shown by the straight l i n e . In the same way that r e l a t i o n s were represented mathematically by a straight l i n e , r e l a t i o n s can be represented by curves of various types. While only one equation Y = a + bX i s used to represent any straight l i n e by determining the proper values to be assigned to the constants a and b, there i s p r a c t i c -a l l y no l i m i t to the d i f f e r e n t kinds of curves which can be s i m i l a r l y described by mathematical 36 equations. For everyday business research, how-ever, many of the common types can be defined by 36. Ezekiel, Methods of Correlation Analysis, p. 76. a small number of rather simple equations. When this i s done the result i s known as a \"type equation\" and each type equation represents an entire family of curves. Figure 1 represents six families of curves often encountered in economic measurement and fore-casting as well as their corresponding type equation. FIGURE I Six .Families of Curves l o o a n t h m i c curve r e t i b r o c a l c u r v e Jog Y \" log + b log X Y = <*\u00E2\u0080\u00A2 + k X 1 1 8 . The c a p i t a l l e t t e r s i n the equations represent the variables while the lower case l e t t e r s represent the parameters. Whether the curves follow the s o l i d or the dashed pattern depends on whether the sign of the parameters i s plus or minus. Which-ever the sign, the basic forecasting problem i s to employ s t a t i s t i c a l procedures to f i n d the values of the p a r t i c u l a r parameters for the p a r t i c u l a r curve of the family which best seems to f i t the data. Multiple Correlation Situations are often encountered where the variation s i n the dependent variable can be explained more f u l l y i f more independent variables are included i n the predicting equation. When two or more independent variables are to be included i n a regression equation the s t a t i s t i c a l analysis i s c a l l e d multiple c o r r e l a t i o n . When two independent variables are used the functional r e l a t i o n s h i p i s expressed conceptually as Y = f(X^# X 2 ) \u00E2\u0080\u00A2 Additional X's may be included to represent additional c o n t r o l l i n g f a c t o r s . The reasons for using multiple c o r r e l a t i o n are the same as the reasons for using simple c o r r e l a t i o n . A regression equation i s found which best f i t s the observed data and thus serves as a foundation f o r future predictions. Multiple c o r r e l a t i o n , l i k e simple c o r r e l a t i o n , may be l i n e a r or c u r v i l i n e a r . A l i n e a r r e l a t i o n -ship between the variables i n the regression equation i s c a l l e d \"linear multiple c o r r e l a t i o n \" , and appears when a straight l i n e best represents the scatter diagrams between the dependent variable and each of the independent variables, that i s , the Y = X^ re l a t i o n s h i p and the Y = X 2 r e l a t i o n s h i p , and so on. For two independent variables involving l i n e a r multiple c o r r e l a t i o n , the regression or predicting equation would take the form Y = a + bX^ + c X 2 ' a n < ^ f o r three independent variables Y = a + bX^ + cX 2 + dX-j. Each X represents a d i f f e r e n t independent variable, and Y i s depend-ent. I f a fourth independent variable was included the equation would have eX^ added on, and i f there was a f i f t h , f X 5 would be added on. The addition of further variables may improve the a b i l i t y of the equation to predict changes i n Y that are caused by va r i a t i o n s i n each of the X's. The c o e f f i c i e n t s b, c, d, etc. represent the rate of change i n the dependent variable per unit change i n each of the independent variables, while the other independent variables are held constant. These c o e f f i c i e n t s are usually c a l l e d \" c o e f f i c i e n t s of net regression\" to d i s t i n g u i s h them from \" c o e f f i c i e n t s of gross regression\" i n simple c o r r e l a t i o n , where no allowance i s made f o r i n d i r e c t influences on the regression. As i n simple c o r r e l a t i o n , the object of the s t a t i s -t i c a l analysis i s to determine the best estimates of the parameters of the regression equation and so enable predictions to be made based on past r e l a t i o n s h i p s . Although l i n e a r multiple c o r r e l a t i o n i s a sa t i s f a c t o r y t o o l f o r most multiple c o r r e l a t i o n problems, c u r v i l i n e a r correlations are e n t i t l e d to at least a short explanation. In the above l i n e a r regression equation, the value of Y changes at a constant rate with respect to changes i n each indep-endent value. In graphic terms the correlations Y = X^, Y = y^2> etc. are straight l i n e s on the scatter diagrams; i n mathematical terms the regression equation involves only the f i r s t powers of the independent v a r i a b l e s . I f Y should change at increasing or decreasing rates with respect to each of the X's, the c o r r e l a t i o n l i n e s on the separate scatter diagrams would be curved rather than s t r a i g h t . Mathematically the regression equation would involve powers greater or less than one for the independent vari a b l e s . Where the dependent var i a b l e shows a c u r v i l i n e a r r e l a t i o n s h i p with one or more of the independent variables we have c u r v i l i n e a r multiple c o r r e l a t i o n . Although the ca l c u l a t i o n s required t c o f i n d the parameters of the regression equation are more complicated and d i f f i c -u l t , the p r i n c i p l e s are the same as i n the l i n e a r case. Some examples of c u r v i l i n e a r multiple regression equations follow: 2 2 Y = a + bX 1 + cX 2 + dx 3 + eX 4 Y = a + bX L + c X 2 2 + dX 3 3 + e X 4 4 Y = a + bXj, + c X 2 2 + dX 3 3 + e X 4 2 As i n simple c o r r e l a t i o n , the problem i n multiple c o r r e l a t i o n i s to choose the equation that represents the r e l a t i o n s h i p between the independent and depend-ent v a r i a b l e s . However, tie theory upon which the equations are based i s often not e n t i r e l y appropriate to the task on hand, and the best that can be accomplished i s to choose the equation that most nearly f i t s the data. When the equations by which values of one variable may be estimated from those of two or more independent variables have been worked out, i t i s frequently desirable to have some measure of how c l o s e l y such estimates agree with the actual values and of how c l o s e l y the v a r i a t i o n i n the dependent variable i s associated with the v a r i a t i o n i n the several independent v a r i a b l e s . The standard error of estimate for a multiple regression equation measures the closeness with which the estimated values agree with the o r i g i n a l values. From the dependent and independent variables can be developed a measure of co r r e l a t i o n , the square of which i s known as the co e f f l e n t of multiple determination. This describes the proportion of t o t a l v a r i a t i o n i n Y explained by the X's i n the equation. Measures of p a r t i a l c o r r e l a t i o n , or degree of association, and of p a r t i a l determination, or proportion of v a r i a -t i o n , can also be developed between the dependent variable and any combination of independent variables i n the regression equation. Advantages of Correlation The use of simple c o r r e l a t i o n presents advantages f o r the user. The forecaster i s always hoping to f i n d an independent variable which fluctuates at a f i x e d i n t e r v a l before sales f l u c t u a t e . I f the c o r r e l a t i o n i s based on a series that leads h i s company sales he does not need to forecast the independent series, he needs only to s c r u t i n i z e anything that may destroy the past relationships 3 7 upon which he based h i s c o r r e l a t i o n . Even i f no lead-lag r e l a t i o n s h i p can be found a good cor-r e l a t i o n can be useful because the related series may be the concern of other forecasters and so be available i n forecasted form. I f sales are corre-lated to a well-known series the forecaster benefits from the opinions of other forecasters. Many persons forecast national income figures, so i f a good r e l a t i o n s h i p can be found between sales and national income the forecaster using national income figures can estimate sales very e a s i l y . The f o r e -casters c a l c u l a t i n g the national income series may be wrong, but the p o t e n t i a l error w i l l probably be less f o r them than for sales forecasters who work in small firms and who lack either the experience or the funds necessary f o r a comprehensive study of the problem. 3 7 . Crawford, Sales Forecasting, p. 3 4 . Although problems are encountered when mul t ip l e regress ion equations or demand functions are developed, the gains may be great enough to j u s t i f y the use of t h i s technique. In the f i r s t place the sales forecast ing i n most f irms i s based l a r g e l y on a combination of i n t u i t i o n , experience and t r a d i t i o n . I f thorough attempts are made to develop demand functions some cherished ideas con-cerning the estimating of sales can be found to be useless , and even i f the r e a l fac tors respons-i b l e for deviat ions cannot be measured, knowing what those fac tors are i s an advantage. Secondly, a poor ly p r e d i c t i n g mul t ip le regress ion equation can sometimes be useful by the nature of i t s misses. The most important demand factors are usua l ly included i n a funct ion , so a miss shows 38 that an unusual;factor i s having an in f luence . Just knowing that such an event i s happening can be u s e f u l . T h i r d , one of the most important reasons for the l i m i t e d use of mul t ip le regress ion equations i n the past has been the lack of s u f f i c i e n t data to p r e d i c t the movement and e f fect of the various 38. Crawford, Sales Forecast ing , p . 45. sales f a c t o r s . The supply of suitable data i s constantly increasing, so forecasters should consider using the function as soon as the data they require are av a i l a b l e . Disadvantages of Correlation Determining the constants for the simple l i n e a r regression equation f o r a given series of data i s c a l l e d \" ' f i t t i n g ' the equation to the data\". Because the simple l i n e a r regression equation i s the simplest of a l l equations to \" f i t \" , i t i s widely used. However, the l i n e a r equation i s extremely li m i t e d i n i t s l o g i c a l meaning. The simple l i n e a r equation can represent only a si t u a t i o n where the change i n the dependent variable, for a unit change i n the independent variable, would be expected to be just the same no matter how large or how small the independent variable was. This i s a narrow, precise r e l a t i o n -ship. In many situations, the r e l a t i o n which would be expected would be a changing r e l a t i o n -ship as the value of the independent variable changed and not t h i s unchanging r e l a t i o n s h i p . F i t t i n g a straight l i n e can be regarded only as an empirical exercise, with no meaning being attached to the constants beyond the purely formal one of specifying the straight l i n e that most nearly represents the data, unless the forecaster has a good l o g i c a l reason to expect the li n e a r 3 9 equation to represent honestly the true s i t u a t i o n . A s i g n i f i c a n t disadvantage that faces the user of both simple and multiple c o r r e l a t i o n i s the danger of leaning too heavily upon s t a t i s t i c a l projections and neglecting to make an independent objective appraisal of the future. Past trends must always be considered when new trends are being predicted, but the forecaster must be ever a l e r t to those factors that can cause sudden and severe deviations from the past. Relationships often f a i l to hold true f o r an extended period of time, and the danger i s always present of the forecaster ceasing to be a l e r t to changes i n the factors that produced the i n i t i a l r e l a t i o n s h i p . Despite the f a c t that multiple regression equations can be useful they w i l l probably never be u n i v e r s a l l y accepted because there are math-ematical weaknesses, l o g i c a l f a l l a c i e s and 3 9 . Ezekiel, Methods of Correlation Analysis, p. 7 3 . 40 pragmatic l i m i t a t i o n s that bar such acceptance. An a r b i t r a r y decision must be made i n choosing the dependent and independent variables to be included i n the regression equation. As a r e s u l t , important independent variables may be overlooked even though an e f f o r t usually i s made to include the most important va r i a b l e s . In addition, the regression equation produces r e l i a b l e r e s u l t s only when the forecasted figures l i e within the range of figures used to develop the forecasting equation. The meaning attached to the successful f i t t i n g of a demand equation i s e a s i l y misunderstood. High co r r e l a t i o n s do not necessarily give proof of cause and e f f e c t and forecasters should not succumb to the temptation to assume that they do. A second weakness i n logic i s the usual assumption that a demand function i s s t a t i c . A s t a t i c re-l a t i o n s h i p means that the r e l a t i o n s between the dependent and independent variables never change. St a t i c relationships simplify computations but seldom r e a l l y exist, or at the most seldom e x i s t f o r any length of time. A s t a t i c equation tends 40. Crawford, Sales Forecasting, p. 47. to average out the e f f e c t s of the variables over a period of time, so the e f f e c t s over l a t e r years, which are the most important ones i n sales f o r e -casting are l i k e l y to be obscured by the e a r l i e r forces. There are other less important weaknesses i n l o g i c . For instance, time i s often included as a variable i n the equation to represent slow moving forces of a minor nature, but the in c l u s i o n i s often u n j u s t i f i a b l e , e s p e c i a l l y over many years, because these forces often have an unknown or changing r e l a t i o n s h i p with the dependent variable and cannot be represented e f f e c t i v e l y by a time v a r i a b l e . F i n a l l y , multiple regression equations sometimes y i e l d nonsense r e s u l t s . Spurious cor-r e l a t i o n s can aris e as the r e s u l t of coincidental appearance of relationships between series of data which have no causal or l o g i c a l connection. There are other weaknesses that may make the use of demand functions impractical. F i r s t , some information, such as competitive strategy, cannot be put i n numerical form and other required figures may not be available i n the form desired. Also, few firms possess suitable sales data for a period of from ten to twenty years, without which the demand function has less chance of y i e l d i n g s a t i s -factory r e s u l t s . Often only annual data are available f o r multiple c o r r e l a t i o n analysis. Since most companies require estimates for shorter periods, the demand function must be supplemented by other techniques. Another weakness i s the f a c t that some variable usually must be forecast. By the time several variables have been forecast the c o l l e c t i v e errors can often equal or even exceed the error of a forecast based on only an opinion. The problems encountered i n the use of multiple c o r r e l a t i o n require such s k i l l i n the solving that few forecasters are prepared to cope with the technique. Generalizations concerning the usefulness of co r r e l a t i o n cannot be made because the p o t e n t i a l i t i e s of the method are dependent e n t i r e l y upon the c i r -cumstances found i n each i n d i v i d u a l s i t u a t i o n . Sometimes cor r e l a t i o n s have been useful, but at other times relationships that appeared to be dependable proved to be useless. Each forecaster must examine h i s own s i t u a t i o n and make h i s own decisions. The forecasters who have not considered the p o s s i b i l i t i e s of using c o r r e l a t i o n have over-looked a useful t o o l . S e r i a l Correlation The r e l a t i o n s h i p between successive observations i n the same series of data i s referred to as s e r i a l c o r r e l a t i o n . This type of c o r r e l a t i o n i s found most often i n the case of time series, where the value of the variable at one period of time i s believed to influence the value i n a succeeding period. The major problem arises i n determining whether or not successive items i n a 41 series are s e r i a l l y correlated. In other types of c o r r e l a t i o n , i n t e r e s t was focused on the magnitude of the c o r r e l a t i o n , but in s e r i a l c o r r e l a t i o n i n t e r e s t i s centered on ascertaining the presence of c o r r e l a t i o n . This i s done because most of the sampling formulas and procedures used i n practice assume that successive observations are independent of each other, and when t h i s assumption no longer holds, most a n a l y t i c a l methods are invalidated. Exect means for measur-ing the extent of bias due to s e r i a l c o r r e l a t i o n are not yet available, as discovering the presence of c o r r e l a t i o n i s the important thing, and measuring 41. R. Ferber, S t a t i s t i c a l Techniques i n Market Research, McGraw-Hill, 1949, p. 402. the magnitude of the c o r r e l a t i o n becomes a second-ary problem. In time series each figu r e i s seldom independ-ent of the preceding f i g u r e . I f s e r i a l c o r r e l a t i o n a f f e c t s a time series to a great enough degree, the future of time series may be forecast with considerable accuracy by a technique that i s mechanical, simple and inexpensive. The f i r s t step i n employing;: the procedure i s to use various s t a t i s t i c a l methods to discover the pres-ence and nature of s e r i a l c o r r e l a t i o n i n the d i f f e r e n t s e r i e s ; the second step i s to develop and t e s t d i f f e r e n t models. A model i s an assump-t i o n stated i n algebraic terms. A sales manager i s using a simple model when he states that sales of every item for six months of the year w i l l equal the sales for the same six month period of the preceeding year. In algebraic terms the manager i s saying that expected sales (E t) equal past sales (&t_2) \u00E2\u0080\u00A2 I f t l i e f i * * 1 1 1 forecast a ten percent increase each year the algebraic terms used would 42 be E. = A t_ 2 + 10% ( A4-_?) \u00E2\u0080\u00A2 S u c h unsophisticated 42. Crawford, Sales Forecasting, p. 43. models as t h i s are seldom s a t i s f a c t o r y , so the forecaster w i l l t r y various methods i n order to discover the best one f o r h i s s e r i e s . Any usefulness provided by an unsophisticated model i s usually i n the form of a yardstick f o r measuring the accuracy of forecasts developed by more sophisticated methods. A forecaster can t e s t an unsophisticated model on h i s past sales, c a l c u l -ate the magnitude of the error, and compare the s i z e of t h i s error with the s i z e of the error that arises from using a more sophisticated technique. He can also use the cost basis for comparing the errors i n order to discover the precise cost of more accurate forecasts. Time Series Analysis At any given point i n time, sales are influenced by four major factors, long-term trends, c y c l i c a l v a r i a t i o n s , seasonal v a r i a t i o n s and i r r e g u l a r f l u c -tuations. Analysis of the h i s t o r i c a l patterns of the f i r s t three factors may prove h e l p f u l i n f o r e -casting sales and business a c t i v i t y ; the fourth factor, i r r e g u l a r fluctuations, has no pattern and defies attempts at systematic forecasting. The analysis of movements of series of data over periods of time i s referred to as \"time series analysis In t r a d i t i o n a l time series analysis, the assumption i s made that any p a r t i c u l a r value i n a series i s the product of factors that can be attributed to the various components. The secular trend of a time series refers to the smooth or regular movement of the series over a long period of time. I n t u i t i v e l y speaking, the trend of a time series characterizes the gradual and cons-4 3 i s t e n t pattern of i t s changes. Some series of data recorded over a given period of time show an upward trend, some show a downward trend and some remain at a l e v e l that i s reasonably constant. Certain factors i n t h i s country have caused the trend of many important economic series to be upward. Upward movements have been apparent in both i n d u s t r i a l production and t o t a l personal income. Trends are of endless v a r i e t y . Some series increase slowly, some increase quickly, others 4 3 . E. Chambers, Economic Fluctuations and Forecasting, Prentice-Hall, Englewood C l i f f s , New Jersey, 1961, p. 7. decrease at varying rates of speed, \tfhile s t i l l others remain r e l a t i v e l y constant for long periods of time. Some series go through a period of growth or decline and then change d i r e c t i o n and go into a period of decline or growth. The easiest v a r i a t i o n of a time series to under-stand i s the seasonal one which consists of patterns regu l a r l y repeated when the length of the pattern i s a year or les s i n duration. The study and measure-ment of seasonal patterns i s e s s e n t i a l i n the analysis of a time s e r i e s . Sometimes the seasonal patterns themselves are very important because a knowledge of seasonal patterns based on adequate s t a t i s t i c a l measures i s necessary as a basis f o r planning and scheduling. At other times the seasonal pattern i s useful as a means of measuring other variati o n s of a time s e r i e s . The point of view i s sometimes taken that i f a time series has i t s trend, seasonal v a r i a t i o n and i r r e g u l a r fluctuations removed, then what remains i s the so-called business c y c l e . This i s probably an ove r s i m p l i f i c a t i o n of the si t u a t i o n , but a common way of measuring a business cycle i s to use t h i s process of elimination. A business cycle can also be described as consisting of a recurrence of the up and down movements of business a c t i v i t y from some sort of s t a t i s t i c a l trend or normal. By normal we mean something i n the nature of a s t a t i s t i c a l average. We are not considering anything p a r t i c u l a r l y perm-anent or universal i n the word normal. Business cycles d i f f e r from seasonal v a r i a t i o n s i n the length of the time period covered, the bus-iness cycle being longer than the seasonal v a r i a t i o n . Further, the fluctuations i n a business c y c l e are thought to have d i f f e r e n t causes than the f l u c t -uations i n seasonal v a r i a t i o n s . Prosperity, reces-sion, depression and recovery are sometimes consid-ered to be the four phases of a business cycle and they are caused by factors other than weather, s o c i a l customs and other s i m i l a r factors that create seasonal patterns. Those business cycles which show enough s i m i l a r i t y to be i d e n t i f i e d as such unfortunately shew so much d i s s i m i l a r i t y as to make predictions of t h e i r future occurrence, length and severity of l i t t l e 44 value. 44. R. McLaughlin, Time Series Forecasting, American Marketing Association, 1962, p. 8. When the fluctuations of a time series are completely unpredictable, or are caused by unrelated but potent factors such as good or bad news, bank f a i l u r e s , elections, floods, earthquakes, s t r i k e s or wars, they are c a l l e d i r r e g u l a r or e r r a t i c v a r i a t i o n s . Some influences cause disturbances that are strongly f e l t , while some cause d i s t u r b -ances that work themselves out before they are strongly f e l t , but both types are c l a s s i f i e d as e r r a t i c . For p r a c t i c a l purposes any v a r i a t i o n that does not account for trend, seasonal or c y c l i c a l move-45 ments i s classed as i r r e g u l a r or e r r a t i c . I f trend, seasonal and c y c l i c a l movements are having an influence they produce c e r t a i n systematic e f f e c t s , while i r r e g u l a r movements, which r e s u l t from chance factors, produce random e f f e c t s which are completely unpredictable when taken one at a time, but which tend to average out over the long run. When an investigator analyzes a time series he i s usually interested i n the v a r i a t i o n s that take place during successive time periods. He may seek 45. McLaughlin, Time Series Forecasting, p. 9. to know i f there i s a recurring seasonal pattern i n the sale of lumber, or to ascertain what the pattern of change i s i n the volume of i n d u s t r i a l production during business c y c l e s . In an endeavor to answer these questions, the investigator t r i e s to i s o l a t e the movements of immediate in t e r e s t from a l l the other movements that influence the series under observation. He uses a process c a l l e d decomposition 46 with t h i s end i n view. The problem confronting the investigator i s , \"How are the d i f f e r e n t move-ments blended together to make up the h i s t o r i c a l series that i s a c t u a l l y recorded?\" The use of de-composition does not necessarily supply the answers. The p a r t i c u l a r process of decomposition that i s employed depends upon c e r t a i n assumptions pe r t a i n -ing to the manner i n which the e f f e c t s of d i f f e r e n t forces are combined. Some of these assumptions may be more acceptable than others. The f i r s t of these movements, the trend, i s commonly used to obtain a general v i s u a l estimate, not to develop a precise s t a t i s t i c a l p r e d i c t i o n . However, the trend can be h e l p f u l used either way. 46. F.C. M i l l s , S t a t i s t i c a l Methods, H. Holt and Co., New York, 1955, p. 323. A r e l i a b l e procedure f o r exposing the trend i s described i n the following paragraph. F i r s t , the sales are plotted on arithmetic and semi-logarithmic graph paper. Second, a study i s made of the general movement. Third, a l o g i c a l state-ment concerning the trend i s developed, based on a knowledge of the h i s t o r y of the series concerned. Fourth, a formula i s selected that embodies the same movement and i s applied to the sales f i g u r e s . The trend has many advantages regardless of the way i n which development takes place. F i r s t , the trend gives the forecaster a base from which to evaluate deviations, f o r by focusing attention on t h i s movement the forecaster i s not influence^too greatly by strong c y c l i c a l or i r r e g u l a r factors of a temporary nature. Second, i f he can develop pro-jections oi; the trend, and also of the c y c l i c a l move-ment and the seasonal movement, he can t h e o r e t i c a l l y t o t a l them f o r an estimate of sales. This i s s e l -dom done, mainly because the c y c l i c a l v a r i a t i o n i s so d i f f i c u l t to i s o l a t e . Third, when there i s a p o s s i b i l i t y of c o r r e l a t i n g sales with some other variable, the c o r r e l a t i o n i s usually more meaningful i f the trend has been i s o l a t e d from both the sales 47 and the other v a r i a b l e . The sales or other data must be consistent over time, that i s , they must be homogeneous. This con-sistency i s usually disregarded where a trend i s sought i n t o t a l company sales. A company seldom produces and s e l l s the same product f o r twenty years or more, so the sales figures for d i f f e r e n t years usually cover d i f f e r e n t items, that i s , they are not homogeneous. To obtain consistent data the trend should be sought either i n sales series for each product or i n t o t a l industry sales for each type of product, e s p e c i a l l y i f the share-of-market fluctuates greatly. Another means of attempting to obtain consistent data i s to remove the var i a t i o n s caused by price.. This i s done because there may be a separate trend f o r pr i c e , which, when added to the unit sales trend produces an o v e r - a l l trend that i s heterogenous. Furthermore, the trend should be stressed for recent years since changes may be occur-r i n g continuously. F i n a l l y , when the forecaster establishes a trend l i n e he must choose the years to be included with care because other groupings of 47. Crawford, Sales Forecasting, p. 31. years w i l l produce d i f f e r e n t trend l i n e s . Despite the apparent value of trend determin-ation the disadvantages l i m i t the usefulness of trend analysis. These disadvantages are a lack of homogeneity and an i n a b i l i t y to remove the p r i c e f a c t o r . Trend c a l c u l a t i o n s do a better job of explaining past behavior than of predicting future behavior, beeausecextending a trend o f f e r s l i t t l e help unless the forecaster has an estimate of future c y c l i c a l movements. The trend seldom predicts accurately because the e f f e c t s are usually obscured during the short run, and the trend often changes 48 d i r e c t i o n over the long run. The most l o g i c a l of a l l trend curves for many series i s the elongated S curve, yet i n pr a c t i c e the curve i s not as valuable as might be expected. The S curve takes into account the nature of growth to s t a r t slowly, accelerate r a p i d l y and l e v e l o f f i n maturity. The application of t h i s curve to business growth has not been as valuable as the application i n s uch f i e l d s as biology because human sources p a r t l y control business data but do not control 48. Crawford, Sales Forecasting, p. 32. 141. b i o l o g i c a l data. A forecaster i s never sure of the p o s i t i o n of h i s sales on the growth curve u n t i l they reach the top plateau, and by that time much of the forecasting usefulness of the curve i s l o s t . The c y c l i c a l movement i s even more d i f f i c u l t to predict than the trend. T h e o r e t i c a l l y , i f the trend i s removed from a sales series and the data are placed on an annual basis, the only fl u c t u a t i o n s remaining are c y c l i c a l or i r r e g u l a r . A smoothing process then removes much of the i r r e g u l a r fluctuations leaving a nearly pure c y c l i c a l movement. In prac-t i c e t h i s process i s almost impossible unless many judgments are made along the way. Even i f a c y c l i c a l movement i s is o l a t e d successfully, the movement i s seldom i n a periodic repeating form. I f a fi x e d periodic pattern i s evident, i t cannot be r e l i e d upon. The forecaster i s then back where he started from, attempting to predict the c y c l i c a l forces that w i l l operate i n the near short-term. Because the forecaster i s o l a t e d the c y c l i c a l movement he derived a better understanding of the past, and so, presumably, 49 i s better able to estimate the future. The extent 49. Crawford, Sales Forecasting, p. 33. to which the c y c l i c a l movement i s useful i n t h i s way to the forecaster i s the extent to which the force i s useful i n sales forecasting. Although the seasonal movement provides the greatest assistance to forecasters, there are d i s -advantages to be considered. Very short-term for e -casts are usually obtained by forecasting for a year, for s i x months or for a quarter, and then by breaking these estimates down into months by a mechanical process based on expected seasonal patterns. There are several ways to compute a seasonal pattern, and to apply the seasonal percentages to an annual forecast requires nothing more than p l a i n m u l t i p l i c a -t i o n . However, d i f f i c u l t i e s are often encountered. The most important i s the a b i l i t y of the s e l l e r to change past seasonal patterns. In addition there may not be enough data available to es t a b l i s h a seasonal pattern, and even i f s u f f i c i e n t data are obtained the seasonal pattern may be changing over time. Consequently, most forecasters make mental estimates of seasonals, but use them as s t a r t i n g points for considering future sales, not as sole determinants of sales. The \"period ogram\" school of business cycle theory i s p a r t l y responsible f o r the development of t h i s questionable method c a l l e d time series analysis. The concept never di d have a firm foundation, and with the emergence of the government as a dominat-ing force i n the economy, the foundation i s s t i l l weaker. Basic changes i n i n s t i t u t i o n s , r i s k s and motivations have altered markedly the continuity' of the stringent conditions necessary for the use of such a projection technique. However, fo r many products t h i s method has a firm economic foundation, e s p e c i a l l y when confined to trend extrapolations and a seasonal pattern, and the forecasting of c y c l i c a l business changes can be l e f t to other methods. In our growing economy, where urbanization and mechan-i z a t i o n are becoming more pronounced, some good long-range forecasts have been obtained f o r some products by merely projecting the trend. VI. MULTIPLE METHOD APPROACH One of the most r e l i a b l e ways to obtain a sound forecast i s to use several techniques. I f forecasts are developed by using the jury method, the sales force composite! ;method and one or more s t a t i s t i c a l means and they a l l agree reasonably well, f a r greater confidence can be placed i n the r e s u l t than i f the forecast was reached by only one approach. In the multiple method approach each forecast acts as a check upon the others. When the differences between two or more independent forecasts are scruti n i z e d , executive judgment i s sharpened and the f i n a l forecast w i l l probably be more r e l i a b l e . When each i n d i v i d u a l forecast i s examined by the executives before being used i n the developing of the f i n a l composite forecast another method i s created \u00E2\u0080\u0094 that of using a jury of executive opinion to select the f i n a l forecast. The multiple method approach i s usually consid-ered to be more r e l i a b l e f o r developing a sales forecast than any single method. This i s also true i n the gathering of data. Since most business a c t i v i t i e s are i n t e r r e l a t e d to some extent, a study of basic forecasting data gathered from many sources w i l l provide several points of cross-50 ^ reference. While the multiple method of fore -casting usually provides best r e s u l t s , a perfect 50. A.C. Scott, \"Finding and Evaluating Basic Data f o r Sales Forecasting,\" i n Sales Forecasting -Uses, Techniques, and Trends, p. 50. and i n f a l l i b l e forecasting method i s unknown, and the development of a forecasting procedure which would r e s u l t i n a more accurate appraisal of future sales prospects i s a universal objective. VII. FORECASTING FROM SCRATCH Today many companies are having to develop sales forecasts with l i t t l e or no background mater-i a l . Many companies have never forecast before and do not possess sales records that are adquate fo r forecasting. In most i n d u s t r i a l f i e l d s there i s a serious lack of s t a t i s t i c a l data, so companies receive l i t t l e assistance from that source. The f i r s t step i n preparing forecasts for the f i r s t time i s to reconstruct the company's sales records and to e s t a b l i s h a system for keeping them up-to-date. This i s a c o s t l y process, but good sales records are so valuable that once they have been prepared t h e i r maintenance i s inconsequential i n comparison with the benefits gained. I f records are available i n a company but are not i n a properly arranged form the necessary arranging may not be unduly time consuming. On the other hand there are instances where thousands of man hours have been spent constructing sales records from invoice f i l e s . I f the forecast i s to be based on the company's po t e n t i a l instead of on the company's past perform-ance, the approximate demand of each t e r r i t o r y must be ascertained. Industry figures should provide the answer. I f such figures are unavail-able the company has to make i t s own estimates. When a company i s i n t h i s p o s i t i o n a market survey must be made i n order to determine the r e -l a t i o n between t o t a l consumption of the company's products and some factor such as regional popul-ation, the number of wage earners employed, pay-r o l l s , the assessed valuation of the plant or i n d u s t r i a l power consumption. I f v a l i d r e l a t i o n -ships can be established, they can be applied to av a i l a b l e data i n order to get an idea of t o t a l demand. Discovering these relationships and deter-mining t h e i r v a l i d i t y requires a good deal of s k i l l . Sometimes one factor w i l l n u l l i f y an otherwise perfect c o r r e l a t i o n and recognizing and c o r r e c t -51 ing f o r the factor can be a d i f f i c u l t task. 51. Thompson, Forecasting Sales, p. 33. A market survey can also be taken i n another way. For instance, i f a manufacturer of commercial equipment lacks external data a survey of several c i t i e s could provide the company with data related to customer preference i n addition to other inform-ation from which could be learned the re l a t i o n s h i p between the sales of the company and the sales of competitors, the sizes of the new and replacement markets, the character and r e l a t i v e importance of the purchasers of the equipment, and the degree to which various markets are saturated. This inform-ation, a l l i e d to an analysis of company sales, could provide a s a t i s f a c t o r y foundation f o r fore-i c a s ting. VIII. FORECASTING THE DEMAND FOR NEW PRODUCTS , Forecasting the demand f o r a new product i s quite unlike forecasting f o r an established product. A product that i s new to the company and to the economy necessitates a c a r e f u l study of the compet-i t i v e and economic c h a r a c t e r i s t i c s of the product, and forecasts that are developed exclusively for that p a r t i c u l a r product. Shaping a forecast to f i t the needs of a new product can be done i n several ways. 1. The evol-utionary approach, may be used. Here the demand fo r the new product may be a projection of the outgrowth and evolution of an e x i s t i n g old product. 2 . In the substitute approach the new product i s analyzed as a substitute f o r a product or service already i n existence. 3. The growth-curve approach u t i l i z e s the growth pattern of established products to estimate the rate of growth and the ultimate l e v e l of demand fo r the new product. 4 . In the opinion-polling approach a sample survey i s taken of the ^ p o t e n t i a l ultimate buyers of the new product and the r e s u l t s of the sampling are blown up to f u l l scale. 5 . I f the sales-experience approach i s used the new product i s offered for sale i n a sample market. 6. When the vicarious approach i s used the product i s placed on the market, and customers' reactions are reported by s p e c i a l i z e d dealers who are depended upon to be informed about customers' needs and the presence of substitute products. The d i f f e r e n t approaches o f f e r varying u s e f u l -ness. The evolutionary approach i s useful only 5 2 . Dean, Managerial Economics, p. 173. when the new product i s l i t t l e more than an improve-ment of an e x i s t i n g product. The demand w i l l probably be close to a projection of the p o t e n t i a l development of the ex i s t i n g product. The primary problem i s to i o ascertain how the demand patterns of the new version w i l l d i f f e r from those of the o l d . The substitute approach i s very h e l p f u l when relevant to the s i t u a t i o n . The old product often indicates the most that can be expected from the p o t e n t i a l market f o r the new product, but the important problem i s to estimate how r a p i d l y the new product w i l l replace the old, not what the po t e n t i a l market w i l l be. Furthermore, each new product usually has several uses, and each use presents a s u b s t i t u t a b i l i t y problem. Also, the sub s t i t u t i o n of the new product for the established one may account for only part of the p o t e n t i a l demand for the new version. The growth-curve approach cannot always be developed, and even where development i s possible, a p p l i c a b i l i t y i s l i m i t e d . This approach i s gener-a l l y used i n the l a t e r stages of demand projection. 150. Wide use has been made of opinion-polling to discover the demand fo r new products. Even f o r established products sampling c a r r i e s inherent problems of r e a l intentions and multiple a l t e r -native choice. For new products the d i f f i c u l t i e s of the approach are accentuated and the additional problem i s encountered of conveying to the p o t e n t i a l customer what the new product i s and what i t w i l l 53 do. When the experiment i s properly controlled, placing a new product on the market on a t r i a l b asis places forecasting on a sounder ba s i s . The s i g n i f i c a n t problem i s deciding the necessary allowance to make for the pec u l i a r c h a r a c t e r i s t i c s of the immature sample market. The vicar i o u s approach features the utmost s i m p l i c i t y but very l i t t l e r e l i a b i l i t y . Estimates from vica r i o u s surveys vary with the a b i l i t y of the dealer to guess what h i s customers w i l l buy and to report h i s guesses i n an unbiased fashion. 53. Dean, Managerial Economics, p. 174. IX. BREAKDOWN OF THE SALES POTENTIAL TO TERRITORIAL UNITS The company sales p o t e n t i a l can often be broken down into t e r r i t o r i a l u n i t s . This involves preparing an index or choosing one already made. Because more than one factor usually influences sales, several indexes are often combined into one. This i s done so frequently that index methods are c l a s s i f i e d broadly on the basis of the number of indexes combined. For instance, i f any i n d i c -a t i o n of the r e l a t i v e p o t e n t i a l of d i f f e r e n t parts of the market for radios i s needed, an index of wealth per ca p i t a may be used, but a more accur-ate estimate can be obtained from a composite index developed by combining indexes of wealth per capita, percentage of urban population and 54 degree of l i t e r a c y . Many index methods are i n use. The reason i s that the indexes are based on a v a r i e t y of assump-tions and they d i f f e r widely i n t h e i r complexity and i n the accuracy of t h e i r r e s u l t s . 54. D.M. Phelps, Sales Management, R.D. Irwin, Homewood, I l l i n o i s , 1953, p. 214. Some companies use a s i n g l e - f a c t o r index method upon which to base t h e i r regional forecasts. These firms assume that the extent to which t h e i r products already permeate the market i s reason-ably close to the p o t e n t i a l of that market fo r the products. For t h i s reason they are s a t i s f i e d to use t h e i r own past sales as the single factor upon which to base t h e i r estimates. This assumption denies the need fo r an objective standard of sales performance. Furthermore, t h i s assumption i s based on other assumptions \u00E2\u0080\u0094 that the company has covered the whole market thoroughly, that a l l parts of the sales organization have been equally success-f u l , and that competition has been of uniform strength i n a l l areas of the market. Even a company i n a quasi-monopolistic p o s i t i o n cannot j u s t i f y such assumptions, e s p e c i a l l y i f competition 55 from substitutes i s a c t i v e . A company i s l i k e l y to be strong i n some areas and weak i n others, and the need to determine the company's r e l a t i v e p o s i t i o n explains the necessity of acquiring objective standards of sales performance. Tv/o multiple-factor index methods are used to e s t a b l i s h r e l a t i v e market potentials, These are the 55. Phelps, Sales Management, p. 215. \" a r b i t r a r y f a c t o r s \" or \"percentage average\" method, and the multiple c o r r e l a t i o n method. The former provides less accurate estimates. The name '!arbitrary f a c t o r s \" i s given to the method because the approach used i n selecting factors i s e n t i r e l y deductive. Personal judgment plays a major r o l e and the r e s u l t s are frequently inaccurate. The method gets the name \"percentage average\" from the way i n which factors are usually combined. The multiple c o r r e l a t i o n method i s named for the s t a t i s t i c a l procedure which i s used to select 56 and combine the factors into a single index. The multiple c o r r e l a t i o n method has many advantages. Use of t h i s method reduces the number of questionable assumptions to be made. The fore -caster does not have to r e l y e n t i r e l y on deductive reasoning when selecting and weighting f a c t o r s . When t h i s method i s used e f f e c t i v e l y the r e s u l t s appear to be more accurate than the r e s u l t s obtained from any other index method. The use of t h i s method as an a i d i n determining market p o t e n t i a l i s a te c h n i c a l matter and should be undertaken only by s t a t i s t i c i a n s . 56. Phelps, Sales Management, p. 239. The multiple factor index i s a series of estimates i n percentage form of the proportion of t o t a l sales which d i f f e r e n t segments of the market would have acquired i f the independent 57 variables had been the only determining f a c t o r s . A well-constructed index based on factors that are reasonably stable and that have had a high r e l a t i o n -ship with sales i n the past may provide a more accurate i n d i c a t i o n of the d i s t r i b u t i o n of sales i n the future than are provided by actual sales i n any preceding year. An index of t h i s kind may be regarded as the market demand pattern which represents the most l i k e l y d i s t r i b u t i o n of sales over the market. Comparing the d i s t r i b u t i o n of company sales with that of both industry sales and the estimated index may provide h e l p f u l information. For such comparisons indexes by sales t e r r i t o r i e s may have to be prepared. The strong and weak t e r r i t o r i e s 58 of a company may be exposed i n t h i s way. Although data made available through the use of multiple 57. Phelps, Sales Management, p. 250. 58. Phelps, Sales Management, p. 252. factor indexes are helpful,,they are not a sub-s t i t u t e f o r , but a supplement to, good judgment. The function of the data i s to expose the p o t e n t i a l s i z e of the market i n the various sales t e r r i t o r i e s , not to t e l l management what to do. Without the information management may view the market only through the sales a c t i v i t i e s and records of the company, and through unorganized observation. With the data, management i s provided with a d d i t i o n a l f a c t u a l material useful i n planning the sales e f f o r t and i n formulating business p o l i c y . The data substitute more accurate assumptions for les s accurate ones and give executives a sounder found-ation for t h e i r planning. Decisions related to marketing should therefore produce fewer errors and greater p r o f i t s . CHAPTER V ADDITIONAL FORECASTING STEPS I. CHOICE OF METHOD There are many d i f f e r e n t forecasting methods and firms should use the techniques that apply best to .-their p a r t i c u l a r s i t u a t i o n s . Many factors must be considered when forecasting procedures are chosen. Some of these factors and the influence they exert upon the choice of the forecasting method are discussed i n the following pages. There are three c r i t e r i a that should apply to a l l methods. The f i r s t states that executives who use the forecasts must understand the methods by which the forecasts are obtained and they must have confidence i n the r e s u l t s . The executive who i s not trained i n mathematics and s t a t i s t i c s cannot understand the more complicated mathematical methods and i f he does not understand how the estimates are obtained he may d i s t r u s t the r e s u l t s . Because the executive must have confidence i n the forecasts i f h i s co-operation i s to be obtained, persons who decide which techniques are to be used should either make sure that the methods are simple enough to be understood or f i n d ways of enlight-ening those who w i l l use the forecasts so they w i l l use them with confidence. The second r u l e i s very simple but i s often overlooked. The method that i s chosen should be f a i r l y accurate, that i s , accurate so far as the siz e of the error i n i n d i v i d u a l forecasts i s concerned and accurate when considered with the number of times the r e s u l t s are unsatisfactory. A few bad forecasts quickly destroy confidence i n what should be a useful t o o l . At the same time, i f the margin of error i s allowed to become too great the usefulness of the forecast i s greatly impaired. Many companies i n s i s t that forecasts be revised frequently so they may be kept i n l i n e with changing conditions. A t h i r d requirement i s that the method chosen must more than repay the company f o r the cost of the program. Not only should the d i r e c t expense of making a forecast be taken into account, but the time devoted to the forecasting problem by executives and salesmen i s an additional expense i n the sense that these people have less time to perform 1 other duties. A good forecast should pay worth-while dividends i f i t i s to j u s t i f y the cost. The types of data available often indicate the method of forecasting that can be used. I f a good supply of data i s available on company sales and on industry sales the data can be analyzed and basic trends extended. S t a t i s t i c a l projections and c o r r e l a t i o n analyses may also be used. I f company records are inadequate but industry records are s a t i s f a c t o r y , sales can be forecast on an industry basis and estimates made of the company's share-of-marinet. I f neither company records nor industry records are adequate other methods must be used such as the jury of executive opinion, the sales force composite, or an analysis made of the general economy, as these methods require only the most elementary past records. The cost of recon-structing past records may be p r o h i b i t i v e , but some companies have found that the cost pays for i t s e l f over a period of years because forecasting can be done with such increased accuracy. 1. C.G. Thompson, Forecasting Sales, National I n d u s t r i a l Conference Board, Studies i n Business Policy, No. 25, 1947, p. 43. The estimating of sales f o r an established product i s d i f f e r e n t from the forecasting of demand for a new product. Even i f past records are poor a company can draw upon personal sales experience to predict the future of a well established product. There i s no such basis f o r estimating the future sales of a new product unless the new product i s very s i m i l a r to the ol d . I f the new product bears no resemblance to any previously made by the company a c a r e f u l study should be made of si m i l a r products on the competitive market. These studies could include market research or campaigns to t e s t sa l e s . I f the new product i s unlike anything previously placed on the market a r e l i a b l e sales forecast may be an i m p o s s i b i l i t y . When a new product i s introduced to the market early sales are watched very c l o s e l y and forecasts are revised with every s i g n i f i c a n t change of consumer or competitor reaction. The sales i n some industries are not seriously affected by changes i n economic conditions and i n such cases f a i r l y r e l i a b l e forecasts can be made merely by making an analysis of industry growth curves and of population changes. Where demand for a product fluctuates, changes i n buying power, national income and pri c e s are among the factors that must be considered. The industry that i s subject to extreme fluctuations i n the demand f o r i t s product must make intensive and elaborate studies of economic conditions and of factors a f f e c t i n g sales. End-use and c o r r e l a t i o n studies may be h e l p f u l and the jury of executive opinion and sales force composite methods may be t r i e d i f the executives and sales force have s u f f i c i e n t knowledge of the company's operations. Forecasting takes a d i f f e r e n t course when a company has some control over factors that a f f e c t sales. During the Second World War and f o r some time after, many companies found that t h e i r sales were l i m i t e d only by the supply of raw materials, the a v a i l a b i l i t y of labour and the capacity of the plant. These companies knew how much of each product they would produce and where the product would be sold. A company whose sales dominate the industry i s sometimes i n a si m i l a r p o s i t i o n . Where t h i s occurs the sales forecast i s developed on the basis of how much the plant can produce and the p r o f i t to be made by each product i n the l i n e . The p a r t i c u l a r technique chosen for the f o r e -cast depends upon the purpose for which the estim-ates w i l l be used. If the forecast i s to be used for f i n a n c i a l planning and plant expansion, basic trends are more important than c y c l i c a l v a r i a t i o n s so a long-range forecast with l i t t l e d e t a i l i s s u f f i c i e n t . For day-to-day planning the forecast must shov; at least the s i g n i f i c a n t short-term fluctuations as well as trends i n the sales of various products and allow f o r temporary factors that need not be considered i n the long-term f o r e -cast. I f the temporary factors cannot be estimated, as i n the case of weather, the f a c t i s noted i n the f i n a l forecast. The e f f e c t that these immeas-urable factors can have on the forecast, as well 2 as the extent of the e f f e c t , should be d e t a i l e d . When the market fo r a c e r t a i n product i s l i m i t e d a study of major customers or consuming industries i s often a good basis f o r a forecast. This i s p a r t i c u l a r l y true i f the consumers do not react i n the same way to the t o t a l consuming market as they do to the general economy. In other words, l o c a l Conditions may have a greater e f f e c t upon 2. Thompson, Forecasting Sales, p. 44. sales than do trends i n general business conditions Here the best technique f o r developing a forecast i s one with a t e r r i t o r i a l b a s i s . Companies with a complex l i n e of products have found they must group together the products that react i n the same way to factors that a f f e c t sales. Some products s e l l poorly when purchasing power i s reduced or spendable income decreases, while some products f i n d a ready market i n depression years. I f c e r t a i n products share the same reaction to factors that have a bearing on t h e i r sales they are grouped together for the forecast. Products that cannot be placed i n a group are forecast separ a t e l y . Many forecasting methods have been developed over the years by manufacturers, wholesalers, and r e t a i l e r s . Not only are some procedures more d i f f i c u l t to use than others, but the same proced-ure may be more d i f f i c u l t to use i n one type of business than i n another. The r e t a i l e r appears to have the easiest forecasting task because he works to a f a i r l y stable seasonal pattern and he s e l l s to a r e l a t i v e l y small l o c a l market. The manufacturer, on the other hand, finds forecasting d i f f i c u l t . He s e l l s to a large and f a r flung market and has d i f f i c u l t y forecasting shipments over long trade channels. The length of a manu-facturer's product l i n e also adds to the complexity 3 of the forecasting. Some firms experience less d i f f i c u l t y i n forecasting than others because some industries are b a s i c a l l y more stable than others. The amount and kind of data available to the various firms influence the ease with which t h e i r f o r e -casting problems can be solved. Companies vary i n size, and other things being equal, the larger company has the advantage of being able to afford s p e c i a l i s t s to study various phases of the fo r e -casting problem. The small firm with no trained personnel f o r the forecasting task often has to use estimates that are based on l i t t l e more than i n t u i t i o n . Although there are many sales forecasting techniques to choose from not a l l techniques can be u t i l i z e d by every company and no one technique i s the best f o r a l l companies. Management must 3. H. Spender and L. Siegelman, Managerial Economics, Homewood, I l l i n o i s , Irwin, 1959, p. 47. decide which method w i l l provide the most accurate and dependable forecast f o r the company concerned. There are several rules that management might consid-er when a forecasting method i s being chosen. 1. Comprehensibility. I f management i s going to use the forecast as a basis f o r planning, the technique used i n developing the forecast must be understood by those who are to use the f o r e -cast. Executives w i l l have no confidence i n a procedure they do not understand and the f o r e -cast w i l l be of l i t t l e value. 2. Accuracy. A forecast cannot be considered accurate i f the turns are not predicted as well as the trends. A forecast that indicates a continuation of a trend implies no change i n plans. As a turn i s approached new decisions must be made and new plans formulated. Since information about a predicted turn i s of v i t a l importance a forecast cannot be considered accurate unless the turning points can be predicted accurately. 3. Timeliness. The forecasting method chosen should make use of the most up-to-date information obtainable and should be capable of u t i l i z i n g new data when conditions change. 4 . U s a b i l i t y . The method that i s chosen should provide forecasts for the same units and groups as the company uses or be e a s i l y converted to those units and groups. For the production department the forecast should be i n physical units, f o r the finance department the pr e d i c t i o n should be i n current or deflated d o l l a r s or whatever form the department requires, and so on f o r other departments. 5. Economy. F i n a l l y the method should be one that the firm can afford and the s t a f f can handle. The value of a forecast i s d i f f i c u l t to c alculate, so management must decide i f the benefits derived warrant the various costs 4 involved. If these c r i t e r i a are taken as a guide and a technique i s chosen that appears most suitable, the forecasting program w i l l not be successful unless a l l departments co-operate i n the development, and management places confidence i n the r e l i a b i l i t y of the forecast. 4 . Spencer and Siegelman, Managerial Econ- omics, p. 4 6 . I I . THE CONCEPT OF A FIRST APPROXIMATION FORECAST Sales forecasting i s described i n business l i t e r a t u r e as a function where a l l pertinent inform-ation and company p o l i c i e s are u t i l i s e d i n order to a r r i v e at a forecast of sales that w i l l be use-f u l throughout the firm. Emphasis has been placed on finding a technique that w i l l provide the most trustworthy forecast. However/ some firms discuss an intermediate step, one that a l l firms i m p l i c i t l y take but few emphasize, and f i n d that when t h i s step i s emphasized the forecast i s more valuable to management. This intermediate step consists of accepting the forecaster's best p r e d i c t i o n as a \" f i r s t approximation.\" The procedure i s as follows. The forecaster studies a l l relevant data and prepares a forecast that he believes i s most accur-ate for the conditions that he has been t o l d or that he assumes w i l l p r e v a i l . Here the new idea makes an appearance. Now the various operational executives discuss the estimate, not i n terms of what they consider l i k e l y , but to discover whether the forecast conforms to the general intentions of the firm and w i l l allow the achievement of desired r e s u l t s . The estimates may provide f o r the s e l l i n g of more of a c e r t a i n item than the firm can buy raw material to make. In another instance the p r o f i t may be too low. The problems are then d i s -cussed i n an e f f o r t to f i n d a more s a t i s f a c t o r y p r e d i c t i o n . This procedure i s one that businessmen would not be expected to overlook, yet many of them do. Some firms merely i n s t r u c t t h e i r forecasters to prepare estimates, to submit these estimates to the sales managers f o r approval and then to send the forecasts to those i n the firm who require them. When the l a t t e r procedure i s followed the executives who develop the plans that are used by the forecaster estimate the res u l t s they expect from each part of the plans. When they make decisions on the plans before the forecaster uses them they do not take into consideration the f a c t that they and the forecaster may not be of the same opinion. The s i t u a t i o n i s the same when the fore-casts are examined by a board of review \u00E2\u0080\u0094 the forecasts are judged on t h e i r accuracy, not on t h e i r p r a c t i c a l i t y . To overcome t h i s weakness some firms appraise not only the forecasts tout the forecasted s i t u a t i o n . The planning stage i s not considered complete u n t i l the f i r s t approximation forecasts have been pre-pared and appraised. The review procedure i n t h i s case i s d i f f e r e n t from the one undertaken when only the forecast i s examined, because a l l the people who had a share i n the o r i g i n a l planning are re c a l l e d to consider the f e a s i b i l i t y of the estim-ates. In t h i s way the forecast becomes an i n t e g r a l 5 part of the planning procedure of the company. The very small firm has probably always followed t h i s course of roundtable discussions by the exec-utive committee. When a company grows large and hi r e s or t r a i n s a forecasting s p e c i a l i s t there i s a tendency to separate planning and forecasting, the \" f i r s t approximation\" concept attempts to reunite these two functions. I I I . POST-PERFORMANCE ACTIVITIES When the estimates have been received and accepted they must be d i s t r i b u t e d throughout the firm and t h i s process can become quite complicated. For example a fir m may make complete forecasts 5. CM. Crawford, Sales Forecasting: Methods of Selected Firms, University of I l l i n o i s , 1955, p. 38. semi-annually and d i s t r i b u t e them i n t h i s manner: 1. the sales department may receive a forecast of company sales by broad commodity groups, by months and by d i s t r i c t s ; 2. the production depart-ment w i l l require a forecast by commodity groups, i n annual units; 3 . the inventory control section may obtain a forecast of company sales by commodity groups, i n d o l l a r s ; and 4 . the budgeting depart-ment may receive a forecast by commodity groups, by months, i n d o l l a r s . Other firms may have simpler or more complicated methods of d i s t r i b -ution, but regardless of the method, d i s t r i b u t i o n 6 i s an important problem. The l a s t performance step, and one that i s neglected too often, i s the recording of the d e t a i l s of the development of the forecast program and the thinking upon which the decisions were based. Fore-casters can study past estimates and learn the reasons f o r t h e i r success or f a i l u r e . Management should be able to trace errors i n the f i n a l fore-cast back to t h e i r source i n order to learn t h e i r cause. I f a forecaster i s new to a firm past forecasts can be of inestimable value to him, so 6 . Crawford, Sales Forecasting, p. 19. management should keep with the company a l l records of past estimates i f the forecaster resigns to go 7 with another organization. IV. PRINCIPLES OF CONTROL APPLIED TO SALES FORECASTING In factory management two types of control can be exercised \u00E2\u0080\u0094 production control and q u a l i t y c o n t r o l . Production i s controlled i n an e f f o r t to gain smooth performance i n the plant from raw material to f i n i s h e d product. Quality i s c o n t r o l l e d to ensure that the f i n a l product reaches the de-s i r e d standard. We can apply t h i s d i s t i n c t i o n to forecasting and f i n d that control which pertains to the development of the sales forecasting program i s a type of production control, and control which pertains to the methods used i n a c t u a l l y preparing the forecast i s a type of q u a l i t a t i v e c o n t r o l . The l a t t e r type w i l l be discussed here. The q u a l i t y of the f i n a l forecast depends upon every step taken from the planning stage to the f i n a l estimates. However, a forecaster can do 7. Crawford, Sales Forecasting, p. 19. three things to augment the q u a l i t y of h i s est imates. F i r s t , he can state the l i m i t s of h i s forecas t . Most people who use a forecast need an exact s ta t e -ment of the best expectations, but some users f i n d they can make be t ter use of a forecasted \"range\". Even those who need an exact statement of estimates often f i n d that a forecasted range helps them i n t h e i r own p lanning . A forecasted range helps an appraiser to know just what the forecast i s i n -tended to por tray , so c o n t r o l i n l a t e r stages i s eas ier to mainta in . A second step the forecaster can take to c o n t r o l the performance of the forecast i s to state the assumptions on which the estimates were based. These assumptions can cover general economic c o n d i -t ions such as na t iona l income, expenditures on producers' durable goods and patterns of consumers' spending, and they may a l so cover the expected r e s u l t s of important company p l a n s . Less s i g n i f -i cant assumptions should a lso be stated because they cannot a l l be remembered and yet reference to 8 them l a t e r may prove very u s e f u l . 8. Crawford, Sales Forecast ing , p . 59. F i n a l l y , a forecaster can t r y to get a l a s t -minute appraisal of h i s forecast before i t i s used. This can be done i n several ways but the most usual ways are to consult with top executives or to use an outside advisory service. A f t e r the forecast i s made some changes usually occur i n conditions which influence sales and even i f none occur some degree of error w i l l usually be found i n the estimates. One factor i n the f o r e -casters' favor i s that the forecast, unlike the product of the factory, can be changed as soon as the estimates are completed. Because of t h i s , forecasters have always used what can be considered a control factor when they compare the predicted r e s u l t s with actual r e s u l t s as soon as the figures come into the company. One way i n which t h i s type of control can be obtained i s by p l o t t i n g the forecast on chart paper and showing the control l i m i t s beyond which the forecast w i l l be too inaccurate to be of value. I f the forecast becomes unreliable, the forecaster must revise h i s estimates. Fore-casts plotted i n t h i s manner can be revised at the beginning of a period as well as at the end, so changes can be made before the forecast loses a l l value. The in t e r e s t i n g aspect of t h i s device i s the f a c t that a control formerly based s o l e l y on judgment can now be accomplished s t a t i s t i c a l l y . This approach can probably be adapted to f i t the conditions i n i n d i v i d u a l firms. A forecast should be revised when the value i s l o s t because of inaccuracy. Therefore, the sooner the company discovers the need f o r change the sooner can the adjustments be made. The most common way to learn when a forecast needs adjusting i s to review the estimates. This brings up the problem of how often forecasts should be reviewed. V. REVIEW AND REVISION Some forecasters believe that reviewing should be done every month. This may seem l i k e an i n -ordinate amount of e f f o r t , but such need not be the case. Even i n a medium sized company where the forecasting i s the r e s p o n s i b i l i t y of one man a monthly review i s f e a s i b l e . Once the forecast i s f i n a l i z e d the forecaster does not i s o l a t e him-s e l f from economic news or from news of the f i e l d sales force f o r months at a time. Furthermore, the review should not consume a disproportionate amount of time because with experience the f o r e -caster becomes s k i l l e d and he becomes accustomed to gathering pertinent information during the month. In addition to a monthly review more and more companies are forecasting twelve months ahead every quarter. This means that the previous estimate fo r the next nine months i s revised, and a f i r s t estimate made for the quarter that i s a year ahead. This system avoids the implication that years should be planned as units, and emphasizes the f a c t that the forecaster i s continually searching 9 for new developments and new information. A short-vor medium-range forecast covering a period ahead of two to eighteen months should always be i n existence and should be reviewed at lea s t every quarter. When the short-range forecast i s reviewed several factors are weighed such as current information on sales, news about the industry, news about the general economy and new plans the company i s making. The r e s u l t of the review i s 9. T.G. MacGowan, \"When and How Forecasts Should be Reviewed and Revised\", i n Sales Forecast-ing - Uses, Techniques and Trends,\" American Manage-ment Association, Special Report no. 16, 1956, p. 100. a general estimate of error i n the forecast or a general estimate of a new forecast. The long-range forecast which predicts f o r a period ahead of f i v e to ten or more years should be revised from every year to every f i v e years. The emergence of a problem p a r t i c u l a r to long-range planning indicates the need fo r added r e v i s i o n . No r u l e can be used f o r deciding when the review of an industry's forecast indicates that a new forecast must be made. Some industries are quite stable, but other industries fluctuate v i o l e n t l y , so general statements cannot be applied i n d i s c r i m i n a t e l y . However, for most industries reforecasting i s advisable when the error i s f i v e percent or more. In other words, i f a new f o r e -cast would vary as much as f i v e percent from the e x i s t i n g one, a new forecast should be made. Either one of two situations may a r i s e that would indicate the need fo r a r e v i s i o n of the f o r e -cast. A new forecast should be made i f the present forecast i s apparently poor, or i f the forecast has been s a t i s f a c t o r y u n t i l the present, but factors s i g n i f i c a n t to sales have obviously changed. I f the forecast i s poor, every e f f o r t should be made to learn the cause. Perhaps the pre d i c t i o n of general business was s a t i s f a c t o r y butLthe estimates of sales were wrong, or perhaps the general economy varied from that predicted because of factors the forecaster did not or could not foresee. Forecast sales and r e a l i z e d sales should be compared i n order to learn when the forecast has gone beyond the maximum expected error, and management should receive notice as soon as possible when t h i s occurs. For t h i s reason the forecast should be broken down by months or by four-week periods i n order to make continual comparisons of sales. This i s true even i n industries where a forecast for a period shorter than a quarter i s impractical. In addition to t h i s information on predicted sales and actual sales the forecaster needs both general and detailed information concerning business conditions. He needs to know the general economic climate, the a c t i v i t i e s of h i s company's competitors, as well as the l a t e s t prices and the l a t e s t knowledge about future p r i c e s . He should know also i f h i s company has made new plans, developed a new product or entered a new market. In short he needs the information that he needed to make the forecast i n the f i r s t place. In order to acquire s u f f i c i e n t data f o r a review of the estimates, the forecaster should have a regular program of conferences with men i n the sales organization. The information the sales s t a f f brings from the f i e l d may provide the f i r s t i n d i c a t i o n that a new development i s taking place. Interviews should be held with the production planning, t r a f f i c , purchasing and any other depart-ments where recent information can be gathered. The forecaster i s interested not only i n the influences that affected sales i n the recent past, but i n developments that w i l l a f i e c c sales i n the future. When recent sales r e s u l t s are to be reviewed they f i r s t should be compared with the estimated f i g u r e s for a given period. The f a c t that some differences w i l l e x i s t i s accepted. The problem i s whether these differences are expected and acceptable. The forecaster should know the probable amount of error and he should know the greatest error that \"chance v a r i a t i o n s \" may cause. These va r i a t i o n s r e s u l t from causes that are unknown. When recent sales records have been reviewed several conclusions are possible. The forecaster may be r i g h t f o r the wrong reasons. Every i n d i v -i d u a l item may be wrong but h i s t o t a l may be nearly correct because errors cancelled each other. He may be wrong for a short-range reason, that i s , a temporary s i t u a t i o n that w i l l soon be changed. For instance, January sales may be higher than expected and February sales correspondingly lower, but the sales f o r the quarter may be as predicted. Another s i t u a t i o n that causes wrong forecasting f o r a temporary reason i s the occurence of \"chance\" fluctuations, which are variations r e s u l t i n g from the sum of minor p o s i t i v e and negative forces which are not considered separately. In theory, these v a r i a t i o n s cancel each other over a.period of time, but i n the short ;r.un either the p o s i t i v e or the negative forces may dominate the s i t u a t i o n and produce a divergence from the forecasted 10 f i g u r e s . The forecast may be wrong fo r a more permanent kind of reason, one that means the forecast w i l l 10. McGowan, Sales Forecasting, p. 103 remain f a u l t y . This may be the r e s u l t of incorrect i n t e r p r e t a t i o n of correct data or be an unfore-seen development i n the general economy. Sometimes such errors appear when the forecaster was unaware of some happening eith e r i n h i s company or i n the industry when he made the forecast. Serious defects of t h i s type usually mean that the forecast must be redone. I t may be possible to adapt the present forecast i f only one factor i s d i f f e r e n t and i f that factor i s so simple i n e f f e c t that a single a l t e r a t i o n w i l l correct the forecast. Such a s i t u a t i o n seldom occurs. The errors usually require that a new forecast be made. An opportunity to study factors a f f e c t i n g sales i s presented whenever a forecast i s reviewed. The forecaster can discover for management what i s happening to the company and the industry. A ' company that has forecast reviews may learn the increasing s i g n i f i c a n c e of a marketing factor long before companies that do not have forecast reviews. I f the company learns of a weakness i n a single product, i n a product group or i n a channel of d i s t r i b u t i o n , the weakness can be r e c t i f i e d before serious consequences r e s u l t . 180. A factor that must be assessed When a forecast i s reviewed i s new information that has not yet affected actual sales. When the forecaster holds conferences with the sales organization, i n t e r -views other departments about product, p r i c e and supply, and d i l i g e n t l y searches for information concerning the industry and the general economy he i s sure to discover some information that was not i n h i s possession when he made the forecast. The new information may be concerned with a v a r i e t y of subjects. The general business outlook may have changed i n an unexpected manner since the forecast was made, the economic s i t u a t i o n or the needs of customers may have changed. The outlook for industries that buy from the industry being forecast needs spe c i a l study. The future trend of the customers* business may be indicated by reports covering the value of construction contracts, by the number of requests f o r mortgage appraisals or by s i m i l a r reports. The forecaster should not f a i l to become f a m i l i a r with estimates made within the industry that buys h i s company's products, because the forecast within the industry w i l l be used as 11 a basis f o r that company's planning and buying. The correct way to proceed with a new forecast i s to s t a r t with the pieces and put them together. Recent sales i n each segment of the industry must be studied i n d e t a i l product by product, channel by channel, and, wherever possible, market by market. These segments are then formed into a t o t a l f o r e -cast . VI. CONTROL AFTER THE FORECASTED PERIOD When a forecast i s discarded f o r a new one, many forecasters have no more i n t e r e s t i n the old forecast. Some forecasters would l i k e to examine the past forecast but do not know what to do. Whether a forecaster wishes to examine the past forecast or not, he cannot unless hekkept de t a i l e d records during the preparation of the forecast. If adequate records have been kept the appraisal can be conducted i n the following manner. F i r s t , the forecaster must obtain from manage-ment a d e f i n i t e opinion concerning the margin of error that i s acceptable i n the forecast. Then department heads should be consulted regarding II. MacGowan, Sales Forecasting,, p. 105. t h e i r opinions on the margin of error acceptable i n connection with operations under their, d i r e c t i o n . Second, the margin of error should be considered i n r e l a t i o n to the cost of the forecasting program because the greater the cost the greater the degree of accuracy that should be expected. Statements are sometimes made i n the business press concerning general l i m i t s of error i n fo r e -casts. Forecasters do not f i n d these statements u s e f u l . Generalizations cannot be made about fore-cast errors, f o r each error depends upon the envir-onment of the firm or industry i n which the error occurred. In every case the acceptable error depends upon such factors as the time period 12 covered and the product mix. The t h i r d step i n the appraisal process i s to apply these opinions to the errors i n the past forecast i n an e f f o r t to learn the cause of each er r o r . These e f f o r t s are not always successful because many variables that cannot be measured are usually present during a forecasted period. However, many of the errors can be explained. 12. Crawford, Sales Forecasting, p. 58. I f the forecaster obtains assistance from others i n the firm and records a l l important d e t a i l s he i s i n a p o s i t i o n to improve h i s subsequent forecasts. A forecaster may take one of two courses of action i f h i s forecasts need to be improved. He may change h i s present forecast or change h i s method 13 of forecasting. I f he changes h i s forecast he may either revise or reforecast, but the l a t t e r procedure i s more common. Reforecasting usually proves s a t i s f a c t o r y provided the forecaster recorded every d e t a i l and opinion that influenced the forming of the estimates, and provided he was thorough i n h i s e f f o r t s to trace the causes of h i s e r r o r s . Almost any forecasting program can be improved, the problem i s to decide how much time and money can be j u s t i f i a b l y spent on the process. I f the decision i s made to change the forecasting method t h i s should be done before the next forecasts~are prepared. In many firms the methods used are temporary arrangements, anyway, except where a forecaster has been employed by a firm f o r several years. Because the records of most forecasters 13. Crawford, Sales Forecasting, p. 60. are incomplete so far as assumptions, methods, and so on are concerned, and because forecasters generally tend to avoid s t a t i s t i c a l techniques, advances in the area of forecasting control are dependent upon a few firms. CHAPTER VI MAKING THE SALES FORECAST This chapter w i l l show the development of sales forecasts of softwood plywood f o r Crown Zellerbach Company f o r the years 1964 and 1968. F i r s t we s h a l l discuss t y p i c a l uses of plywood and mention materials that are competitors of plywood. This w i l l be follow-ed by a development of various predicting equations and the equation that appears most desirable w i l l be used for forecasting sales i n the plywood industry. The next step w i l l be to forecast plywood sales f o r 1964 and 1968. Then we s h a l l predict the share of the t o t a l market that Crown Zellerbach can expect to capture. I. PRODUCT HISTORY The f i r s t commercial production of plywood i n Canada started at Fraser M i l l s , B r i t i s h Columbia, i n 1910. In 1961 there were seventy-five plants through-out Canada classed i n the Veneer and Plywood industry 1 group as follows: 1. Dominion Bureau of S t a t i s t i c s , Veneer and Plywood Industry, Queen's Printer, Annual, Cat.No. 35-206, July, 1963, p. 3. Province Number of Plants Gross Value of Shipment! Ontario 26 $ 22,444,000 Quebec 23 24,484,000 B r i t i s h Columbia 19 91,494,000 Other Provinces 7 5,377,000 To t a l 75 $143,719,000 There are two basic types of plywood-softwood and hardwood. Most of the softwood plywood that i s produced i s made from Douglas f i r but a small amount of plywood i s made from hemlock and white pine. Nearly a l l the softwood plywood produced i n Canada i s manufactured i n B r i t i s h Columbia. Hardwood plywood i s produced i n the eastern provinces and i s made from bir c h , maple, basswood and elm. I I . TYPICAL USES OF PLYWOOD By a sizeable margin the construction industry i s plywood's biggest customer. Some of the uses to which plywood i s put i n construction are presented below. 1. Form Work - Strong, waterproof glue f i r plywood panels are required f o r form work. Because the panels are large, uniform i n size, and of r e l a t i v e l y l i g h t weight, they can be handled e a s i l y and erected quickly on a va r i e t y of jobs from the construction of conventional house foundations to complex bridges, dams and highway overpasses. A f i r plywood panel can be used f o r as many as f i f t y concrete pours and s t i l l be used f o r a long time on the construction s i t e f o r ramps, 2 runways, fences or crew huts. 2. Sheathing - A large volume of f i r plywood i s sold as sheathing for roofs, walls and f l o o r s . Sheath-ing must have outstanding bracing a b i l i t y and possess great resistance to racking, and i t must have high l a t e r a l nail-bearing strength and be e a s i l y handled f o r f a s t erection. Plywood i s r i g i d , possesses nail-bearing strength due to cross-laminated panel construction and i s e a s i l y handled because of lightness and size, so makes 3 i d e a l sheathing. 3. Prefabrication - The factory production of large-sized b u i l d i n g components i s c a l l e d p r e f a b r i c -a t i o n . F i r plywood has properties that lend themselves p e r f e c t l y to pr e f a b r i c a t i o n . In f a c t p r e f a b r i c a t i o n has made great s t r i d e s ever since f i r plywood was developed. In North America, prefabricated walls, f l o o r s , c e i l i n g and roof panels f o r houses and commercial buildings are 2. Plywood Manufacturers' Association of B r i t i s h Columbia, Canadian Douglas F i r Plywood, Unpublished Report, Vancouver n.d., p. 7. 3. PMABC, Canadian Douglas F i r Plywood, p. 9. 188. nearly always b u i l t of s o l i d wood framing members 4 covered with plywood skins. Structural components made of plywood i n arch and many other forms are enabling contemp-orary architects to use new forms of expression and at the same time are providing b u i l d i n g developers with opportunities to economize i n material and labor costs. 4. Agriculture - On the farm plywood can be used i n more ways than any other b u i l d i n g material. F i r plywood can be used for inside and outside jobs i n the construction of barns, feeders, s i l o s and storage u n i t s . There i s wide scope f o r a c l e a r -span structure that can be erected by farm hands and such a structure can be b u i l t of lumber frames joined by plywood gussets to form a three-5 hinged arch. 5. I n d u s t r i a l - The combination of great strength and l i g h t weight makes f i r plywood an excellent choice for i n d u s t r i a l purposes. High resistance to racking combined with low weight makes f i r plywood i d e a l for a l l types of i n d u s t r i a l packaging. 4. PMABC, Canadian Douglas F i r Plywood, p. 9 . 5. PMABC, Canadian Douglas F i r Plywood, p. 10. Plywood crates are s c i e n t i f i c a l l y designed f o r shipping goods of a l l kinds from f r a g i l e p r e c i s i o n instruments to j e t engines. The use of plywood p a l l e t boxes makes the handling and storing of many kinds of foods much easier. Plywood i s sold f o r making temporary screens or p a r t i t i o n s , and f a r o u t s e l l s any other material for making shelves. Because of good s t r u c t u r a l properties at a reasonable p r i c e plywood i s used i n thous-ands of ways by North American i n d u s t r i e s . I I I . COMPETITIVE PRODUCTS Plywood has many competitors. For exterior sheathing plywood competes with shiplap, masonry and stucco; f o r i n t e r i o r use plywood i s superceded by p l a s t e r except i n basement and a t t i c ; f o r sub-floor-ing and underlayment f i r plywood and shiplap are equally popular. Some newer products such as f i b e r -glass, aluminum, s t e e l and p l a s t i c s are taking a portion of the market. The r i g i d i n s u l a t i n g board industry, which competes with the plywood industry, covers asphalted sheathing board, b u i l d i n g board (in natural or coated 6. PMABC, Canadian Douglas F i r Plywood, p. 11 panels), roof i n s u l a t i o n board, decorative board ( t i l e or plank form, including acoustical t i l e ) , other boards and asbestos. Plywood\u00E2\u0080\u00A2s p r i n c i p a l competitor i s probably asphalted sheathing board which i s believed to be used mainly f o r house wall 7 sheathing. (See Table I) Semi-hardboard competes with plywood f o r use under t i l e f l o o r s , with sanded grades of plywood for wall panelling, and sometimes with plywood for grain 8 b i n l i n i n g s . :(See Table II) 7. McConnell, Eastman and Company, Marketing and Promotion Study, Unpublished Report, Vancouver, n.d.,p 8. McConnell, Eastman and Company, Marketing and Promotion Study, p. 5 . TABLE I Domestic Shipments of R i g i d Insulat ing Boards (Square feet 1/2\" basis) B u i l d i n g Asphalted Lath for Plaster Roof Insu la t ion Other R i g i d Boards Sheathing Board Boards Boards Boards 1952 106,900,768 36,421,085 16,684,679 47,684,705 27,755,188 1953 108,997,145 52,895,230 19,243,862 65,854,751 30,310,923 1954 100,273,269 55,928,778 15,001,148 76,555,220 33,084,453 1955 97,552,187 70,178,551 12,211,787 86,952,548 34,866,291 1956 111,697,922 83,560,324 7,112,871 100,013, 364 43,420,464 1957 90,306,788 68,143,153 5,013,069 98,354,839 51,739,446 1958 84,487,060 91, 712,889 4,405,754 108,601,065 76,881,333 1959 74,592,115 89,414,382 152,275,921 78,371,472 1960 61,290,396 78,909,381 156,069,631 70,362,103 1961 57,188,897 83,863,624 . 149,238,124 89,308,596 1962 54,519,168 91,081,701 181,868,995 97,563,784 1963 59,377,398 100,907,522 185,032,631 101,809,283 Source: Dominion Bureau of S t a t i s t i c s , Rig id Insulat ing Board, Queen's P r i n t e r , Monthly, C a t . No. 36-002, 1952-1963. TABLE II Canadian Semi-Hardboard Production and Shipments Production Shipments Domestic Export (Sq. f t . , 1/8\" basis) 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 17,012,468 23,072,928 29,234,912 79,330,419 105,531,480 104,531,385 105,715,675 123,324,122 127,067,173 144,809,587 174,475,530 159,675,136 15,983,132 20,576,741 21,725,269 45,825,893 64,806,041 67,137,236 92,214,502 102,442,287 117,328,657 137,827,246 155,531,795 147,383,866 373,264 8,481,024 35,452,572 40,233,440 39,969,152 15,756,142 23,135,816 11,580,339 15,615,530 19,237,514 24,411,948 Sources Dominion Bureau of S t a t i s t i c s , Hard Board, Queen's Printer, Monthly, Cat. No. 36-001, 1952-1963. IV. HISTORY OF CROWN ZELLERBACH BUILDING MATERIALS, LIMITED Crown Zellerbach Building Materials Limited, presently owned by Crown Zellerbach (Canada) Limited, originated i n the l a t t e r part of the nineteenth century on the banks of the Fraser River and was c a l l e d the McLaren M i l l . From 1905 to 1908 the McLaren m i l l was r e b u i l t and enlarged under the name of Fraser River Sawmills. In 1909 there was a shortage of s k i l l e d sawmill workers so the company brought out from Quebec one hundred ten workmen and t h e i r f a m i l i e s . These people s e t t l e d i n an area close to the millsifce and started the French-speaking community of M a i l l a r d v i l l e . In 1910 the company was reorganized under the name of Canadian Western Lumber Company, 9 Limited. During the next three years the sawmill was modernized, a door factory was added, more homes were b u i l t i n the townsite known as Fraser M i l l s . Approximately one hundred f i f t y r e t a i l outlets were 9. Crown Zellerbach (Canada) Limited, History of Canadian Western Lumber Company, Limited, Unpublished, Report, September, 1957. purchased to include the p r a i r i e provinces i n the company's market. In 1913 the plywood plant was completed and was the f i r s t of i t s kind i n Canada. Canadian Western became an a f f i l i a t e of Crown Zellerbach, Canada, Limited i n 1953 and the name was changed to Crown Zellerbach Building Materials, Limited. Today t h i s plant i s a modern up-to-date plant barking approximately 1,400 logs per day, cutti n g 550,000 feet of lumber per day, turning out 420,000 square feet of plywood per day (3/8\" basis) and cutting 400 squares of shingles per day. The company's head o f f i c e i s i n Vancouver, and other f i e l d o f f i c e s are i n Calgary, Winnipeg, Waterloo, Toronto and Montreal. Sales represent-atives cover the entire country. In 1961 the Fraser M i l l s plywood plant of Crown Zellerbach, Canada, Limited embarked on a $1.5 m i l l i o n expansion program. In 1962 the plywood plant set an a l l - t i m e production record of 103 m i l l i o n square f e e t . A $2 m i l l i o n modernization of the Fraser M i l l s sawmill d i v i s i o n was started during 1962, c r e d i t i n g the company with two major projects i n 10 two years. V. PREPARING TO FORECAST To make forecasts of plywood sales the technique to be used must be selected. To prepare 1964 and 1968 forecasts of plywood sales f o r Crown Zellerbach Company procedures that culminate i n regression equations were chosen. At present, forecasts of plywood sales i n t h i s company are prepared by sales forecasters who use personal judgment based on an analysis of a l l available pertinent information, as well as on the opinions of f i e l d representatives. Personal judgment plays an important r o l e i n the method presently used by the company, but i n t h i s study personal judgment i s placed i n a minor r o l e with the major r e s p o n s i b i l i t y f o r the forecast being placed i n s t a t i s t i c a l equations. The object Of the study i s to f i n d one or more estimating equations which show the r e l a t i o n s h i p between the dependent variable, domestic softwood plywood sales, and one or more independent v a r i a b l e s . The 10. \"$1.5 m i l l i o n extension despite tax\", The Province, Vancouver, June 29, 1963, p. 15. independent variables that have been chosen were selected because on the basis of knowledge and lo g i c they appear to be the major factors c o n t r o l -l i n g the volume of plywood sales. The r e s u l t i n g predicting equations are developed on the basis of s t a t i s t i c a l analysis. The independent variables used i n preparing the forecasting equations are gross national product, personal expenditure on consumer goods and services, r e s i d e n t i a l construction, i n d u s t r i a l construction, commercial construction and i n s t i t -u t i onal construction. These factors were chosen because they are believed to play an important part i n determining the sales of plywood. The s p e c i f i c reasons for choosing each of these factors are discussed i n the following paragraphs. Gross National Product About 27 per cent of t o t a l plywood sales are made to homeowners who use the product for 11 renovations, repairs and do-it-yourself projects. 11. The percentage of t o t a l plywood sales purchased by various groups i n the economy was provided i n a personal interview with Mr. D. Owen of Crown Zellerbach Company. The amount of money spent on plywood by homeowners i s believed to be related to t h e i r incomes. Indiv-id u a l incomes are incorporated i n the national income and the l a t t e r bears a close r e l a t i o n s h i p to the Gross National Product. Therefore there appears to be a r e l a t i o n s h i p between plywood sales to homeowners and the Gross National Product. Farmers purchase approximately 11 per cent of the plywood sold i n Canada. The amount of money that farmers spend on building and repairing i s determined to a considerable extent by t h e i r incomes. Again there i s a re l a t i o n s h i p between income and Gross National Product and again there i s j u s t i f -i c a t i o n f o r assuming that a re l a t i o n s h i p e x i s t s between plywood sales to farmers and Gross National Product. Industry i s estimated to use 27 per cent of the plywood sold. The amount of plywood purchased by industry i s influenced by the pace of i n d u s t r i a l a c t i v i t y . Since the i n d u s t r i a l a c t i v i t y of the country i s r e f l e c t e d i n the Gross National Product there again appears to be a d e f i n i t e r e l a t i o n s h i p between the amount of plywood used by industry and the Gross National Product. 198. Gross National Product figures have been available f o r some time and w i l l continue to be avai l a b l e . Gross National Product figures of the past, though based on a large number of estimates, are generally considered to be reasonably accurate. In addition, estimates of these figures are continually being made available by economists. Residential Construction. Residential contractors use about 23 per cent of the plywood sold. Figures for r e s i d e n t i a l construction have been published by the Dominion Bureau of S t a t i s t i c s for many,years. Similar figures w i l l be available i n the future and they are cert a i n to be as accurate as the figures have been i n the past. Non-residential Construction About 13 per cent of the plywood sold was used fo r non-residential construction and a large part of the 13 per cent appears to have been used for i n d u s t r i a l , commercial and i n s t i t u t i o n a l construction. For many years now the Dominion Bureau of S t a t i s t i c s has published annual d o l l a r figures f o r these three segments of the construction industry. Personal Expenditure on Consumer Goods and Services As previously stated, there appears to be a close r e l a t i o n s h i p between Gross National Product and plywood sales and t h i s close r e l a t i o n s h i p j u s t i f i e s the incorporation of GNP into an estimating equation f o r plywood sales. But when other factors are incorporated into the equation, such as r e s i d e n t i a l , i n d u s t r i a l , commercial and i n s t i t u t i o n a l construction, the in c l u s i o n of GNP produces a problem, because GNP includes the t o t a l d o l l a r figures spent on r e s i d e n t i a l , i n d u s t r i a l , commercial and i n s t i t u t i o n a l construction. As a r e s u l t the in c l u s i o n of both GNP and construction figures i n the same equation r e s u l t s i n some double counting. In an attempt to decrease the amount of double counting, GNP can be replaced with Personal Expenditure on Consumer Goods and Services. When t h i s i s done only the r e s i d e n t i a l construction figures are counted twice. Selection of the Estimating Equations Current annual d o l l a r figures and constant 1957 annual d o l l a r figures were acquired for GNP 200. and personal expenditure on consumer goods and services. Current annual d o l l a r figures also were obtained f o r r e s i d e n t i a l , i n d u s t r i a l , commercial and i n s t i t u t i o n a l construction, In addition, Canadian annual softwood plywood consumption figures were acquired. For a l l seven series of data the figures covered the years 1943 to 1962. The construction figures were then re-stated i n 12 terms of 1957 d o l l a r s . (see Tables I I I and IV). When a l l seven series of data were examined, the trend l i n e that seemed to f i t each series of 13 data best was a straight l i n e . As a r e s u l t t h i s investigator decided that simple and multiple l i n e a r regression equations would be developed i n order to forecast softwood plywood sales f o r the years 1964 and 1968. 12. The deflators that would convert GNP current d o l l a r figures into constant 1957 d o l l a r figures were determined. After t h i s was done, the construction figures i n current d o l l a r s were mu l t i p l i e d by t h e i r appropriate d e f l a t o r s . This resulted i n the construction figures being re-stated i n constant 1957 d o l l a r s . 13. The f i r s t step i n the selection of the curve type was the p l o t t i n g of each series of data on graph paper. When t h i s was done, i t was possible by inspection to discover that i n each case the approp-r i a t e l i n e to be f i t t e d was a straight l i n e . This conclusion was v e r i f i e d by determining that the \" f i r s t - o r d e r differences\" f o r each of the various series of data were constant. 201. It was decided that four regression equations would be developed. In a l l four equations the dependent variable i s plywood sales. The f i r s t equation i s a simple l i n e a r regression equation with GNP as the independent v a r i a b l e . The second, t h i r d and fourth equations are multiple l i n e a r regression equations. The independent variables i n the second equation are r e s i d e n t i a l , i n d u s t r i a l , commercial and i n s t i t u t i o n a l construction. In the t h i r d equation the independent variable personal expenditure on consumer goods and services i s added to the four construction s e r i e s . The fourth equation contains f i v e independent variables - GNP, r e s i d e n t i a l , i n d u s t r i a l , commercial and i n s t i t u t i o n a l construc-14 t i o n . 14. The simple regression equation was developed by hand but the multiple regression equa-tions were produced by an IBM 1620 el e c t r o n i c computer. Instructions necessary f o r informing the computer how to prepare the desired multiple regression equations were programmed int o the computer. In addition, i t was decided that c e r t a i n information about the equations should be made ava i l a b l e . The c o e f f i c i e n t of c o r r e l a t i o n and the standard error of estimate were to be deter-mined f o r the simple regression equation. The c o e f f i c i e n t s of c o r r e l a t i o n , the c o e f f i c i e n t s of multiple determination, and the standard errors of estimates were t o be acquired - with the aid of the computer - f o r the multiple regression equations. TABLE III Dependent and Independent Variables GNP RESIDENTIAL INDUSTRIAL COMMERCIAL INSTITUTIONAL (000) (000) (000) (000) (000) 1943 $11,088,000 $ 63,684 $142,516 $ 38,873 $ 13,148 1944 11,850,000 83,928 75,862 29,918 21,006 1945 11,835,000 125,524 88,743 39,682 30,448 1946 11,850,000 193,627 157,573 84,894 48,624 1947 13,165,000 233,303 204,964 143,246 73,361 1948 15,120,000 255,756 242,832 166,073 121,421 1949 16,343,000 356,562 215,664 199,266 174,462 1950 18,006,000 508,525 274,849 211,763 206,219 1951 21,170,000 1,042,000 393,000 359,000 291,000 1952 23,995,000 1,029,000 422,000 326,000 314,000 1953 25,020,000 1,297,000 402,000 502,000 343,000 1954 24,871,000 1,400,000 364,000 546,000 377,000 1955 27,132,000 1,735,000 398,000 513,000 464,000 1956 30,585,000 1,902,000 604,000 571,000 455,000 1957 31,909,000 1,813,000 611,000 656,000 519,000 1958 32,894,000 2,189,000 396,000 689,000 550,000 1959 34,915,000 2,183,000 415,814 759,065 569,109 1960 36,254,000 1,913,000 451,568 738,127 615,188 1961 37,421,000 1,951,000 410,868 754,659 647,056 1962 40,401,000 2,115,100 492,676, 733,300 809,840 Sources: Dominion Bureau of S t a t i s t i c s , National Accounts Income and Expenditure, Queen's Printer, . Annual, Cat. No. 13-201, 1950-1963. Dominion Bureau of S t a t i s t i c s , Construction i n Canada, Queen 's Printer, Annual, Cat. No. 64-201, 1944-1963. TABLE I I I (Continued) Personal Expenditures on Plywood Bales Consumer goods and Services (sq. f t . 3/8\" basis) (000) (000) 1943 $ 5,727 162,500 1944 C\ 6,187 181,667 1945 6,811 170,000 1946 7,977 190,000 1947 9,173 246,667 1948 10,112 281,667 1949 10,963 250,000 1950 12,029 290,833 1951 13,273 332,503 1952 14,366 361,939 1953 15,112 450,541 1954 16,175 516,857 1955 17,389 652,270 1956 18,833 738,559 1957 20,072 701,761 1958 21,245 848,669 1959 22,591 845,560 1960 23,512 915,477 1961 24,486 1,038,246 1962 25,749 1,117,401 to O CO TABLE IV Variables i n Constant 1957 Dollars GNP RESIDENTIAL INDUSTRIAL COMMERCIAL INSTITUTIONAL PERSONAL EXPENDITURE ON CONSUMER GOODS AND SERVICES (000) (000) (000) (000) (000) (ooo: 1943 $20,317,000 $1,116,669 $261,089' $$71,215 $ 24,087 $10,492,000 1944 21,071,000 149,224 134,883 53,194 37,349 11,000,000 1945 20,575,000 218,161 154,235 68,967 52,919 11,838,000 1946 20,177,000 329,747 268,347 144,574 82,807 13,585,000 1947 20,439,000 362,320 318,309 222,461 113,930 14,246,000 1948 20,821,000 352,176 334,380 228,683 167,197 13,924,000 1949 21,626,000 471,732 285,323 263,629 230,813 14,504,000 1950 23,114,000 652,946 352,906 271,904 264,785 15,443,000 1951 24,531,000 1,207,678 455,487 416,081 337,269 15,384,000 1952 26,514,000 1,137,045 466,310 360,230 346,970 15,874,000 1953 27,525,000 1,426,700 422,000 552,200 377,300 16,623,000 1954 26,714,000 1,503,600 390,936 586,404 404,898 17,372,000 1955 29,018,000 1,856,450 425,860 548,910 496,480 18,606,000 1956 31,508,000 1,959,060 622,120 588,130 468,650 19,393,000 1957 31,909,000 1,813,000 611,000 656,000 519,000 20,072,000 1958 32,284,000 2,147,409 388,476 675,909 539,550 20,841,000 1959 33,398,000 2,089,131 397,934 726,425 544,637 21,620,000 I960 34,144,000 1,802,046 319,465 695,316 579,507 22,348,000 1961 35,023,000 1,826,136 384,457 706,361 605,644 22,919,000 1962 37,195,000 1,934,100 453,755 675,369 745,829 23,715,000 Source: Developed from Table I I I by multiplying the various figures by t h e i r appropriate d e f l a t o r s . 205. Least-squares Method The method that i s used i n t h i s study f o r f i t t i n g the straight l i n e s to the numerical data i s known as the method of least squares. The c r i t e r i o n of least squares demands that the l i n e that i s f i t t e d to the data be such that the sum of the squares of the v e r t i c a l deviations from the points to the l i n e be a minimum. While t h i s method i s used almost u n i v e r s a l l y f o r f i t t i n g l i n e s to numerical data i t has not escaped c r i t i c i s m because the assumptions t h i s method makes about the data are not always cor r e c t . The lea s t squares method requires that the independent variables be independ-ent of each other, that i s , that they not be 15 strongly i n t e r c o r r e l a t e d . There should be no auto-correlation within the various independent s e r i e s . I t also i s assumed t a c i t l y that there are no observational errors, that i s , that the s t a t i s -t i c a l data dealt with are based on measurements 16 that are exact. This method assumes that the depend-ent variable i s dependent and plays no part i n influencing the siz e of the independent v a r i a b l e s . 15. H. Wold, Demand Analysis, New York, J . Wiley, 1953, p. 28. 16. Wold, Demand Analyses, p. 38. 206. When these assumptions are not met the regression equation that i s developed from the l e a s t squares 17 technique can produce biased r e s u l t s . However, the extent of error may not be very serious and a forecasting equation developed by another tech-nique may produce equally biased r e s u l t s . Since the l e a s t squares technique i s well known, easy to use and usually produces r e s u l t s as acceptable as any other method, the investigator i s j u s t i f i e d i n using t h i s method fo r f i t t i n g straight l i n e s (and other curves) to numerical data. End-use Index Method The independent variables used i n t h i s study are not the only variables that could be or have been used to forecast plywood sales. In Business and Economic Forecasting a p r e d i c t i n g equation was 18 developed to forecast plywood sales i n the U.S. One of the independent variables incorporated into the equation was an end-use index. In order to develop the end-use index the various markets for plywood were defined and measured. Then the respective 17. Wold, Demand Analysis, p. 28. 18. M. Spencer, C. Clark, and P. Hoguet, Business and Economic Forecasting, R. Irwin, 1961, p. 331. markets were weighted by the proportion of t o t a l plywood sales consumed by each of these markets during the early n i n e t e e n - f i f t i e s and were then combined i n an additive manner to form the f i n a l end-use index. The end-use index method takes i n t o account the varying rates of growth i n the d i f f e r e n t end-use markets but the method used i n t h i s essay does not do t h i s . Moreover, neither method takes i n t o account changes i n the percentage of the t o t a l plywood sales consumed by the various markets over a period of time. The end-use index method i s not used i n t h i s essay because c e r t a i n necessary information i s not ava i l a b l e . To use t h i s method the forecaster must know what the various end-use markets are and he must possess t o t a l yearly consumption figures f o r each of the markets. This type of information i s not available on plywood sales i n Canada. The forecaster must also know the percentage of t o t a l plywood sales consumed by each of the end-use markets, and t h i s information i s not available f o r the Canadian plywood market. As a r e s u l t , a fore-cast based on an end-use index was not developed because c o l l e c t i n g information about the t o t a l d o l l a r consumption figures for each of the plywood markets and the percentage of t o t a l plywood sales consumed by each of the end-use markets would require a separate extensive study. VI. MAKING THE INDUSTRY FORECASTS A presentation and discussion of the four regression equations and pertinent information about these equations w i l l now be provided. The f i r s t equation to be developed was the simple l i n e a r regression equation. A simple l i n e a r regression equation i s of the form y = a -f bx. Here a*.and b are numerical constants and once they are known a predicted value of y for any given value of x can be calculated by d i r e c t s u b s t i t u t i o n . In the simple l i n e a r equation developed for t h i s study GNP i s the independent v a r i a b l e and plywood sales the dependent v a r i a b l e . In t h i s p a r t i c u l a r case the regression equation takes the form Y = -282, 148 + .03317 X x The standard deviation about a l i n e of average relati o n s h i p , being a measure of the accuracy of the estimates, i s c a l l e d the standard error of 19 estimate. Given an approximately normal d i s t r i b -ution of items about the l i n e of rela t i o n s h i p , 68 per cent of a l l the cases w i l l be within a range of +S, 95 per cent w i l l f a l l within +2S and 99 per cent w i l l f a l l within +3S. I f there were no scatter about the l i n e f i t t e d to the points representing the corresponding value of X and Y, S would have a value of zero, and the value of Y could be estimated from the value of X with perfect accuracy. The l e s s the dispersion about the least-squares l i n e , the smaller the value of S. The value of S serves, therefore, as an in d i c a t o r of the si g n i f i c a n c e and usefulness of the l i n e which describes the r e l a t i o n s h i p between the two v a r i a b l e s . The standard error of estimate for the simple regression equation i s Sy = 36,853,000 sq. f t . The c o e f f i c i e n t of c o r r e l a t i o n i s a measure of the goodness of f i t of the least-squares l i n e . I f the f i t i s poor, the c o e f f i c i e n t of c o r r e l a t i o n w i l l be close to 0. If the f i t i s good, the 19. F. M i l l s , S t a t i s t i c a l Methods. H. Holt and Company, 1938, p. 330. 20. M i l l s , S t a t i s t i c a l Methods, p. 332. c o e f f i c i e n t of c o r r e l a t i o n w i l l be close to +1 or 21 -1. The c o e f f i c i e n t of c o r r e l a t i o n (r) f o r the equation i s r = .994 Since r i s close to 1, we can say that the f i t of the least-squares l i n e i s extremely good. There i s a strong l i n e a r r e l a t i o n s h i p between X and Y. A multiple l i n e a r regression equation i s of the form y = a + bx^ + c x 2 + dx^ The second equation, a multiple l i n e a r regression equation, takes the form Y = 162,693 + .06637X3 - ,55849X4 + .03608XP + 1.34605X, 5 6 where X, i s r e s i d e n t i a l construction, X. i s i n d u s t r i a l 3 4 construction, X i s commercial construction and X. 5 6 i s i n s t i t u t i o n a l construction. A f t e r working out the equation by which the values of one variable may be estimated from those 21. J . Freund and F. Williams, Modern Business S t a t i s t i c s , Prentice-Hall, 1958, p. 309. of several independent variables, i t i s desirable to have some measure of how c l o s e l y such estimates agree with the actual values and of how c l o s e l y the v a r i a t i o n i n the dependent variable i s associated with the v a r i a t i o n i n the Independent: v a r i a b l e . The standard error of the estimate f o r a multiple regression equation measures the closeness with which 22 the estimated values agree with the o r i g i n a l values. Sy = 93,281,000 sq. f t . The c o e f f i c i e n t of multiple determination i s the square of the c o e f f i c i e n t of multiple c o r r e l a t i o n . The c o e f f i c i e n t of multiple determination (R 2) o f f e r s a measure of the proportion of the v a r i a t i o n i n the dependent factor which can be explained by, or i s 23 associated with, v a r i a t i o n i n the independent f a c t o r s . R 2 = .930 2 Since R i s close to 1 the f i t of the least-squares l i n e i s extremely good. The t h i r d equation, a multiple l i n e a r regression equation, takes the form Y = -659,579 + .07452X2 -f .14958X3 - .33454X4 - .51256X,. + .21947X,. 5 o 22. M. Ezekiel and K. Fox, Methods of Correlation and Regression Analysis, New York, J . Wiley and Sons, 1959, p. 188. 23. M. E z e k i e l and Fox, Methods of Cor r e l a t i o n and Recrression Analysis, p. 191. where i s personal expenditure on consumer goods and services, X^ i s r e s i d e n t i a l construction, X^ i s i n d u s t r i a l construction, X 5 i s commercial construction and X- i s i n s t i t u t i o n a l construction. The Standard 6 Error of Estimate i s Sy = 69,183,000 sq. f t . The C o e f f i c i e n t of Multiple Determination i s R 2 = .964 Since R i s close to 1 the f i t of the l e a s t squares l i n e again i s extremely good. The f i n a l multiple l i n e a r regression equation takes the form Y = -185182 + . 0 5 0 7 9 X 3 ^ - . O 8 O 8 6 X 3 - .301213^ + .21172XC + .23431X^ 5 6 where X^ i s Gross National Product, X 3 i s r e s i d e n t i a l construction, X i s i n d u s t r i a l construction, X i s commercial construction and X\u00E2\u0080\u009E i s i n s t i t u t i o n a l construc-6 t i o n . The Standard Error of Estimate i s Sy = 59,101,000 sq. f t . The C o e f f i c i e n t of Multiple Determination i s R 2 = .974 Since R 2 i s close to 1 the f i t of the l e a s t squares l i n e i s extremely good. In a l l four equations the least squares technique produces an excellent f i t . However, i n each of the multiple regression equations one or more of the constants associated with the independent variables are negative. Since the c o e f f i c i e n t s of c o r r e l a t i o n between each of the independent variables and the dependent va r i a b l e are p o s i t i v e (see Table V), one would expect a l l of the constants to be p o s i t i v e . Because the multiple regression equations contain these negative figures when a l l of the c o e f f i c i e n t s of c o r r e l a t i o n are p o s i t i v e , one or more of the assump-tions underlying the use of the lea s t squares technique and the regression equation has not been met. Prom Table V i t can be seen that the c o e f f i c i e n t of c o r r e l a t i o n between p a i r s of the independent variables are extremely high. This indicates that there i s a considerable amount of i n t e r c o r r e l a t i o n between these v a r i a b l e s . But a basic assumption underlying the use of the l e a s t squares technique i s that there must be no i n t e r c o r r e l a t i o n between p a i r s of independent v a r i a b l e s . The f a i l u r e to meet t h i s prerequisite probably accounts for the undesirable presence of the negative figures i n the equation. None of the multiple l inear regression equations produce a better f i t to the data than the simple regression equation and they contain negative figures that should not be present. Only the simple l inear regression equation i s reasonable and log i ca l . As a result , i t seems desirable to select this regression equation for use as the forecasting model. 01 02 03 04 05 06 07 TABLE V Coefficients of Correlation For the Dependent and Independent Variables 01 GNP 1.000 .940 .656 .940 .969 .981 02 Personal Expenditure 1.000 .927 .614 .955 .975 .975 03 Residential 1.000 .755 .974 .951 .906 04 05 06 I n d u s t r i a l Commercial Institu** t i o n a l 1.000 .735 .713 .583 1.000 .962 .918 1.000 .952 07 Plywood Sales 1.000 Sources Analysis of the s i x independent series of data and the single dependent series of data. VII.CANADIAN ECONOMY Since the simple regression equation containing Gross National Product as the independent variable i s to be used to forecast plywood sales, i t seems desirable that the recent h i s t o r y of the Canadian economy be reviewed. A knowledge of Canada's economic h i s t o r y may shed some l i g h t on the future path to be followed by the economy and thus make a Gross National Product p r e d i c t i o n more accurate or at l e a s t make i t easier to j u s t i f y a forecasted Gross National Product. Since World War I I , growth i n the Canadian economy has been phenomenal. In constant d o l l a r terms the GNP increased by 124 per cent between 1939 and 1945, which i s an annual compound rate increase of 5.16 per cent. The per capita Gross National Product increased i n r e a l terms by 61 per cent, or at an annual compound rate of 3.04 per cent. During t h i s same period the index of i n d u s t r i a l production i n -creased by 142 per cent or at an annual compound rate 24 of 5.68 per cent. Growth of t h i s magnitude was possible because of concurrent growth i n the basic 24. Canada, Royal Commission on Canada's Economic Prospects, Queen's Printer, 1958, p. 79. 217. factors o f production. The population increased r a p i d l y r e s u l t i n g i n an increase i n the labor force, and a large amount of c a p i t a l became avail a b l e f o r the exploration and development of natural resources. The greatest increase i n economic expansion, however, was i n the area of secondary manufacturing. The two factors l a r g e l y responsible f o r t h i s expansion w e r e the absence of foreign competition and the development of improved methods and materials as a r e s u l t of i n t e n s i f i e d research. I t i s estimated that between 1939 and 1944 the output of secondary manufacturing 25 Industries increased by 160 per cent. As might be expected, government expenditure provided the c h i e f stimulus to economic expansion during the war years. Government expenditure on goods and services, including exports d i r e c t l y or i n d i r e c t l y financed by government, represented about 26 40 per cent of the Gross National Product. After the end of the war, when the economy was being converted to peacetime production, and on through 25. Canada, Canada's Economic Prospects, 1958, p. 87. 26. Canada, Canada's Economic Prospects, 1958, p. 87. 218. 1949 when the t o t a l output again began to r i s e , the p r i n c i p a l stimulus to economic growth and a c t i v i t y 27 appears to have been consumer expenditure. At the war's end there was a large backlog of u n s a t i s f i e d consumer demand and a sizeable accumulation of personal savings. As l e v e l s of consumer income remained f a i r l y constant the consumer demands and a v a i l a b l e savings combined to cause such a stream of consumer spending that the economy s h i f t e d from a wartime to a c i v i l i a n basis without either an undue decline i n output or a disturbing increase i n unemploy-ment. Although consumer demand seems to have provided the main economic stimulus from 1945 to 1949, priva t e investment and external demand exerted a strong influence during those years. New housing developments appeared on the edges of Canadian c i t i e s and these housing developments have been increasing ever since. Farmers bought combines and t r a c t o r s i n such numbers that farming procedures i n the p r a i r i e provinces were completely changed. The outbreak of war i n Korea i n 1950 resulted i n 27. Canada, Canada's Economic Prospects, 1958, p. 88. extensive export demand accompanied by heavy c a p i t a l spending and another economic boom came in to be ing . Th i s per iod saw the cont inuat ion of a strong demand for consumer goods. The fac t that there was a wide d i s t r i b u t i o n and a l e v e l l i n g up of incomes throughout the populat ion enabled consumers to s a t i s f y many of t h e i r d e s i r e s . By 1955 Canada had developed a much more d i v e r s i f i e d economy than she had i n 1939. A large number of new i n d u s t r i e s were f i r m l y es tabl i shed and secondary manufacturing occupied a stronger p o s i t i o n i n the economy than i t had ever done before . In a d d i t i o n , the sources of raw mater ia l s were grea t ly extended. Commodities exported i n s i g n i f i c a n t volume covered a wider range than they had done prev ious ly , but t o t a l exports accounted for a much smaller segment of the Gross Nat ional Product than they had done i n 1939. Th i s was caused by the d e c l i n i n g importance of a g r i c -u l t u r a l exports i n r e l a t i o n to t o t a l output. During 1955 Canada recovered r a p i d l y from the m i l d contract ion that ex i s ted from the middle of 1953 to the middle of 1954. In the l a t t e r part of 1954 the output of goods and services began to r i s e and continued to r i s e so that the output for 1955 was 28 ten per cent greater than the output f o r 1954. In 1955 and 1956 there was a great increase i n the nation's production hut i n 1957 there was a noticeable l e v e l l i n g o f f i n the rate of economic a c t i v i t y . The great increase i n the outlays of investment c a p i t a l that characterized 1955 and 1956 moderated i n 1957, investment i n machinery and other equipment declined i n the l a s t three-quarters of the year and outlays f o r non-residential construction decreased. However, a f t e r a series of declines, r e s i d e n t i a l con-st r u c t i o n began to increase during 1957. During 1958 economic a c t i v i t y as a whole increased, and the Gross National Product gradually resumed an upward course. By the second quarter of the year the upward trend was more f i r m l y established and by 29 the c l o s i n g quarter i t had begun to gather momentum. During 1959 improvement continued on a broad f r o n t . There was a slackening of economic a c t i v i t y during the f i r s t quarter of 1960 followed by a s l i g h t decline i n the second quarter which caused a small decrease 28. Dominion Bureau of S t a t i s t i c s , Canada Year Book, 1956, Queen's Printer, 1957, p. v i i i . 29. Dominion Bureau of S t a t i s t i c s , Canada Year Book, 1958, Queen's Printer, 1959, p. i x . i n the Gross National Product, when seasonal factors are taken into account. Nevertheless, aggregative measures of a c t i v i t y were higher than those for the same period of 1959. The slackening of a c t i v i t y during the f i r s t h a l f of 1960 was related to a major decrease i n housebuilding and to a decrease i n the rate of accumulation of stocks. By the fourth quarter of I960, however, economic a c t i v i t y once again resumed an upward trend. Support f o r t h i s trend came mainly from expansion i n government outlays f o r goods and services, and from a sharp increase i n the demand f o r Canadian export products. The advance i n Gross National Product was resumed 30 i n 1961. The f i r s t quarter of the year was weak but the l e v e l of a c t i v i t y moved on a r i s i n g trend f o r the rest of the year. One of the c h a r a c t e r i s t i c s of t h i s upward trend was a sharp increase i n imports. Furthermore, during the l a t t e r h a l f of the year business inventories showed a modest improvement and consumer expenditures increased at a moderate rate . Outlays f o r goods and services by the government continued to b o l s t e r the economy throughout 1961. 30. Dominion Bureau of S t a t i s t i c s , Canada Year Book, 1961, Queen's Printer, 1962, p. 1058. 1962 proved to be another good year i n terms of output. The Gross National Product increased eight per cent i n value and s i x per cent i n physical volume, and r e a l output per person increased markedly fo r the f i r s t time since 1956. The volume of production increased during the f i r s t h a l f of the year and p r i c e s increased during the second h a l f of the year, the l a t t e r due p a r t l y to the depreciation of the Canadian 31 d o l l a r . The increased a c t i v i t y was widespread geographically and i n v a r i e t y of employment and re-ceived further impetus because of a good wheat crop i n the f a l l . The important stimulus, however, were a sizeable increase i n merchandise exports and an important increase i n the outlays of goods and services by p r o v i n c i a l and municipal governments. Business conditions i n Canada remained favorable throughout 1963. The flow of exports exerted an influence that permeated business and industry. The construction industry operated successfully and farm income was good. There was l i t t l e incentive to stock up on inventories and the supply of goods and labor was adequate. 31. Canadian Imperial Bank of Commerce, Commercial Letter, May-June, 1963, p. 1. The Gross National Product for 1963 was about 6.6 per cent above the 1962 l e v e l . There was some increase i n p r i c e s as higher costs worked t h e i r way through the p r i c e system, but most of the increase i n GNP was i n r e a l output. Production, employment, incomes, p r o f i t s , consumption and foreign trade reached record high l e v e l s during the year. Capital spending i n the business sector of the economy was at an a l l - t i m e 32 high i n terms of current d o l l a r s . At a forum of members of the National I n d u s t r i a l Conference Board a panel of economists predicted that Canada's Gross National Product f o r 1964 w i l l be 33 $44.6 b i l l i o n s , a gain of 4.5 per cent over 1963. Only about one per cent of t h i s increase i s l i k e l y to r e s u l t from p r i c e r i s e s . The panel further predicted that spending f o r consumer goods and services w i l l t o t a l $28 b i l l i o n , a r i s e of four per cent over the estimated expenditure f o r 1963. One economist expressed the opinion that spending on consumer durables which reached a peak i n the second quarter of 1963 when spending ran at an annual rate of 3.1 b i l l i o n , w i l l 32. Canadian Imperial Bank of Commerce, Commercial Letter, May-June, 1964, p. 1. 33. \"Guidelines to '64,\" The F i n a n c i a l Post, Toronto, November 1, 1963, p. 17. 34 probably not reach that rate again i n 1964. Another economist decided that the boom i n the car market i s losing momentum. There w i l l probably be a four per cent gain i n spending on non-durables, one per cent of which w i l l be caused by p r i c e increases. Pood pric e s are expected to r i s e two per cent and clothing 35 pr i c e s one per cent. Government spending i s sure to strengthen the economy during 1964, but the second h a l f of the year may show a firmer trend than the f i r s t h a l f . Federal expenditures on goods and services w i l l increase s l i g h t l y to $3.2 b i l l i o n , while p r o v i n c i a l and munic-i p a l expenditures w i l l increase by approximately 9.5 per cent to $5.7 b i l l i o n , which i s a r i s e of seven 36 per cent over the 1963 estimated l e v e l . One economist suggested that 1964 be regarded as the threshold to another period of vigorous investment expansion, another gave 1965 as the s t a r t i n g year. 34. \"Guidelines to \u00E2\u0080\u00A264\", The F i n a n c i a l Post Toronto, November 1, 1963, p. 17. 35. \"Guidelines to \u00E2\u0080\u00A264\", The F i n a n c i a l Post Toronto, November 1, 1963, p. 17. 36. \"Guidelines to '64\", The F i n a n c i a l Post Toronto, November 1, 1963, p. 17. Differences of opinion regarding the timing of t h i s expansion resulted i n estimates of a one per cent decline f o r 1964 to an eight per cent increase i n 37 c a p i t a l spending. An averaging of the forecasts c a l l e d f o r an investment of $7.7 b i l l i o n i n 1964, which i s a gain of four per cent over 1963. Housing s t a r t s are expected to go down to about 130,000 i n 1964, and housing expenditures are expected to s l i p by two per cent or three per cent. To t a l exports of goods and services are expected to r i s e 5.5 per cent to $9.3 b i l l i o n , while imports w i l l probably r i s e three per cent to $9.5 b i l l i o n . This would leave a current account d e f i c i t of $200,000,000. The consensus of opinion was that 1964 w i l l be a good year with 1965 or 1966 even be t t e r . The Canadian economy i s expected t o expand more ra p i d l y i n the years 1954 to 1979 than i t d i d i n the 38 years 1928 to 1953. Three reasons may be given i n support of these expectations. 37. \"Guidelines to '64\", The F i n a n c i a l Post, Toronto, November 1, 1963, p. 17. 38. O.J. Firestone and others, ed., Growth and Future of the Canadian Market, Ottawa, 1956, p. 112. 1. Less Serious Disturbances Expected During the early part of the period 1928 to 1953 Canada experienced the worst depression i n her h i s t o r y . Most other nations of the world suffered from the same disturbance. Canada's long-terra economic expansion was seriously retarded f o r a long time by t h i s depression. This same period included s i x years of World War II which stimulated greatly both i n d u s t r i a l expansion and d i v e r s i f i c -ation, but which also channeled a s i g n i f i c a n t proportion of the nation's energies into m i l i t a r y purposes. Neither serious depression nor a l l - o u t 39 war are expected during 1954 to 1979. 2. More Research and S c i e n t i f i c Work In d u s t r i a l research and s c i e n t i f i c work influence economic development i n Canada f a r more than they ever d i d before and they are expected to have considerable e f f e c t upon the Canadian 40 economy. 39. Firestone, Growth and Future of the Canadian Market, p. 112. 40. Firestone, Growth and Future of the Canadian Market, p. 113. 3. Greater Experience and Confidence Canadian management i s more experienced and e f f i c -ient than i t was i n the early part of the twentieth century; Canadian workers have acquired greater and more varied s k i l l s and c a p i t a l equipment i s ava i l a b l e i n larger quantities and Canadian c i t i z e n s are more confident concerning the country's 41 future economic achievements. Despite the knowledge and research that are u t i l i z e d i n making these predictions the predictions cannot be guaranteed for there are important factors that could prevent Canada from r e a l i z i n g the anticipated development. The economy could be disrupted by a series of crop f a i l u r e s , by serious labor s t r i f e or by a breakdown i n i n t e r n a t i o n a l trade. I f investment expenditures decreased' or i f growth were temporarily halted a slowdown i n Canada's economic growth would r e s u l t . V I I I . PLYWOOD SALES FOR 1964 The preparation of the domestic softwood plywood sales forecast for 1964 can now be described. Gross 41. Firestone, Growth and Future of the Canadian Market, p. 113. National Product, the independent variable i n the fore-casting equation, must be determined for the year 1964. The components of GNP are consumption, government spend-ing, investment and foreign trade. These components were c o l l e c t e d i n current d o l l a r s f o r the years 1948 to 1962, and then, with the a i d of the deflators prev-io u s l y mentioned, were converted into constant 1957 d o l l a r s . (see Table V I ) . Linear trend l i n e s were f i t t e d to these figures using the leaist squares technique and estimates of the various components were obtained f o r 1964. For con-sumption and government and investment spending the figures for the years 1948 to 1962 were used; f o r foreign trade the figures f o r the years 1953 to 1962 were used. A shorter time span was used f o r foreign trade because a basic change seems to have taken place i n t h i s area during recent years. Since 1953 there has been a large annual foreign trade d e f i c i t . The figures i n 1957 d o l l a r s f o r the various components of the 1964 GNP are as follows: Foreign Trade Consumption Government Investment 25,530,000,000 7,576,000,000 7,630,000,000 -1,005,000,000 Gross National Product $ 39,731,000,000 Reference has already been made to the f a c t that Canadian economists estimate a 1964 GNP of $44.6 b i l l i o n i n current d o l l a r s . In terms of 1957 d o l l a r s t h i s amount i s approximately $39.7 b i l l i o n , so the two independent estimates a r e tlje same i n t o t a l . Since the trend l i n e produced the same estimated GNP figure as the economists, the GNP figur e produced by the trend l i n e w i l l be used f o r forecasting 1964 plywood sales. The c a l c u l a t i o n of the domestic softwood plywood sales forecast f o r the year 1964 i s given below. The simple l i n e a r regression equation previously developed and accepted was used f o r t h i s purpose. Y = -282,148 + .033175^ where X^ ^ i s Gross National Product and Y i s domestic softwood plywood sales. Y = -282,148 + .03317 ($39,731,000,000) Y = 1,317,880,000 sq. f t . The sale of softwood plywood i n Canada during 1964 i s expected to t o t a l , 1,317,880,000 sq. f t . of 3/8\" basis.' TABLE VI Components of GNP i n Constant 1957 Dollars Consumption Government Investment Foreign (000) (000) (000) (000) 1948 $ 13,049 $ 2,839 $ 4,043 517 1949 13,637 3,175 4,390 189 1950 14,542 3, 349 5,146 -260 1951 14,759 4,188 5,659 -375 1952 15,775 5,250 5,675 -55 1953 16,658 5,251 6,268 -561 1954 17,044 5,098 5,095 -485 1955 18,304 5,319 6,137 -931 1956 19,478 5,664 8,100 -1588 1957 20,072 5,722 7,566 -1422 1958 20,707 6,113 6,484 -948 1959 21,711 6,205 6,900 -1448 1960 22,357 6,255 6,576 -1136 1961 23,070 6,544 5,977 -710 1962 23,915 6,761 L 6,806 -490 Source: Components of GNP were multiplied by t h e i r appropriate d e f l a t o r s \u00E2\u0080\u00A2 to CO o IX. PLYWOOD SALES FOR 1963 231. The simple l i n e a r regression equation can be used to forecast sales of plywood i n Canada during 1968. The s i z e of the independent variable, GNP, must be estimated f o r t h i s year. For t h i s forecast a d i f f e r e n t technique i s used to ca l c u l a t e GNP. Assumptions are needed of the rate of growth of f i v e economic factors -population, labor force, persons employed, hours worked per week, and output per man hour. At t h i s point the factors that have determined the s i z e of Canada's population i n the past and w i l l determine the rate of growth i n the future should be mentioned. Canada's Population Many forecasts have been made concerning the siz e of Canada's population a few years hence. The fore-casts vary g r e a t l y . The Dominion Bureau of S t a t i s t i c s has produced population projections f o r 1971 showing that Canada may have a population i n that year of twenty-one to twenty-four m i l l i o n . I f such a population i s to materialize there must be annual growth r a t i o s of 1.97 per cent or 2.73 per cent (compound) respectively. An approximate rate of population growth can be determined by examining past rates of growth and by considering some of the factors that w i l l have a bear-ing on the expansion of Canada's population during the 42 next few decades. 1. B i r t h Rate Canada's present b i r t h rate i s higher than i t was before the war. Although there are sure to be fluctuations t h i s upward trend i s expected to continue. The second h a l f of the s i x t i e s are expected to show more rapid population growth than the f i r s t h a l f of the s i x t i e s . 2. Death Rate There has been a steady decline i n the Canadian death rate and t h i s decline i s expected to continue. 3. Immigration Since the war l e v e l s of immigration have varied with the prosperity of the country. Immigration Has at a high rate when the economy was expanding r a p i d l y and was at a low rate when the economy was expanding slowly. 4. Emigration So long as the economy expands most Canadians 42. Firestone, Growth and Future.of the Canadian Market, p. 102. w i l l f i n d increasing opportunities to prosper i n t h e i r own country, so emigration i s not expected to r i s e to any great extent. TABLE VII Number of Families and Number of Households i n Canada No. of Families 1951 3,282,445 1952 3,413,000 1953 3,477,000 1954 3,595,000 1955 3,685,000 1956 3,705,000 1957 3,849,000 1958 3,953,000 1959 4,038,000 1960 4,138,000 1961 4,140,384 1962 4,239,000 No. of Households 3,409,295 3,561,000 3,675,000 3,785,000 3,891,000 3,974,000 4,055,000 4,173,000 4,303,000 4,404,000 4,509,000 4,592,000 Sources: Dominion Bureau of S t a t i s t i c s , Estimates of Households and Families i n Canada, Queen * s Print e r , Cat. No. 91-204, 1952-1956. Dominion Bureau of S t a t i s t i c s , Estimates of Families i n Canada, Queen's Printer, Annual, Cat. No. 91-204, 1957-1963. Determination of Gross National Product f o r 1968 i n 1957 D o l l a r s . The population of Canada has increased s t e a d i l y over the l a s t ten years. The average annual increase i n population between 1953 and 1962 was 414,000. (see Table V I I I ) . I f thjs average increase continues u n t i l 1968, Canada's population i n that year should be 21,100,000. Sometimes the growth of Canada's labor force has been more rapid than the population growth and sometimes less r a p i d . Between 1953 and 1962 the labor 43 force averaged 36.5 per cent of the population. I f t h i s r e l a t i o n s h i p p r e v a i l s i n 1968 the labor force w i l l be 7,701,500. Unemployment averaged f i v e per cent of the popul-44 ation during the same period of 1953 to 1962. I f the unemployment i s f i v e per cent of the labor force i n 1968 the unemployed w i l l t o t a l 335,000. The number of persons employed w i l l then be 7,316,425. In 1953 the number of hours worked per week i n Canadian manufacturing industries averaged 41.3. Assuming that employees worked 50.0 weeks per year, the employees worked a t o t a l of 2,065 hours per year. 43. Canadian Imperial Bank of Commerce, Commercial Letter, May-June, 1963, p. 8. 44. Canadian Imperial Bank of Commerce, Commercial Letter, May-June, 1963, p. 8. In 1962 the number of hours worked per week i n Canadian manufacturing industries averaged 40.7. If employees worked 49.5 weeks per year, the employees worked a t o t a l of 2,015 hours per year. The output per employed person i n 1953 was $5,258, and i n 1962 the output per employed person was $5,983. These figures are i n 1957 d o l l a r s . If the output per employed person was $5,258 i n 1953 and the hours worked per employed person were 2,065 i n 1953, the output per man per hour was $2,546. In 1962 the output per employed person was $5,983 and the hours worked per employed person were 2,015, therefore, the output per man per hour was $2,983. The increase i n output per man hour from 1953 to 1962 resulted i n an average increase per year of $0,047. Thus the increase i n output per man hour and per man year was 1.7 per cent. The t o t a l reduction i n working hours per year between 1953 and 1962 was f o r t y hours. With ah average reduction of 5.5 hours per year and an average number of hours worked per year of 2,045 there was a 0.2 per cent reduction i n hours per year of work. As a r e s u l t : 1.7 per cent - 0.2 per cent = 1.5 percent increase i n output per year +0.1 percent adjustment f o r expected improve-ment. = 1.6 percent increase i n output per year. Therefore the output per employed person i n 1968 should be $6,581. With an employed labor force of 7,316,425 the gross national product i n 1968 i n 1957 d o l l a r s w i l l be $48,149,000,000. TABLE VIII Structure and Performance of tfee Economy 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 Gross National (millions of dollars) Product At Market Prices 25,020 24,871 27,132 30,585 31,909 32,894 34,915 36,254 37,421 40,401 Constant(1957)dollars 27,525 26,714,,29,018 31,508 31,909 32,284 33,398 34,144 35,023 37,195 Population and Employment To t a l Population > (June 1) 14,845 15,287 15,698 16,081 16,610 17,080 17,483 17,870 18,238 18,570 C i v i l i a n Labor Force (Annual Avg.) 5,397 5,493 5,610 5,782 6,003 6,127 6,228 6,403 6,518 6,608 Emplyment (Annual Avg.) 5,235 5,243 5,364 5,585 5,725 5,695 5,856 5,955 6,049 6,217 Unemployment (Annual Avg.i per cent of labor force) 3.0 4.6 4.4 3.4 4.6 7.1 6.0 7.0 7.2 6.0 Source: Canadian Imperial Bank of Commerce, Commercial Letter, May-June, 1963, p. 8. to The c a l c u l a t i o n of the domestic softwood plywood sales forecast f o r 1968 i s given below. The simple l i n e a r regression equation was used for t h i s purpose. Y = -282,148 + .OSSITX^^ where i s Gross National Product and Y i s domestic softwood plywood sales. Y = -282,148 + .03317 ($48,149,000,000) Y = 1,597,100,000 sq. f t . The sale of softwood plywood i n Canada during 1968 i s expected to be 1,597,100,000 sq. f t . , 3/8\" b a s i s . X. CROWN ZELLERBACH SALES The t o t a l sales to be achieved by Crown Zellerbach can be determined now that the industry sales estim-ates are a v a i l a b l e . Crown Zellerbach management have indicated that they plan t o r e t a i n t h e i r share of the plywood market i n Canada. As previously pointed out they have increased the capacity of t h e i r manu-facturing plant and have b u i l t up a national sales organization. The percentage of the t o t a l domestic market r e a l i z e d by Crown Zellerbach i s c o n f i d e n t i a l inform-ation. However, since the company knows t h i s per-centage and the expected industry sales of plywood, the company forecaster can c a l c u l a t e the amount of plywood to be sold i n 1964 and 1968. For the sake of discussion l e t us suppose that Crown Zellerbach has at present 12 per cent of the industry sales. Then # In:1964 t h e i r domestic s o f t -wood plywood sales should be: 12 per cent of 1,317,880,000 sq. f t . , 3/8\" basis = 158,145,600 sq. f t . , 3/8\" basis and i n 1968 t h e i r sales should be: 12 per cent of 1,597,100,000 sq. f t . 3/8\" basis = 191,652,000 sq. f t . 3/8\" b a s i s . CHAPTER VII SUMMARY AND CONCLUSION A l l companies face a future that holds an endless array of problems f o r them. An increasing awareness of t h i s f a c t and the increasing dimensions of the problems have induced management to place an increased emphasis on corporate planning. E s s e n t i a l to good planning i s forecasting - forecasting the economy, the industry volume and the company sales - f o r the short term or over an extended period, or both. With the assistance of such guidelines v i t a l marketing, f i n a n c i a l and production plans ultimately emerge, together with t h e i r supporting schedules. Without adequate forecasting these plans and schedules can go woefully astray. Sales forecasts emanate from a number of sources. The salesman or h i s d i s t r i c t manager i s involved i n the preparation of f i e l d sales forecasts. Sometimes a senior company executive, such as the general sales manager, prepares the forecasts. In a number of companies the market research department i s respons-i b l e . In some companies an economist or an economics department provides sales forecasts. Sometimes a l l these people can be involved. In large companies espec i a l l y , the number of personnel or departments involved i n forecasting can be numerous. However, only one person should have o v e r - a l l r e s p o n s i b i l i t y for the forecasting task and the extent to which he i s responsible should be c l e a r l y defined. To be e f f e c t i v e , t h i s i n d i v i d u a l must acquire not only a detailed understanding of company a c t i v i t i e s but a thorough knowledge of the character-i s t i c s of a sound forecasting operation. Not only must the forecaster be f a m i l i a r with the techniques or methods available for developing forecasts, but he must understand also the necessity of carrying out various pre-performance and post-performance a c t i v i t i e s . While an executive can determine how accurate a forecast i s a f t e r the f a c t , only the soundness of the forecasting operation can be appraised before the f a c t . Frequently a simple naive approach i s used. Company sales are forecast by reference only to past sales with modifications based on opinions gathered from the f i e l d . What general economic conditions are l i k e l y to be, or what the industry i s l i k e l y to s e l l , or how the company's markets are going to fare -these factors are overlooked a l l too often. Since no company exists i n a vacuum, a forecast prepared i n a vacuum i s prone to excessive error because i t f a i l s to r e l a t e the company sales a c t i v i t i e s to the opportunities i n the market place. In forecasts covering a f u l l year ahead, substan-t i a l deviations i n even the best of forecasts w i l l occur, i f not i n the t o t a l then i n some product or t e r r i t o r i a l component of the forecast. When deviations a r i s e and there i s no time-table for review and r e v i s i o n management hesitates to develop a new f o r e -cast. Under these circumstances, i f a r e v i s i o n f i n a l l y i s decided upon, i t probably w i l l be drawn up h a s t i l y and c a r e l e s s l y . On the other hand, when a r e v i s i o n i s deferred i n d e f i n i t e l y , a number of \"private\" forecasts may a r i s e i n d i f f e r e n t departments. A s i t u a t i o n of t h i s type leads to independent, unco-ordinated decisions, and planning deteriorates. Scheduled reviews of forecasts with the opportunity to revise them when necessary undoubtedly are d e s i r -able i f the forecasts are to be of maximum usefulness. In addition there i s a need for post-mortems. Si g n i f i c a n t deviations between actual and forecasted figures should be appraised and explained. Not only should a forecaster acquire a knowledge of the c h a r a c t e r i s t i c s of a good forecasting operation, but he should possess also an appreciation of the lim i t a t i o n s and problems inherent i n forecasting. A forecaster cannot assume that h i s forecasted figures are exact and c e r t a i n . He faces a s i t u a t i o n s i m i l a r to the production executive. The production man has h i s tolerances: he permits a deviation within a c e r t a i n range of values, and h i s f a m i l i a r i t y with p r o b a b i l i t y i n q u a l i t y control r e s u l t s i n h i s expectation of the occasional f a u l t y product. The forecaster faces many problems. For example, much of the government data used to forecast the economic and industry outlook consists of estimates, and these estimates contain e r r o r s . Sometimes these errors are large, sometimes they are small, but always they are present. The forecaster's awareness of t h i s type of problem prompts him to caution i n using and projecting f i g u r e s . This thesis has presented a general review of sales forecasting l i t e r a t u r e with p a r t i c u l a r attention t o the preparing of the forecast, the pre-planning and the review. In addition, forecasts were prepared showing the expected sales of domestic softwood p l y -wood to be achieved by the plywood industry and by Crown Zellerbach Company for the years 1964 and 1968. A simple regression equation and three multiple regression equations were produced with the intention of using them to forecast industry plywood sales. The three multiple regression equations contained high c o e f f i c i e n t s of multiple determination, but each of the equations also possessed one or more unaccept-able negative constants. One of the basic assumptions underlying the development of the multiple regression equations apparently had not been met. Because of t h i s d i f f i c u l t y a l l of these equations were rejected f o r use i n preparing a forecast. A simple regression equation was developed which possessed an extremely high c o e f f i c i e n t of c o r r e l a t i o n and a small standard error of estimate. Since t h i s equation contained these desirable features and seemed to incorporate no underlying f a l l a c y , t h i s Simple regression equation was the one used to forecast industry plywood sales. This simple equation i s acceptable i f one wishes to estimate accurately one variable when another v a r i a b l e i s given, but i f the objective i s to obtain an explanation of one va r i a b l e as a function of one or more other variables, the simple equation cannot be considered e n t i r e l y acceptable. Gross National Product must be viewed as providing an explanation of plywood sales i n only the broadest sense. I f the multiple regression equations had been acceptable, the objective of obtaining an explanation of one variable as a function of one or more other variables would have been r e a l i z e d to a much greater degree. However, the high c o e f f i c i e n t of c o r r e l a t i o n , the low standard error of estimate and the reasonableness of the simple regression equation make i t acceptable fo r predicting industry plywood sales. Since the management of Crown Zellerbach Company has taken action necessary f o r reta i n i n g t h e i r share-of-market i n an expanding industry, the assumption that they w i l l obtain the same percentage of the industry market as they now possess seems reasonable. This j u s t i f i e s multiplying the t o t a l projected industry sales figures by a percentage which represents the company's present share of the t o t a l market. The development of 1964 and 1968 plywood sales f o r e -casts for Crown Zellerbach r e s u l t s i n the achievement of the objective stated at the beginning of the study. Since the simple equation explains plywood sales i n only the broadest sense and does not d i r e c t l y r e l a t e the dependent variable to various known independent variables, the management of Crown Zellerbach should not depend e n t i r e l y on t h i s equation when forecasting plywood sales. Forecasts developed by using other techniques should be made before a f i n a l forecast i s chosen. A desirable s t a t i s t i c a l technique to use would be the end-use index method described i n Chapter VI. In t h i s method the various markets for plywood are defined, measured, weighted and then combined i n an additive manner to form the f i n a l end-use index. Unfortunately, the necessary information about the plywood markets i s not available, so t h i s method cannot be used at the present time. A judgment method which produces s a t i s f a c t o r y r e s u l t s i s used by the Crown-Zellerbach Company fore-caster. He consults with salesmen, sales managers, dealers and others who have knowledge about the plywood market. The forecaster receives information about the plywood industry from the plywood manufacturers association and he receives information about general business conditions from various sources. The f o r e -caster evaluates the information that he has gathered and determines how much plywood Crown-Zellerbach should s e l l i n the coming year. Without a doubt the forecaster owes a good deal of h i s success i n using t h i s method to h i s own extensive experience i n forecasting plywood sales and to the experience of the men with whom he consults regarding the condition of the plywood market. The company forecaster should continue to employ the judgment technique that he presently uses, but he would follow a better course i f he used one or more s t a t i s t i c a l methods i n addition to h i s present method. A f i n a l forecast could be selected a f t e r an analysis had been made of the forecasted figures developed by the various techniques. BIBLIOGRAPHY I. BOOKS Abrarason, A., and R. Meek, ed., Business Forecasting i n Practice, John Wiley, 1956. American Management Association, Materials and Methods of Sales Forecasting, Special Report no. 27. American Management Association, Sales Forecasting -Uses, Techniques and Trends, Special Report no. 16. Bassie, V., Economic Forecasting, McGraw-Hill, 1958. Boyd, H., and R. Westfall, Marketing Research, R. Irwin 1956. Canada, Royal Commission on Canada's Economic Prospects Ottawa, Queen's Printer, 1958. Chambers, E., Economic Forecasting, Prentice-Hall, 1961 Crawford, C , Sales Forecasting; Methods of Selected Firms, Bureau of Economic and Business Research, University of I l l i n o i s , 1954. Dean, J., Managerial Economics, Prentice-Hall, 1951. Eze k i e l , M., Methods of Correlation Analysis, New York, John Wiley, 1941. Ezekiel, M., and F. Fox, Methods of Correlation and Regression Analysis, J . Wiley, 1959. Ferber, R., S t a t i s t i c a l Techniques i n Marketing Research, McGraw-Hill, 1949. Ferber, R., and P.J. Verdoorn, Research Methods i n Economics and Business, New York, MacMillan, 1962. Firestone, 0., Growth and Future of the Canadian Market, Ottawa, 1956. Freund, J., and F. Williams, Modern Business S t a t i s t i c s Prentice-Hall, 1958. 2. Gordon, R., Business Fluctuations, New York, Harper, 1952. Howard, J., Marketing Management, R. Irwin, 1957. Johnston, J., Econometric Methods, New York, McGraw-H i l l , 1960. \u00E2\u0080\u00A2^Koontz, H., and C. O'Donnell, P r i n c i p l e s of Management, New York, McGraw-Hill, 1959. Kuznets, S., \"Concepts and Assumptions i n Long-Term Projections of National Product,\" i n Studies i n Income and Wealth, V o l . XVI, Long-Range Economic Projections, New York, National Bureau of Economic Research, Princeton University Press, 1954. McLaughlin, R., Time Series Forecasting - A New Computer Technique for Company Sales Forecasting, American Marketing Association. M i l l s , F., S t a t i s t i c a l Methods, New York, Holt, 1938. Moore, G., ed., Business Cycle Indicators, V o l . I, National Bureau of Economic Research, Princeton University Press, 1961. Nemmers, E., Managerial Economics, New York, J . Wiley, 1962. Newbury, F., Business Forecasting, New York, McGraw-H i l l , 1952. Nevmian, W., Administrative Action, New York, Prentice-H a l l , 1951. Parten, M., Surveys, P o l l s and Samples, New York, Harper, 1950. Phelps, D., Sales Management, Chicago, R. Irwin, 1951. Prochnow, H., ed.. Determining the Business Outlook, New York, Harper, 1954. Schultz, H., The Theory and Measurement of Demand, Chicago, University of Chicago Press, 1938. Spencer, M., C. Clark, and P. Hoguet, Business and Economic Forecasting, Homewood, I l l i n o i s , R. Irwin, 1961. 3. Spencer, M. H., and L. Siegelman, Managerial Economics, R. Irwin, 1959. Thompson, C , Forecasting Sales, Studies i n Business Po l i c y no. 25, National I n d u s t r i a l Conference Board, 1947. Tintner, G., Econometrics, New York, John Wiley, 1952. Wold, W., Demand Analysis, New York, John Wiley, 1953. Wright, W., Forecasting f o r P r o f i t , New York, John Wiley, 1947. I I . GOVERNMENT PUBLICATIONS Dominion Bureau of S t a t i s t i c s , Canada Year Book, Queen's Printer, 1955-1963. Dominion Bureau of S t a t i s t i c s , Construction i n Canada, Queen's Printer, 1944-1963. Dominion Bureau of S t a t i s t i c s , Estimates of Families i n Canada, Queen's Printer, 1957-1963. Dominion Bureau of S t a t i s t i c s , Estimates of Households and Families i n Canada, Queen's Printer, 1952-1956. Dominion Bureau of S t a t i s t i c s , Hard Board, Queen's Printer, 1952-1963. Dominion Bureau of S t a t i s t i c s , National Accounts Income and Expenditure, Queen's Printer, 1950-1963. Dominion Bureau of S t a t i s t i c s , Rigid Insulating Board, Queen's Printer, 1952-1963. Dominion Bureau of S t a t i s t i c s , Veneer and Plywood Industry, Queen's Printer, July, 1963. 4 * I I I . REPORTS Crown Zellerbach (Canada) Limited, History of Canadian Western Lumber Company, Limited, Unpublished Report, September, 1957. McConnell, Eastman and Company, Marketing and Promotion Study, Unpublished Report, Vancouver, n.d. Plywood Manufacturers Association of B r i t i s h Columbia, Canadian Douglas F i r Plywood, Unpublished Report, Vancouver, n.d. IV. JOURNALS AND PUBLICATIONS Beckett, A., \"Economic Indicators: How to Use Them i n Business Forecasting,\" Cost and Management, A p r i l , 1960, pp. 123-126. Bund, H., and J . C a r r o l l , \"The Changing Role of the Marketing Function,\" Journal of Marketing, January, 1957, pp. 268-325. Canadian Imperial Bank of Commerce, Commercial Letter, May-June, 1963, pp. 1-12. Canadian Imperial Bank of Commerce, Commercial Letter, May-June, 1964, pp. 1-12. Dhalla, N., \"How to Develop a Sales Forecast,\" Canadian\u00E2\u0080\u00A2Business, September, 1961, pp. 48-52. E d s a l l , R., \"How to Forecast Your Sales;\" \"Forecasting Business\u00E2\u0080\u00A2Conditions i n General;\" \"Forecasting Industry and Company Sales,\" I n d u s t r i a l Canada, January, February, March, 1962. Ferber, R., \"Sales Forecasting by Sample Surveys,\" Journal of Marketing, July, 1955, pp. 1-12. Knowles, J . W., \"Relation of Structure and Assumptions to Purpose i n Making Economic Projections,\" Proceedings of the 116th Annual Meeting of the American S t a t i s t i c a l Association, 1956, pp. 7-8. Livingston, J., \"How Wrong Can Economists Be?\" The Reporter, May 26, 1953, pp. 17-20. Oxenfeldt, A., \"How to Use Market-Share Measurement, Harvard Business Review, January-February, 1959, pp. 59-68. Paranka, S., \"Marketing Predictions from Consumer A t t i t u d i n a l Data,\" Journal of Marketing, July, 1960, pp. 46-51. . Schweiger, J., \"Forecasting Short-Term Consumer Demand from Consumer Expectations,\" Journal of Business, A p r i l , 1956, pp. 90-100.-Tosdal, H., \"Bases f o r the Study of Consumer Demand, Journal of Marketing, July, 1939, pp. 3-15. V. NEWSPAPERS \"Guidelines to '64,\" The F i n a n c i a l Post, Toronto, November 1, 1963, p. 17. \"$1.5 m i l l i o n extension despite tax,\" The Province, Vancouver, June 29, 1963, p. 15. "@en . "Thesis/Dissertation"@en . "10.14288/1.0102394"@en . "eng"@en . "Business Administration - Marketing"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "Sales forecasting in the plywood industry"@en . "Text"@en . "http://hdl.handle.net/2429/37604"@en .