UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Product life cycle and determinant attributes : an investigation of the Japanese structural panel market Welbourn, Derek George 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2001-0290.pdf [ 10.59MB ]
Metadata
JSON: 831-1.0075351.json
JSON-LD: 831-1.0075351-ld.json
RDF/XML (Pretty): 831-1.0075351-rdf.xml
RDF/JSON: 831-1.0075351-rdf.json
Turtle: 831-1.0075351-turtle.txt
N-Triples: 831-1.0075351-rdf-ntriples.txt
Original Record: 831-1.0075351-source.json
Full Text
831-1.0075351-fulltext.txt
Citation
831-1.0075351.ris

Full Text

P R O D U C T LIFE C Y C L E A N D D E T E R M I N A N T A T T R I B U T E S -A N INVESTIGATION O F T H E J A P A N E S E S T R U C T U R A L P A N E L M A R K E T by D E R E K G E O R G E W E L B O U R N Bachelor of Commerce, The University of British Columbia, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Faculty of Forestry Department of Wood Science W e accept this thes is as conforming to the required s tandard The University of British Columbia April 2001 © Derek George Welbourn, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Derek Welbourn Department of Wood Science The University of British Columbia Vancouver, Canada Date ABSTRACT Japan is the largest structural panel importer in the world and currently consumes predominantly plywood. Oriented strand board (OSB) has been available in Japan since the early 1990's but has made only modest inroads. For example OSB represented 2.5 percent of the Japanese structural panel market in 1999 and has averaged approximately 2 percent for the previous 5 years. By utilizing the product life cycle concept to compare the development of the Japanese structural panel market to that of North America, there is evidence to suggest that OSB could be positioned for significant growth in Japan. However, whether OSB becomes a dominant structural panel in the Japanese market will depend on how well it can meet the critical needs of panel customers. In order to assess which characteristics of a structural panel have a significant effect on Japanese customers', a determinant attribute analysis was done. Questionnaires were filled out during personal interviews with Japanese structural panel purchasing companies that included trading houses, distributors and builders. The determinant attribute analysis identified the critical characteristics that Japanese panel customers consider when making purchasing decisions as being (in order of significance): Competitive Price, Thickness Swell, Panel Size, Shipment Arrives In Good Condition, Long-term Supply, Formaldehyde Emissions, Environmental Forestry Practices, Uniform Panel Thickness and Linear Expansion. The determinant attribute analysis suggests that OSB is not ready to immediately enter into the growth phase of the product life cycle in Japan. OSB needs to improve its performance on the technical determinant attributes and maintain competitive pricing relative to competing panel products. The relevance of these results could be substantially altered if the demand for structural panels in Japan was to change significantly, which could be the case with the trend towards structurally improved building methods in Japanese traditional homes. ii TABLE OF CONTENTS Abstract " Table of Contents Hi List of Tables vii List of Figures ______ ix Acknowledgements x 1 Introduction 1 1.1 Research Objectives 2 2 Literature Review 3 2.1 Product Life Cycle Theory 3 2.1.1 Development Stage 4 2.1.2 Growth Stage : 4 2.1.3 Maturity Stage . 5 2.1.4 Decline Stage . . 5 2.1.5 Product Li fe Cycle A s A Management T o o l 5 2.1.5.1 Product L i f e Cycle Limitations 6 2.2 Product Life Cycle And The North American Panel Market 6 2.2.1 Product Li fe Cycle Analysis . 7 2.2.2 Notable Trend 12 2.3 Product Life Cycle And the Japanese Panel Market \ 12 2.3.1 Japanese Structural Panel Demand 13 2.3.2 Product L i fe Cycle Analysis 14 2.4 Research Methodology 18 2.4.1 Customer Needs A n d Marketing Strategy • 19 2.4.2 Determinant Attribute Theory 20 2.4.2.1 Approach Descriptions 20 2.4.2.2 The Theory o f Determinant Attributes A n d Direct Dual Questioning 21 2.4.2.3 Direct Dual Questioning A n d Attribute Classification 22 iii 2.4.2.4 Direct Dual Questioning - Attribute Analysis and Results 23 2.4.2.4.1 Attribute Selection [ 24 2.4.2.4.2 Attribute Measurement 25 2.4.2.4.3 Determinance Score 25 2.4.2.4.4 Response Bias . 26 2.4.2.4.5 Identification 27 2.4.2.4.6 Direct Dual Questioning - Respondent Analysis And Results 28 2.4.2.5 Attitudes Versus Reality 29 3 Japanese Market Research Method 30 3.1 Population And Sample Frame 30 3.1.1 Sampling Method . 30 3.1.2 Sample Size 31 3.2 Sampling Procedure And Questionnaire Development 33 3.2.1 Phase-One Questionnaire 34 3.2.1.1 Attribute Selection . . 34 3.2.2 Phase-Two Questionnaire 36 3.2.2.1 Importance 36 3.2.2.2 Difference 36 3.3 Results Analysis 36 4 Research Results 38 4.1 Demographic And Descriptive Statistics 38 4.1.1 Respondent Categories And Size . 38 4.1.2 Panel Type And Usage 38 4.1.3 Panel Quality Difference . 42 4.1.4 Home Builders__ . 42 4.2 Determinant Attribute Identification 43 4.2.1 Attribute Selection 43 4.2.2 Attribute Frequency . 44 4.2.3 Importance 47 4.2.4 Difference . 47 4.2.5 Determinance 48 4.2.6 Response Bias 49 4.2.7 Determinant Attributes - All Attributes . . 51 4.2.7.1 Analysis of Variance Assumptions . 51 iv 4.2.7.2 Analysis of Variance 52 4.2.7.3 Attribute Identification 53 4.2.7.4 Further Attribute Identification 54 4.2.8 Determinant Attributes - B y Attribute Category 55 4.2.8.1 Technical Attribute Determinance 55 4.2.8.2 Marketing Attribute Determinance 56 4.2.8.3 Service Attribute Determinance 57 4.2.8.4 Attribute Category Importance _ _ _ _ _ 57 4.2.9 Determinant Attributes - B y Respondent Category 58 4.2.9.1 Distributors 58 4.2.9.2 Builders . 60 4.2.9.3 Trading Houses 61 4.2.9.4 Distributors A n d Trading Houses 63 5 Disucussion 65 5.1 Determinant Attribute Comparison by Respondent Category 65 5.1.1 Competitive Price A n d Thickness Swell 66 5.1.1.1 Competitive Price 66 5.1.1.1.1 Changing From Plywood T o O S B . 67 5.1.1.1.2 Product Differentiation 67 5.1.1.1.3 Exchange Rates 68 5.1.1.2 Thickness Swell . 68 5.1.2 Panel Size . 69 5.1.3 Shipment Arrives In G o o d Condition 70 5.1.4 Long-term Supply 72 5.1.5 Formaldehyde Emissions 72 5.1.6 Environmental Forestry Practices 73 5.1.7 Other Determinant Attributes 73 5.2 Determinant Attribute Comparison By Attribute Category 73 5.3 Research Limitations And Suggestions 74 5.3.1 Judgement Sample 74 ' 5.3.2 Sample Size 75 6 Conclusion 76 7 Literature Cited 78 Appendix I: Interview Introduction Letter 82 v Appendix II: Phase-one Questionnaire 83 Appendix III: Phase-two Questionnaire 89 Appendix IV: Phase-one Results (Raw Data) 96 Appendix V: Anova - Technical Attributes 100 Appendix VI: Anova - Marketing Attributes 101 Appendix VII: Anova - Service Attributes 102 Appendix VIII: Anova — Distributor Attributes 103 Appendix IX: Anova - Builder Attributes 104 Appendix X: Anova - Trading House Attributes 105 Appendix XI: Anova - Distributors and Trading House Attributes 106 vi LIST OF TABLES Table Page 1. Possible scores and attribute classification for direct dual questioning approach 23 2. List of respondents interviewed 32 3. Structural panel attribute categories and type 35 4. Attribute selection process for the technical attribute category during the phase-one questionnaire 35 5. Japanese construction usage of structural panels 39 6. Structural panel purchased or considered by respondents 39 7. Respondent primary panel thicknesses ; 41 8. Respondent primary panel dimensions 41 9. Builders survey, average home sizes 43 10. Technical (objective) attributes 43 11. Marketing (subjective) attributes 44 12. Service (subjective) attributes 44 13. Additional technical (objective) attributes identified by the respondents during the phase-one questionnaire 44 14. Mean importance scores for technical, marketing and service attributes for structural panels in Japan 47 15. Mean difference scores for technical, marketing and service attributes for structural panels in Japan 48 16. Mean determinance scores for technical, marketing and service attributes for structural panels in Japan 48 17. Row centered mean determinance scores for technical, marketing and service attributes for structural panels in Japan 50 18: Single factor summary and ANOVA table for the 14 different attributes 52 19. Homogeneous attribute groups and determinant attributes 53 20. Attributes by category 55 21. Homogenous attribute groups, technical attribute category 56 22. Homogenous attribute groups, marketing attribute category 56 23. Homogenous attributes, service attribute category 57 24. Category importance ; 58 25. Homogenous attribute groups and determinant attributes for distributors 60 26. Homogenous attribute groups and determinant attributes for builders 61 27. Homogenous attribute groups and determinant attributes for trading houses 63 28. Homogenous attribute groups and determinant attributes for distributors and trading houses 64 29. Determinant attribute results, comparison by respondent category 65 vii 30. Breakdown of in transit panel damage 71 31. Determinant attribute results, comparison by attribute category 74 32. Example of plywood and OSB price differentials 76 33. OSB's estimated performance relative to other panel products 77 viii LIST OF FIGURES Figure Page 1. Stages of the product life cycle 3 2. North American structural panel production capacity by product 7 3. Western plywood product life cycle 8 4. Southern pine plywood product life cycle 10 5. OSB product life cycle 11 6. The Japanese structural panel product life cycle 13 7. Domestic hardwood plywood product life cycle 15 8. Imported hardwood plywood product life cycle . 16 9. Domestic softwood plywood, OSB and Canadian softwood plywood life cycle 17 10. Products viewed as a group of attributes (i.e., birthday cake) 20 11. Generic direct dual determinant attribute approach 24 12. Possible effects of a missed attribute 24 13. Alternative approaches to identifying determinant attributes from direct dual questioning _ 27 14. Basic research design 30 15. Two-phase sampling process 33 16. Respondents surveyed by category 38 17. Total respondent panel volumes by category 38 18. Respondent import panel consumption by country 40 19. Japanese panel imports by country 40 20. Respondent panel consumption by end-use 41 21. Builders surveyed by home type 42 22. Attributes and respondent selection frequency 45 23. Attributes with determinant frequencies 46 24. Row centered mean determinance scores for technical, marketing and service attributes for structural panels in Japan 50 25. Determinant attributes 54 26. Determinant frequency selection for distributors 59 27. Determinant frequency selection for builders 60 28. Determinant frequency selection for trading houses 62 29. Determinant frequency selection for distributors and trading houses 63 30. Imported hardwood plywood packaging 71 31. Bankyo floor, N21 OSB brochure 72 ix ACKNOWLEDGEMENTS The author thanks the following individuals: • Michael Ainsworth of Ainsworth Lumber Co. Ltd. for creating the opportunity to conduct this research, • Dr. David Cohen of University of British Columbia for his support and guidance during the research design, field work and thesis writing, • Dr. Robert Kozak of University of British Columbia for his guidance and statistical knowledge, • Glen Wilson of Interex Forest Products for his assistance with setting up the field research in Japan, • Seji Omote of Interex Forest Products (Japan) for his time and support during the Japanese field research, • Yoshi Fujihira of Interex Forest Products (Japan) for his time and support during the field research, • Hiroshi Yokoyama of Canfor Woodsales Co. Ltd. for his time and support during the field research, and • Karen Bothwell for her assistance with literature research and overall support. The author also thanks the following organizations for financial and technical support: • Ainsworth Lumber Co. Ltd., • Interex Forest Products, • Canfor Woodsales Co. Ltd., and • University of British Columbia. x 1 INTRODUCTION The North American structural panel market is the largest and most developed in the world. For the past decade, the structural panel product, oriented strand board (OSB) has grown at a rapid pace displacing competing North American1 products. Since 1996, OSB has held a larger market share than any other panel product and continues to expand (RISI 2001). Despite rapid growth in North America some innovative producers have targeted Japan as a potential new market for OSB. By studying the North American market, insight can be gained into the future of the less developed Japanese market. Japan is the largest panel importer and is the second largest panel consuming market in the world2 (FAO 2001). However, to date, OSB's inroads to the Japanese market have been modest (Interex 2000). As such, if OSB were to have the same growth patterns as seen in North America, Japan would represent a substantial opportunity for OSB manufacturers. OSB's future success in Japan is dependent on whether it can more effectively meet the Japanese customer's needs relative to competing panel products. The product life cycle (PLC) is a strategic concept that can be used for understanding the evolution of a product and market. With the utilization of PLC analysis, the North American and Japanese panel markets can be compared, shedding light on the future of OSB in Japan. The PLC comparison between these two markets indicates that OSB in Japan is at a critical point and is either poised to become a dominant high volume product or a smaller niche player. The product life cycle concept and corresponding analysis of the North American and Japanese markets are completed in Chapter 2, Sections 1 to 3. The future of OSB in Japan primarily depends on how well its attributes fit with the needs of Japanese panel customers. Market research can be carried out to identify which customer needs are critical when making a structural panel purchasing decision. Determinant attribute analysis is a proven research technique for identifying which of a product's attributes has a direct influence on customers' purchasing behavior. A detailed description of determinant attribute analysis is provided in Chapter 2, Section 4. 11ncludes: United States and Canada. 2 Japan and China trade places as being the second and third largest structural panel markets depending on the year and source. 1 In 1998, a determinant attribute analysis was completed on the Japanese panel market and the critical attributes effecting purchasing decisions were identified. The specific research and analysis technique is explained in Chapter 3, while the results of the market research are presented in Chapter 4. If OSB can meet the Japanese panel customers' needs that have been identified by the determinant attribute analysis, it has the potential to become a dominant structural panel. Unfortunately, projecting a product's performance on critical attributes relative to competing products is not an easy task nor is it guaranteed to be accurate as there are unpredictable outside forces that have a direct effect on a product's performance. For example, competitors can improve their product's performance on particular attributes. The most determinant attributes and the issues surrounding them are discussed in Chapter 5. Finally, concluding remarks on the structural panel market and the estimated relative performance of OSB is provided in Chapter 6. 1.1 R E S E A R C H O B J E C T I V E S The research undertaken in Japan was to gain an understanding of the needs of Japanese structural panel customers. The study was designed to profile the Japanese respondents and to identify the attributes of a structural panel that have a significant effect on their purchasing decisions. With this knowledge, structural panel producers will be able to focus on the issues that will best lead to market penetration and profitability. Specifically, the research objectives for the Japanese fieldwork were to: • gather information about Japanese structural panel customers providing a profile of the different market segments, • identify attributes that have a direct influence on purchasing decisions, • create a market research technique that can be replicated in the future. 2 2 LITERATURE REVIEW The product life cycle (PLC) is an established marketing concept that can be a useful tool for understanding market evolution and developing strategic marketing plans. The North American structural panel market is highly evolved and consists of three major product groups, all of which are at different stages in the PLC. Japan, while an established market for structural panels, is clearly behind North America in market evolution. Whether the Japanese market will follow the same evolution patterns as North America will depend on how well the new products entering the Japanese market fit customer needs (Levitt 1965). Understanding customer needs is at the forefront of any successful marketing strategy and the long-term success of structural panel products, to a large extent, will depend upon how well the product is aligned with customer demands (Thomas and Waterman 1982, Seward and Sinclair 1988). Determinant attribute analysis has proven itself as a successful method for identifying customer needs that are crucial to purchasing decisions (Armacost and Hosseini 1994, Sinclair and Stalling 1990 and Lumpkin et al. 1985). 2.1 PRODUCT LIFE CYCLE THEORY The PLC is a concept which advocates that most successful products live their lives through a distinguishable pattern of 4 recognizable stages, development, growth, maturity and decline, see Figure 1. Figure 1. Stages of the product life cycle. Source: (Adapted from Beckman et al. 1982 and Porter 1980) 3 2.1.1 Development Stage The development stage begins by bringing a new product to market. At this time there is typically no established customer base for the newly released product and the goal of the firm is to stimulate demand. The producer is focused on introducing the new product and creating demand with a promotional campaign. The campaign is normally focused on promoting the product in general and not necessarily on a product or brand specific to the firm. "Effective promotion is important as the world does not automatically beat a path to the person with a better mousetrap. The world has to be told, coddled, enticed and even romanced." (Levitt 1965:5). Most importantly, the product must offer superior attributes in order to become a viable long-term option over other existing products. During the development stage sales increase gradually while the market becomes familiar with the new product. Often producers who first release new products must endure negative profits as the costs of product development, new manufacturing facilities and marketing activities are not covered by the low sales volumes. The dashed line in Figure 1 illustrates negative profits during the development stage. Generally, the greater the required shift in the customer's usual way of doing things, the longer the development stage will last. According to Levitt (1965) how long the development stage will last depends on the new product's complexity, degree of newness, the presence of competitive substitutes and its fit into customer needs. Arguably, customer needs is the most important criteria for a new product to perform well on. The closer a product meets a purchaser's requirements, relative to other products, the less competitive substitute products will be. Typically, if a product efficiently meets the needs of the customers over competing products, it will eventually overcome any complexity or unfamiliarity. "It has been demonstrated time after time that properly customer-oriented products are one of the primary conditions of sales and profit growth" (Levitt 1965:3). As the development stage progresses, the introductory firm should welcome the addition of competitors. With additional suppliers to choose from the customers will feel more comfortable adopting the new product. Generally, the more firms participating in promotional activities, the quicker the product will work through the development stage and enter the profitable growth stage. 2.1.2 Growth Stage Once the market has become familiar with the new product and its characteristics, demand begins to accelerate and the size of the total market expands rapidly (Levitt 1965). The growth stage is the most profitable of the product life cycle stages and is where producers see significant returns on their 4 investments. With the dramatic increases in sales and profits the product becomes highly attractive and many new competitors begin to enter the market. Some of the new competitors offer the same product, while others make improvements. With the increase in competition, profit margins start to fall but the increase in sales volumes more than make up for any reductions in price. Individual producers begin to switch marketing strategies changing from one of generic product promotion to one of promoting their particular product. 2.1.3 Maturity Stage At the beginning of the maturity stage the industry sales volumes continue to escalate, although at a reduced rate relative to the growth stage. At this point many new producers have entered the market raising capacity and putting continued downward pressure on price. As the maturity stage progresses, there is evidence of market saturation and competition on price becomes intense, industry profits drop and eventually sales revenues begin to decline (Beckman et al. 1982). The maturity stage generally lasts as long as there are no changes in primary demand or entrance of competitive substitutes. Market competitiveness reaches the highest level yet. Producers are forced to compete heavily on price and on subtle product differences, often switching to a product differentiation strategy (Levitt 1965). Depending on the product, competing on service may be the most effective strategy. 2.1.4 Decline Stage In the decline stage, the product begins to lose customer appeal and sales start to consistently decrease. The over-capacity that was affecting prices and levels of competition in the maturity stage now becomes catastrophic. Some producers are forced from the market due to poor economic performance, while others attempt to improve efficiency and their competitiveness. Regardless of the tactic, over capacity and the fight for survival drives industry profits lower and lower. Typically, the decline stage of a product coincides with the growth stage of a superior substitute. 2.1.5 Product Life Cycle As A Management Tool The PLC concept is established in marketing literature and has been proven with numerous examples. By utilizing this concept it forces producers to analyze where their products are in the life cycle and which marketing strategies should be considered. When done correctly it can lead the producer to the optimal strategy for launching a new product or for keeping an existing product in the most profitable growth stage 5 for as long as possible (Beckman et al. 1982). With proper consideration, PLC analysis can help the producer predict if and when they should consider manufacturing new products and diversifying. By creating new uses or finding new markets the product life cycle can literally start over at the development and/or growth stage. This strategy is known as product stretching or an extension strategy because it can literally extend the growth stage of the product life cycle (Levitt 1965 and Beckman et al. 1982). 2.1.5.1 Product Life Cycle Limitations With the suggestion of utilizing the product life cycle for planning purposes, its limitations should be considered. PLC, like any widely held concept, has attracted some criticism. Some of the strongest negative views have challenged the heart of the idea and suggest that employing it as a management tool can and has lead to costly mistakes (Dhalla and Yuspeh 1976). However, the criticisms are based on the fact that many product evolutions do not follow the pattern suggested by the product life cycle. Critics point out that products can start off with explosive growth, never reach maturity and then plunge into obscurity with their development and decline stages barely perceptible (Dhalla and Yuspeh 1976). A classic example of a product like this would be a fad in the fashion industry. The PLC concept is like any tool or piece of advice; it must be applied to the specific situation. PLC analysis is only as good as the thought that has been put into understanding the forces effecting the product. One of the chief criticisms of the PLC concept is that it suggests there is a foregone outcome for each product, when in fact producers can have a significant affect on the entire industry product curve (Dhalla and Yuspeh 1976). However, it is precisely this fact that makes PLC a useful management concept. By recognizing where the product is in the life cycle, a producer can use the optimal strategy to position their particular product and even affect the industry curve to their benefit. The final important aspect of the PLC concept is to recognize that the overall industry curve for the product life cycle is not necessarily the same curve as for a particular producer. The product life cycle curve is the cumulative sum of all producers selling that product. There are many examples where firms perform well in a mature or declining market and equally as many examples of producers that can not operate profitably during a market growth stage. 2.2 P R O D U C T LIFE C Y C L E AND THE NORTH AMERICAN PANEL M A R K E T The evolution of the North American structural panel market is an excellent example of products that have followed and are continuing to follow a classic product life cycle. In fact, the product life cycle concept has 6 been used in various studies to describe the progression of the North American structural panel market (Shook 1999, Shook et al. 1998 and Sinclair 1991). North America is the largest structural panel market in the world, consuming over 70 percent of the world's softwood plywood and over 95 percent of OSB (FAO 2001, RISI 2001 and APA 2000). Since its inception, the North American structural panel market has continuously expanded with an average annual growth rate of 4 percent and annual volume of 40 million cubic meters (RISI 2001). However, the continually expanding market hides the dramatic expansions and contractions experienced in the underlying products that service this market (Sinclair 1991). Displayed in Figure 2 are the production capacities of North American structural panel producers for the past 75 years. Figure 2. North American structural panel production capacity by product. North A m e r i c a n St ructura l P a n e l C a p a c i t y 45,000 1 1 40,000 - ^ O S B —*- Southern Pine Plywood - m - Western Plywood Total Source: Adapted from Spelter 1994 and RISI 2001. 2.2.1 Product Life Cyc le Ana lys is Western plywood, the first panel product to be used for structural applications, was first produced in 1905 and went through a long development period that lasted approximately 40 years. During the lengthy product 7 development stage, the manufacturing process was refined and producers explored the different potential end uses. During the 1930's and early 1940's western plywood was primary used for industrial applications such as draw bottoms and door panels. In 1933 the Douglas Fir Plywood Association was chartered and began to establish uniform grading rules which helped create a higher quality, standardized product (Cour 1955). Through the World War II period, plywood expanded into the military applications such as landing craft construction, ammunition boxes and field tables (Cour 1955). By the early 1940's, plywood use was starting to be dominated by structural applications in the residential construction market. The consumption of plywood started to expand with the growing demand for housing and by 1945 the total market was estimated to consume over 2 million cubic meters annually (Shook 1999 and Cour 1955). Western plywood had proven itself as a much more efficient and cost effective system for home construction, relative to sheathing a home with boards which was the standard construction method at the time. By 1945, it was well established and demand began to expand rapidly up until 1965 increasing from 2 million cubic meters to over 12 (Spelter et al. 1994 and RISI 2001). During this period western plywood had entered the growth stage of its product life cycle and experienced an average annual growth rate of approximately 14 percent (Shook et al. 1998). Figure 3 shows the product life cycle that western plywood has experienced. Figure 3. Western plywood product life cycle. Source: Adapted from Spelter 1994 and RISI 2001. 8 Western plywood's success from 1945 to 1965 created a strong attraction for competing products. At the end of 1965 western plywood's capacity additions began to slow down and by 1972 started to decline. Synonymous with western plywood's deceleration in growth, two new substitutes entered the market, southern pine plywood and OSB/waferboard3, refer back to Figure 2. In 1966, facing rapidly expanding southern pine plywood capacity, western plywood entered the maturity stage with slowing sales and stagnant market share. By 1988, enduring vigorous competition from both southern plywood and OSB as well as increasing production costs due to a reduction in fiber supply, western plywood entered the decline stage. In 1988, there were 102 operating western plywood plants a significant decrease from the 1965 peak of 184, and today this decline has not yet stabilized. Total capacity in 2000 was estimated to be 10 million cubic meters down considerably from its 1973 peak of 14 million (Spelter et al. 1997 and RISI 2001). Southern pine plywood's development phase was almost non-existent, illustrated in Figure 4. As Levitt (1965) described, a new product that offers some desirable benefits and very little newness will have a short development stage and grow rapidly. Southern pine plywood benefited from offering a similar product to western plywood at a more competitive price. Furthermore, the products were virtually identical in the eyes of the customer because of a change in building codes that allowed the two different panels to be used interchangeably (Fassnacht 1964). The market was not expecting such a dramatic expansion in southern pine plywood capacity, but it is an excellent example of how rapidly a product can expand when it is matched with customer desires. Fassnacht (1964:22), a supporter of the development of southern pine, stated that, "the talk about dozens of new plywood plants in the South during the next decade does not seem justified". Despite the skepticism, within 8 years of the first plant's startup there were 55 operating southern pine plywood mills and an estimated capacity of 6 million cubic meters (Spelter et al. 1997). 3 For analysis purposes Waferboard, the predecessor of OSB, is deemed as an early stage OSB product and is thus considered as one product group, OSB. 9 Figure 4. Southern pine plywood product life cycle. Southern Pine Plywood Capacity Source: Adapted from Spelter 1994 and RISI 2001. Southern pine plywood continued its growth stage until OSB had worked through its development phase and began to be viewed as a viable plywood substitute. For OSB, the development stage was much slower than that experienced by southern pine. OSB entered the market at approximately the same time as southern pine plywood, but didn't start to experience any significant market growth until the early 1980's. Figure 5 illustrates the product life cycle that has been experienced by OSB. Despite a significant cost advantage over both western and southern pine plywood, OSB did not immediately become the product of choice. OSB had to overcome inferior product performance, a dramatically different physical appearance and an exclusion from existing building codes. 10 Figure 5. OSB product life cycle. Source: Adapted from Spelter 1994 and RISI 2001. During the development stage, OSB's structural properties improved with advancement in processing technology and by the early 1980's its performance was equivalent to plywood in most applications. One of the key events that shifted OSB from the development stage to the growth stage was a change in the product standards from specification to performance based standards. Previously the product standards were based on the plywood manufacturing process and specified the exact makeup of a panel, including resin type, veneer grade and veneer thickness. However, instead of being based on product manufacturing specifications, the new standards were based on structural performance criteria (Sinclair 1991). It no longer mattered how a panel was produced, but rather how it performed. OSB could meet these new structural standards and thus could now be used interchangeably with plywood. In the early 1980's OSB entered its growth stage and has since consistently increased in volume at an estimated rate of 14.9 percent per annum, superceding the growth experienced by southern pine plywood (RISI 2001). By 1996 OSB's annual production capacity had surpassed all other structural panel products at 15 million cubic meters (Spelter et al. 1997 and RISI 2001). OSB had a group of attributes that were desirable to structural panel customers and correspondingly the demand for plywood began to decrease. Furthermore, the growth of OSB also benefited from a significant increase in total market demand, which enabled OSB consumption to rise without directly displacing the same proportion of plywood. In 2001, OSB was estimated to have a market share of 53 percent and is expected to continue to expand to 67 percent by 2005 (RISI 2001). Additionally, there are no other competing products that appear to have the potential to push OSB into the next phase of the product life cycle. 11 2.2.2 Notable Trend As can be seen in the previous analysis, there are three distinct products that supply the North American structural panel market and each of them has experienced periods of rapid growth and prosperity. When reviewing these products' life cycles the fundamental concept that has brought success to one product over another has been equivalent structural performance for a more competitive price (i.e. attribute of competitive price). 2.3 P R O D U C T LIFE C Y C L E AND THE J A P A N E S E PANEL M A R K E T Japan is a diverse market. With limited domestic fiber supply, Japan imports from almost every forest products exporting country in the world. Japan's large demand for panel products is met with imports from many different countries including Indonesia, Malaysia, New Zealand, South Africa, United States and Canada (APA 2000). For analysis purposes, only panel products that have been consumed in significant4 volumes have been included in the following life cycle analysis. As such, the following analysis focuses only on the main trends seen in the Japanese panel market. In the late 1960's, as was the case in North America throughout the 1940's, the use of structural panels in Japan began to accelerate as they became an established component in residential construction. Refer to Figure 6 for an overview of the development of the Japanese structural panel market. 4 Only panel products that have been consumed in quantities greater than 200,000 cubic meters per annum. 12 Figure 6. The Japanese structural panel product life cycle analysis. Japanese Structural Panel Capacity 12,000 -i —•— Domestic Hardwood Plywood -G— Domestic Softwood Plywood —x— Imported Hardwood Plywood -_— Imported OSB Total Structural Panel Market —•— Canada Softwood Plywood Source: Adapted from APA 2000, FAO 2001, Interex 2000 and Statistics Canada 2001. 2.3.1 Japanese Structural Panel Demand Japan is one of the largest structural panel consuming countries in the world, second only to North America and roughly equivalent in size of China (FAO 2001). In 1999, it was estimated that Japan consumed 8 million cubic meters of structural panels (APA 2000). Interestingly, North America and Japan are the only two regions in the world that have a cultural preference for wooden housing. Unlike North America, total Japanese demand for structural panels has not continuously expanded since inception and has been rather stagnant since the late 1960's, refer to Figure 6 and back to Figure 2. This is a noteworthy point as a fundamental change in overall demand can significantly alter a product's life cycle (Levitt 1965). The relatively stagnant growth of structural panels in the construction of Japanese traditional homes, the largest wooden housing segment in Japan, could increase due to changes in the demand for 13 housing performance. In Japan the construction methods have focused on wooden cross bracing as opposed to structural panels, even though structural panels have proven to increase the performance of housing in structural stability and overall performance efficiency. With the advent of 2x4 and hybrid housing and events such as the Kobe earthquake, housing systems based on structural panels is starting to come to light. Currently, Japan is in the middle of a significant change in the regulations that govern residential construction (Cohen and Gaston 2001). Although it is unclear what effects these code changes will have, their focus is predominantly on improving home stability and longevity. "The revised building standards law should accelerate the focus on structural stability and increase the use of hybrid homes" (mix of traditional post & beam and 2x4 construction) (Cohen and Gaston 2001 :iii). In addition, it appears that the codes are progressing towards performance based standards as opposed to specification, a switch that dramatically assisted OSB and other engineered wood products to expand in North America. All of these changes bode well for a potential increase in the use of structural sheathing in residential construction. If a substantial increase in Japanese structural panel demand were to occur, it would assist with the development of substitute panel products. With the creation of new demand these substitute products could claim market share without the difficulty of displacing existing panel products. As well, any additional sales could assist in moving the product from the development stage to the growth stage of the product life cycle. 2.3.2 Product Life Cyc le Ana lys is The Japanese structural panel market was founded on hardwood plywood produced with imported logs, predominantly from Indonesia and Malaysia. The life cycle of domestically produced hardwood plywood in Japan parallels that of North America's western plywood. Domestic hardwood plywood expanded rapidly, along with total demand, until the early 1970's. Domestic hardwood plywood consumption peaked in 1974 at 8.6 million cubic meters (APA 2000). Although, some inconsequential amounts of plywood were being imported from mainly, Indonesia, Malaysia and others; At this point, domestically produced plywood was the dominant product with close to 100 percent market share, refer back to Figure 6. The consumption of domestically produced plywood stayed relatively unchanged, indicating that it had reached the maturity stage of the product life cycle and remained there from 1973 to 1988. Japanese domestically produced plywood using south seas hardwood logs had a lengthy maturity stage as during this time period there were no other competitive panel products in which to compete with, refer to Figure 7. Indonesia and Malaysia, although rich with fiber, had not made the integration into plywood manufacturing. 14 In the late 1980's, domestically produced hardwood plywood started to come under increasing competitive pressure and entered the decline stage of the product life cycle, as can be seen in Figure 7. This was a direct result of Indonesia, the dominant Japanese log supplier, implementing policies that encouraged the development of their domestic forest industry. In 1980, the Indonesian government implemented policies to phase out log exports by 1985 and instructed all timber rights holders to develop down stream plywood mills (Elias 2000). The result was a rapid increase in Indonesian plywood capacity and plywood exports from 1980 to 1993. During this period Indonesian plywood exports were reported have increased at a rate of 36 percent per annum (APKINDO 1991). Figure 7. Domestic hardwood plywood product life cycle. Source: Adapted from APA 2000, FAO 2001 and Interex 2000. Once plywood was available from Indonesia, the demand for imported hardwood plywood in Japan increased rapidly, see Figure 8. Imported and domestic plywood, given that they were produced with the same fiber, were very similar products. The barriers that imported plywood had to overcome were a Japanese cultural preference for domestically made products and a slightly lower quality of Indonesian made plywood due to less experienced producers (Elias 2000). An additional benefit that was provided by the imported plywood was lower cost due to the equivalent Japanese producers having to endure Japan's high cost of land and labor (Elias 2000). Additionally, as the Indonesian producers continued manufacturing, their product quality began to meet the Japanese standards. As a result, imported hardwood plywood began to increase its share of the Japanese structural 15 panel market and pushed it into the growth stage during the later 1980's with an estimated annual increase of 14 percent. From 1985 onwards, the growth in Japan of imported hardwood plywood was very similar to what was experienced in North America by southern pine plywood, i.e. the characteristics of the two plywood products were the same with a near identical product being substituted for a higher cost product. Refer to Figure 8 for illustration of the product life cycle of Japanese imported hardwood plywood and refer back to Figure 4 for the product life cycle of North American southern pine plywood. Figure 8. Imported hardwood plywood product life cycle. Imported Hardwood Plywood Consumption 10,000.00 9,000.00 8,000.00 IT 7,000.00 b o o 6,000.00 12 5,000.00 » S u 4,000.00 !5 3 3,000.00 u 2,000.00 1,000.00 Matur ity i i . Growth Development / K ,„„,,,..,, y\J ».~ *-* dpN d?? dp3 d£ dp4 <£N d? d^ d^ dpN dS? d^ d£ d?s d^1 <# Source: Adapted from APA 2000, FAO 2001 and Interex 2000. In 1997 imported hardwood plywood had reach a peak of 7 million cubic meters (APA 2000). It appears that imported hardwood plywood has reached the maturity stage as demand has been fluctuating without consistent growth or decline since 1997. However, in the near future, there does not appear to be any established competing products that will push imported hardwood plywood into decline. There are three relatively new products that have entered the Japanese structural panel market, domestic softwood plywood, Canadian softwood plywood and OSB. However, none of these products have claimed much of the overall market. Total annual consumption of each of these products has all been below 1 million cubic meters or less than 8 percent of the total market (APA 2000 and Interex 2000). Currently it is unclear as to where they are in their product life cycles or what their futures will be, refer to Figure 9 and back to Figure 6. 16 Out of the three products, domestic softwood plywood is the latest to enter the market and has experienced the most growth, refer to Figure 9. It is possible that domestic softwood plywood will replicate the swift expansion seen by imported hardwood plywood in Japan and southern pine in North America. According to PLC theory, domestic softwood plywood would have the potential to become a dominant product in the Japanese panel market as there is very little product "newness" for customers. Domestic softwood plywood is highly similar to the other panel products that Japan is already familiar with, being a standard plywood product, and it is even being offered by many of the same Japanese producers that previously offered domestic hardwood plywood. It is thought that imported softwood plywood's success will depend on its price competitiveness with other competing panel products. Japanese softwood plywood is manufactured from imported Russian larch logs which have, to date, been relatively price competitive. Based on the success displayed by southern yellow pine in North America and pending fiber availability, which looks secure, high growth is expected from domestic softwood plywood production. Figure 9. Domestic softwood plywood, OSB and Canadian softwood plywood product life cycles. Softwood Plywood and OSB Consumpt ion 1,000 -<>- Domestic Softwood Plywood Imported OSB -•-Canadian Softwood Plywood Source: Adapted from APA 2000, FAO 2001, Interex 2000 and Statistics Canada 2001. North American OSB and Canadian softwood plywood producers have been shipping modest volumes to Japan since the early 1990's. It is unsure how the future of these products will progress. By comparing the North American product life cycles to Japan, it would be expected that Canadian softwood plywood is currently in the decline stage and will eventually all but disappear in both North America and Japan. Canadian softwood plywood was included in the North American western plywood production numbers and is suffering from increasing production costs and, as a result, decreasing demand. In addition, Canadian 17 softwood plywood's appeal is to a limited portion of the Japanese market because it is not available in traditional Japanese sizes. The future of OSB as a dominant product in Japan is uncertain. Does OSB have the group of attributes that are desirable and will it be a viable substitute product? There is evidence from the North American market that this is the case. The relationship between OSB and domestically produced softwood plywood in Japan is similar to what was seen in North America between OSB and southern yellow pine. It is possible that OSB went through a prolonged development phase in Japan because it has not been supported by many different OSB producers (i.e. has had limited product promotion) and, as a result, has taken considerable time to be accepted by Japanese panel customers. There have only been two OSB suppliers that have consistently supported the Japanese market since the product was first offered to Japan (Interex 1999). Thus, according to the product life cycle theory, this would elongate the development phase due to a lack of promotional efforts and would create a negative view of the long-term supply of OSB (Levitt 1965). In order to determine whether OSB is currently working through a product development phase and has the potential to become a dominant panel product in Japan, it must determined whether OSB can meet the needs of the Japanese structural panel customers. To understand the needs of Japanese structural panel customers, various marketing research approaches were reviewed and a methodology was designed to access the Japanese market. 2 . 4 R E S E A R C H M E T H O D O L O G Y Understanding customers' needs in the forest business can lead to successful products and product growth. Determinant attribute analysis has proven itself to be a method for understanding what aspects of a product are important to customers when making a purchasing decision. Trinka et al. (1992) discussed that it is essential to understand the needs of customers when developing a substitute structural panel in today's highly competitive market place. However, the historical development of the forest products industry has been driven by reactions to rising costs and potential loss of competitiveness and not by striving to meet customer needs. The technological history of the forest business has been driven by the increasing scarcity of timber and the rising cost of large logs, which has induced innovations that have facilitated the use of smaller logs and inferior fiber sources (Rosenberg et al. 1990). The results of the technical development have created an industry that has become highly automated, capital intensive and increasingly flexible. The increased technical nature of manufacturing has made it 18 easier to modify products even during production, often requiring no plant down time. For example, during the processing of OSB, resin content and panel density along with other parameters can be changed to alter the properties of the finished product without incurring any stoppage in production. Mr. Costa, Technical Manager of an OSB facility, says that "as computer technology continues to advance, production changes become even more systematic and routine" (in Ijtt., Feb 2001). Composite structural panels have a significant amount of flexibility during manufacturing and are routinely adapted to meet different product specifications. 2.4.1 Customer Needs And Marketing Strategy The ability to alter physical characteristics of new products through adjustments in the manufacturing process offers the opportunity to put customer needs back into the forefront of product development in the wood products industry (Trinka et al. 1992). However, the ability to enhance the success of those products by adjusting to customer desires is limited unless there is an awareness of which product characteristics are influential in customer purchasing decisions. When a flexible manufacturing process is complemented with specific knowledge of customers' attitudes, a highly successful marketing strategy can result. Understanding customers' attitudes and product expectations is critical for successful marketing strategies and overall product success. Through research of America's most successful companies Thomas and Waterman (1982) stated in their classic book, In Search of Excellence, "that excellent companies really are close to their customers. That's it. Other companies talk about it; excellent companies do it." However, old-line manufacturing businesses such as forest products often lack the skills to analyze customer desires, market segments and product positioning. It seems that forest products companies de-emphasize customer analysis despite the fact that it is becoming more and more important as the industry is facing over-capacity and fierce competition from both foreign and domestic producers (Doyle and Saunders, 1985). Determinant attributes can be defined as the attitudes toward a product's features or attributes which are most closely related to preference and actual purchasing decisions; the remaining features, no matter how positively viewed by the respondents, are not determinant. It is obviously a significant benefit for marketers to know which attitudes or features lead to or "determine" buying behavior. Determinant features of a product are what a successful marketing strategy should be built around as when positively ranked attributes are consistently reinforced, the result is a commitment to that product (Alpert and Myers 1968, Parasuraman 1983). Furthermore, "a marketing strategy built on determinant attributes can result in a market place competitive advantage" (Trinka et al. 1992: 386). Therefore, it is clearly important to understand those attributes. Additionally, determinant attribute analysis can be particularly useful as a 19 managerial tool for old line industries hoping to penetrate new markets with commodity like products, such as the market for structural panels in Japan (Sinclair and Stalling 1990). 2.4.2 Determinant Attribute Theory Determinant attribute theory is designed around the premise that a product is viewed by a potential purchaser as a group of attributes, as illustrated in Figure 10. The customer then determines their purchasing decision based on how well a product performs on the attributes that they feel are the most important, relative to the performance of competing products. Figure 10. Products viewed as a group of attributes (i.e., birthday cake). Determinant attribute analysis is the technique that establishes which of the product's attributes has the strongest influence on a customer's purchasing decision. The objective is twofold: (1) to understand which factors affect a product's image in the eyes of a customer and (2) determine which of those factors have a significant effect on a customer's purchasing decision. When analyzing a product's image the factors that define that particular product are broken down into a list of attributes. The product's image is then assessed by rating each attribute according to the customer's view. "Those attributes projected by the product's image which lead to the choice of that product are called determinant, since they determine preference and purchasing behavior" (Alpert 1971:184). 2.4.2.1 Approach Descriptions Through market research there are a number of approaches for identifying which of a product's attributes are determinant. The approaches for identifying determinant attributes can be broadly classified into three categories: (1) direct questioning, (2) indirect questioning and (3) observation/experimentation (Alpert 1971, Armacost and Hosseini 1994). The direct approach simply asks respondents which attributes they feel are the most important and the ones chosen the most frequently are deemed determinant. The indirect approach can be as simple as 20 replacing the direct questions from the direct approach to a third person context; for example, asking a question such as, "what most people feel" as opposed to "what v_u feel" (Alpert 1971:1985). The indirect approach can also be highly complex incorporating motivational research techniques. For example, a customer's rating towards a product could be compared to an overall product rating and the respondent's own purchasing behavior. A multiple regression analysis technique could then be used to rank the attributes in order of descending contribution relative to the dependent variable, such as overall preference, thereby identifying the determinant attributes (Alpert 1971). Observation/experimentation is a more familiar method that is typically used in scientific research. Generally, this method is where a certain' activity is observed and recorded. A treatment or a force that alters the characteristic being studied is applied and any changes from the originally observed behavior are identified (Kozak 1999). Understandably, this method is difficult to utilize in market research and particularly in the study of widely used industrial products, as it would be costly and difficult to alter products and then observe the difference in purchasing behavior. Given the obvious importance of customer purchasing behavior a significant amount of research has focused on understanding customer's attitudes and determinant attributes. However, assessing these attributes and selecting a method for identification, as discussed above, a researcher is faced with a variety of data collection techniques and analysis approaches (Alpert 1971, Armacost and Hosseini 1994). Alpert (1971) as well as Heeler et al. (1979) compared several of the most prominent customer choice analysis techniques and found strong support for the direct approach and more specifically, a modification of the direct approach known as the direct dual questioning approach. Subsequently, determinant attribute analysis and the direct dual questioning approach have gained acceptance in marketing research and has been used to identify many different factors determining product choice of both household and industrial products (Armacost and Hosseini 1994, Sinclair and Stalling 1990, Lumpkin et. al. 1985 and Bearden 1977). Given its effectiveness the direct dual questioning approach has been chosen for the market research carried out in this thesis. 2.4.2.2 The Theory of Determinant Attributes And Direct Dual Questioning Direct questioning - the technique of identifying determinant attributes by asking respondents what they consider to be important when making a purchasing decision - although proven to be effective is founded on two highly questionable assumptions. First, it is assumed that the respondent knows why they buy or prefer one product over another and, second, it is assumed that the respondent will willingly 21 disclose this information. "The plain fact appears to be that customers often don't understand their own reasons for purchasing and even when they do, they are unwilling to admit what may make them look foolish or irrational" (Alpert and Myers 1968:16). In order to overcome the limiting assumptions that direct questioning has been founded on and to recognize the two-dimensional makeup of attributes, the direct dual questioning approach has been developed. Rather than asking a customer what they feel is the most important factor when making a purchasing decision a product's determinance is broken down into two parts, importance and difference. To identify determinance, the customer is asked to rank each attribute on its level of importance when considering that product, for example from "very important to not important". The customer is then asked to rank the difference that they perceive between competing product's performance on each attribute, for example from "very different to very similar". Thus, each attribute is given two scores according to the customer's attitude, one for importance and one for difference. A determinance score is then calculated by combining the importance and difference ratings. An attribute is considered determinant if it has a high determinance score, meaning that it ranks high in both importance and difference. The higher the determinance score, the higher the relative determinance of the attribute. Thus, a determinant attribute is viewed by customers as being both important to their purchasing decision and different among alternative products. Logically, a determinant attribute would have considerable influence on which product a customer purchases. 2.4.2.3 Direct Dual Questioning And Attribute Classification The concept of a determinant attribute is dynamic. For example, a product can possess attributes that are highly important but not different between competing products. This is to say that a product may perform highly on an important attribute, but is given no special credit by the customer as all of the other competing products offer the equivalent performance (Alpert and Myers 1968). This would be taken into account in direct dual questioning as the attribute would score high in importance and low in difference resulting in a low determinance score and thus a non-determinant attribute. For the purposes of this study, attributes that score high on importance but low on difference are referred to as "required attributes". An example of a required attribute can be seen in work done by Alpert (1971) when carrying out an analysis on cars. Alpert found that one of the most highly rated attributes in terms of importance was vehicle safety. However, vehicle safety was found to be a very poor predictor of vehicle sales. Further research revealed that customers felt all of the vehicles in the class being studied performed equally well on safety and thus safety was not a determinant attribute. Correspondingly, if any one particular product began to perform poorly on an attribute which all of the 22 products previously performed equally well on, that attribute's rating on difference would increase along with its determinance score and overall determinance. Another possibility is that an attribute scores low on importance and high on difference. This attribute would clearly be non-determinant and is classified as an "exceeded" attribute. Although the product is different among alternatives there is little effect on a customer's purchasing decision as the customer doesn't highly value that particular attribute (i.e. low importance) and the competing products that perform highly on this attribute are given no special credit, their performance has exceeded customers' requirements. The final possibility is that an attribute scores low on importance and low on difference. These attributes are classified as "immaterial" as they are not important and all of the competing products perform equally on this characteristic. For a summary of the different possible attribute classifications refer to Table 1 below. Table 1. Possible scores and attribute classification for direct dual questioning approach Importance Difference Attribute Type Low Low Immaterial - Non Determinant Low High Exceeded - Non Determinant High Low Required - Non Determinant High High Determinant 2.4.2.4 Direct Dual Questioning - Attribute Analysis and Results Although the approach for utilizing direct dual questioning is somewhat flexible and is implemented in a slightly different manner from study to study, the established concept doesn't change. Generically, as shown in Figure 11, attribute analysis can be broken down into four major steps. 23 Figure 11. Generic direct dual determinant attribute approach. 1. Attribute Selection Identification of product attributes that customers consider when making a purchasing decision. 2. Attribute Measurement Respondents score each attribute on importance and on difference. 3. Determinance Scoring Determinance scores are calculated and response bias is considered. I 4. Identification The attributes with the highest determinance scores are identified as determinant. 2.4.2.4.1 Attribute Selection The selection of product attributes to be included in the determinant attribute analysis is crucial, as all of the attributes that a customer considers when making a purchasing decision must be included. The importance lies with the fact that a missed influential attribute will at best create bias and at worst has the potential to render misleading results. Obviously, the two possibilities for a missed attribute that is considered when a customer makes purchasing decision, as shown in Figure 12, are that the attribute would have been non-determinant or determinant. Figure 12. Possible effects of a missed attribute. Missed Attribute? Possibilities Non-determinant Bias Determinant Misleading Results The lesser of the two possibilities would be a missed non-determinant attribute which would effect the mean level of determinance and thus the identification of the determinant attributes, as attributes are deemed determinant by a high determinance score relative to other attributes. If the missed attribute would have been determinant, a significant result of the study would be missing as the determinant attributes that were identified would either be short an attribute or accepted 24 incorrectly. Whichever the case, a missed attribute has a significant negative effect on the validity of the analysis and steps should be taken to avoid excluding any attributes that have influence on a customers purchasing decision. As noted by Sinclair and Stalling (1990) determining which attributes to include in a determinant attribute study is critical, for if evaluative factors important to the customer are overlooked the usefulness of the study will be severely limited. Obviously, when using the results of a determinant attribute analysis to base marketing decisions on, one would not want to be excluding an issue that has a direct effect on a customer's purchasing decision. Generally, to ensure that important attributes are not missed when developing a determinant attribute study, a review of previous research relevant to the topic under consideration and exploratory research should be completed (Sinclair and Stalling 1990, Sinclair and Hansen 1993 and Anderson et al. 1976). 2.4.2.4.2 Attribute Measurement Once the list of attributes has been established a questionnaire is developed to score each attribute on importance and difference. Typically each attribute is scored on an attitude rating ranging from four to six-points. Although there is justification for using a simplified three-point scale for research that is concerned with averages over respondents, results are improved with the use of a five or six point scale, providing that the additional decision making does not create respondent fatigue and increased non-response bias (Lehmann and Hulbert 1972). Obviously, this is much less of a concern for questionnaires that are administered through personal interviewing. 2.4.2.4.3 Determinance Score After attribute measurement the importance and difference scores are combined to derive a determinance score. The researcher has a choice of utilizing a multiplicative or additive model for calculating the scores. A multiplicative model is suggested as the superior method and is the standard used in market research. The choice of an additive model would imply that there was direct equality between importance and difference scores and could result in low determinance scores being out weighted by one of the two factors. As central to the determinant attribute theory the determinacy of an attribute in the decision process is a function of importance and difference, supporting the multiplication model (Moriarty and Reibstein 1986 and Anderson et al.). 25 The determinance scores are calculated by multiplying the importance score by the perceived difference rating as shown in equation (1) and as described by Alpert, (1971). in >\ -1 >: Dij = determinance score for attribute i and respondent j lij = importance score for attribute i and respondent j Yy = difference rating for attribute i and respondent j 2.4.4 Response Bias A second important consideration after the calculation of the determinance scores is the potential for response bias. The response bias is created by the different interpretations that each respondent may have towards the ranking scales for both the importance and the difference. For example, one customer's understanding of "very important" may be different than the next respondent's, thereby creating a bias from one respondent's score to the next (Franke 1985). Considering that the bias is created by each respondent interpreting the ranking scales personally, the determinance scores can be adjusted to compensate for the bias potential by row centering, or normalizing each of the respondent's results (Schaninger and Buss 1986, Howell 1987 and Marshall 1998). In research literature, response bias has long been recognized as introducing measurement error to rating-scale scores. The underlying issue is whether or not each of the respondent's responses should be adjusted to remove response bias and if this adjustment warrants the potential of losing legitimate differences within individual's opinions (Schaninger and Buss 1986). Although this topic is still under debate there is considerable evidence that adjusting for response bias improves accuracy (i.e., a reduction of within respondent variance) (Bass and Wilkie 1973). Moriarty and Reibstein (1986) proved that after adjusting for response bias the results were more closely related to purchasing intent. There are three generally accepted methods for response bias adjustment, row centering, standardization and normalization. Row centering subtracts a respondent's average score from each of the respondent's individual's scores, standardization is the transformation of each individual's response into unit standard deviation with a mean of zero and normalization is where each individual's response is divided by their average score (Schaninger and Buss 1986, Sinclair and Stalling 1990 and Bass and Wilkie 1973). The most preferred transformation is row centering because response bias is reduced and the variability within the respondent's answers is maintained (Green and Carmone 1978 and Trinka et 26 al. 1992). During the determinant attribute analysis, response bias would effect both the importance and difference scores. However, considering that the respondents would treat both the scales equally the response bias can be adjusted for in the determinance scores. The'procedure for row centering is shown below in equation (2) as described during the work by Trinka et al. (1992). (2) DNV=(DV-Xj) DN,j= row centered (normalized) determinance score Dy = determinance score for attribute i and respondent j X j - mean value of Dij for all i of respondent j 2.4.5 Identification Statistically the methods for identifying attribute determinance are relatively straightforward. As previously described, the attribute scores are a mean score calculated from each of the individual responses. Thus, any statistical method that can differentiate between one mean and another can be used to identify determinant attributes. The underlying task is to take the mean determinance scores for each attribute and identify which are significantly higher and thus the most determinant. After the review of previous research there are generally, with some small exceptions (Armacost and Hosseini 1994), two different approaches for identifying determinant attributes from direct dual questioning, as shown in Figure 13. Figure 13. Alternative Approaches to identifying determinant attributes from direct dual questioning. Test Against Mean Determinance Level Z or T-Test Test For Significant Differences In Means Analysis of Variance 1 Group Means Into Homogenous Subsets Scheffe Duncan's Multiple Range Bonferoni Tukey 27 The simplest method and the one used most often was noted by Alpert (1971). Alpert suggested the determinance mean for each attribute could be tested versus the population determinance mean, whereas the population determinance mean would be estimated by the grand determinance mean for all attributes. A one-tailed Z or T-test can then be applied to identify which attributes fall significantly above the grand determinance mean (Alpert 1971, Lumpkin et al. 1985 and Ainsworth 1995). Alpert notes that there is a potential to introduce bias. However, to the extent that there is any bias present, it would result in a test which is too conservative because highly determinant attributes would pull the grand mean towards their sample means and could result in underestimating an attributes determinance. Another apparent weakness of Alpert's suggested method is defining what the grand level of determinance stands for. Mathematically, it is the average level of determinance that is experienced when selecting an attribute. The difficulty is understanding what effect the grand determinance value has on an individual's purchasing decision and whether it is a relevant bench mark for evaluating other attributes. It appears that this limitation has been noted by stating that this method at least allows determinant attributes to be chosen systematically rather than through mere "eyeballing" (Alpert 1971). A somewhat alternative approach would be to utilize a one-way analysis of variance to identify if any of the attribute means from each of the respondent's individual scores are statistically different from one another (Bearden 1977 and Seward and Sinclair 1988). If the means prove to be different, there are four alternative methods for breaking the means into heterogeneous groups (ie. the group with the highest scores being the most determinant). The four methods are a Scheffe test, Duncan's Multiple Range test, Bonferoni T-tests and a Tukey test, refer back to Figure 13 (Trinka et al. 1992, Forbes et al. 1994). 2.4.2.4.6 Direct Dual Questioning - Respondent Analvsis And Results Understandably, direct dual questioning and determinant attributes can be effective tools to identify differences within respondent groups. Furthering this concept, determinant attribute analysis has been suggested and effectively used as a market segmentation tool (Sweitzer 1975, Sinclair and Stalling 1990 and Dole and Saunder 1985). 28 2.4.2.5 Attitudes Versus Reality When evaluating the results of determinant attribute analysis and utilizing the information to support marketing decisions, it is essential to keep in mind that determinant attributes are based on how a customer perceives a product. A determinance score is a measure of a customer's attitudes, which may or may not reflect how the product actually performs. Customer perceptions do not always correspond to what manufacturers believe about their own products, yet it is precisely these perceptions which determine success in the market place (Sinclair 1991). The respondent's attitudes are the most important information as they are what purchasing decisions are actually based upon. The fact that there may be a "gap" in how a product performs versus how a product is perceived to perform is critical information for further market development. For example, if a product is perceived to be not as durable as competing products, but laboratory results clearly show that all products perform equal on durability, a marketing communication program designed to enlighten customers to this fact maybe the optimal solution. Contrarily, if a product is perceived to be low on durability relative to other products and laboratory tests confirm these results, perhaps a technical program designed to enhance product durability is the best solution. 29 3 JAPANESE MARKET RESEARCH METHOD To meet the research objectives, a sampling of Japanese structural panel customers was taken and surveyed with a two-phase questionnaire. The first questionnaire was designed to collect descriptive statistics on each of the respondents while the second questionnaire was based around direct dual questioning and gathered the necessary data for a determinant attribute analysis. The basic research design is outlined in Figure 14 below. Figure 14. Basic research design. Sample of Population I Phase 1 Questionnaire I Phase 2 Questionnaire I Data Summary and Statistical Analysis Respondent descriptive statistics and determinant attribute analysis. 3.1 POPULATION AND S A M P L E F R A M E The population frame was defined as all companies that purchase structural panels in Japan including trading houses, builders and distributors. The population was chosen because it is the actual customer base for panel producers exporting to Japan and thus would be the most relevant for North American producers. 3.1.1 Sampl ing Method A judgement sample of the population was taken with the assistance of Interex Forest Products Japan Ltd. (Interex) and Canfor Woodsales Co., Ltd. (Canfor). Both Interex and Canfor are marketing companies based in Japan that actively import and sell structural panels to a broad group of Japanese customers. The companies to be included in the sampling were selected on the basis of being a significant customer of 30 structural panels in Japan and having enough of a relationship with Interex or Canfor that it was possible to meet the personnel responsible for the panel products purchasing decisions. For the direct dual questioning it was decided that the most effective method for carrying out the survey was to administer questionnaires through personal interviews. It was felt that personal interviews would be optimal given the amount of clarification that would be required to ensure that each of the respondents interpreted the questions in the same manner. Personal interviews would also overcome the difficulty of a low response rate that would have otherwise been expected with a mail survey. Additionally, it was felt that with the personal presence of familiar entities, Interex and Canfor, the respondents would be more comfortable answering questions and the validity of the results would be enhanced. 3.1.2 Sample Size Budgetary and resource constraints were the main determining factors of sample size. A negative aspect of the personal interview approach was the relatively high costs and binding budgetary constraints that limited the time available in Japan and correspondingly the sample size. Traveling in Japan for interviewing was extremely costly and consumed significant resources from Interex and Canfor personnel whom were required for respondent introductions, appropriate meeting etiquette and translation. Given the budgetary constraints, only two weeks of interviewing could be done and with the amount of time required for travel and meetings, approximately two respondents could be interviewed per day. With the two weeks available, nineteen companies were interviewed (n=19). See Table 2 for a list of respondents and their classification. Although this is was a relatively small sample size, it was expected to capture a large portion of the market in regards to structural panel market share as most of the dominant Japanese forest products importing companies were included. 31 Table 2. List of respondents interviewed l n t e r v i e w s ( n =19) O r g a n i z a t i o n C l a s s i f i c a t i o n 1 Maruyoshi Co. Ltd. Trading House 2 Tostem Corporation Builder 3 Mitsubishi Corporation Distributor 4 Venichu Co., Ltd. Distributor 5 Homest - (Shokusan Jutaku) Builder 6 Nichimen Corporation Trading House 7 Bankyo Floor Builder 8 Shingu Shoko., Ltd. Distributor 9 Shimada & Co., Ltd. Distributor 10 Iwatani International Corporation Builder 11 Dainippon Ink and Chemicals Inc. Builder 12 Nissho Iwai Trading House 13 Marubeni Trading House 14 GL Home Builder 15 Mitsui Home Co. Ltd. Builder 16 Sekisui Global Trading Corp Builder 17 Sumitomo Forestry Builder 18 Nichiei Distributor 19 Sanei House Co., Ltd. Builder 32 3.2 SAMPLING P R O C E D U R E AND QUESTIONNAIRE D E V E L O P M E N T In order to utilize a determinant attribute analysis model and the direct questioning approach, an innovative two-phase survey process was developed. The two-phase sampling procedure enabled the collection of descriptive statistics and accurate data for a sound determinant attribute analysis. Refer to Figure 15 for an overview of the sampling procedure. Figure 15. Two-phase sampling process. Judgement Sample of 19 Companies Introductory Letter and Phase 1 Questionnaire Faxed to the 19 Respondent Companies Phase 1 Written Questionnaire Answer Descriptive Statistic Questions Respondent Profile i Identify Panel Attributes j jThat Have Influence on | Purchasing Decisions Attributes For Determinant Analysis Phase 2 Interview Questionnaire Scoring selected attributes on importance and difference. Data Summary and Statistical Analysis Respondent descriptive statistics and determinant attribute analysis. Once a company was selected as part of the judgement sample, the respondent was contacted by telephone to schedule an interview and was subsequently sent a letter. The letter which was translated into Japanese, was sent one week prior to the interview. The purpose of the letter was to remind the respondent of the interview, describe the nature of the research and to introduce the participants. It was thought that this approach would ensure an efficient meeting schedule and further the responding companies willingness to participate. See Appendix I for a copy of the letter that was sent to the participating companies. 33 3.2.1 Phase-One Quest ionnaire The phase-one questionnaire was faxed to the respondent and collected during the personal interview. The questionnaire gathered descriptive information from each of the respondents and had them identify the panel attributes that had influence on their purchasing decisions. It also included questions that were straight forward, requiring limited explanation and could be clearly explained in a written format. The phase one questionnaire helped keep the time required for the interviews in phase-two to a manageable amount by collecting the descriptive information prior to the meeting and also gave the respondent additional time to access any internal information which may have been required for answering company specific questions. See Appendix II for a copy of the phase-one questionnaire. 3.2.1.1 Attribute Selection An effective determinant attribute analysis, based on direct dual questioning, requires each attribute that a respondent considers during the purchasing process to be part of the study. The selection of structural panel attributes to be included in the survey is an extremely important step as any missed attributes that influence the respondent's purchasing decisions will have significant negative effects on the research results, as described in section 2.4.2.4.1 (Sinclair and Stalling 1990). Literature suggests that exploratory research and a careful review of prior studies should be done to ensure that any influential attributes are included. Exploratory research, due to time and budget constraints, was not done as it would have added significant expense. To overcome the risk of excluding important attributes, prior research on structural panel attributes were reviewed and detailed discussions were held with structural panel marketing and production personnel from Ainsworth, Interex and Canfor (Trinka et al. 1992, Seward and Sinclair 1988 and Ainsworth 1995). The research review and discussion produced a substantial list of attributes such that they were broken down into three separate groups, technical, marketing and service. A technical category was created for those attributes that could be empirically tested and where considered objective, such as panel strength. A marketing category was created for those attributes that were based on the overall market aspects such as competitive pricing. The third category, also subjective, was based on service attributes, such as manufacturer's reputation, refer to Table 3. 34 Table 3. Structural panel attribute categories and type Category Attribute Type Technical Attributes Objective Marketing Attributes Subjective Service Attributes Subjective Each of the respondents were then given the attribute list in the phase-one questionnaire and asked to pick "the five most important attributes for each category". In addition, there were four spaces available, per category, where a respondent could add important attributes that they felt were missing. See Table 4 for an example of the attribute selection procedure. The attributes that were chosen in phase one were then transferred to phase-two where they were included in the interviews and determinant attribute analysis. This novel two-phase process ensured that no important attributes were excluded and eliminated the need for exploratory research. Table 4. Attribute selection process for the technical attribute category, during the phase one-questionnaire For each category indicate THE FIVE MOST IMPORTANT ATTRIBUTES. Please add in any important attributes that you feel are missing in the spaces provided at the end of each section. Technical Attributes: Attribute: Importance • Linear Expansion ( ) 1» Panel Stiffness ( ) • Panel Strength ( ) 1* Internal Bond ( ) • Density of Panel ( ) ]• Nail Withdrawal 7 T . . . . T~ ( ) • Uniform Panel Thickness ( ) [» Physical Appearance ( ) • Formaldehyde Emissions ( ) 1* Thickness Swell ( ) • Other (please specify) ( ) 1« Other (please specify) ( ) • Other (please specify) ( ) f«~ Other (please specify) .__ ( ) The open ended selection process also helped to eliminate bias by allowing respondents to select attributes according to their preferences and not forcing them to select from a static list. Additionally, this will assist 35 with the study being replicable overtime as any changes that occur in overall preferences will be included in the study. 3.2.2 Phase-Two Quest ionnaire The phase-two questionnaires were filled out during personal interviews that lasted an average of one and a half hours. The interviews began by acquiring the results from phase-one and transferring the attributes chosen to the phase-two questionnaire. The phase-two questionnaire was designed to score the attributes, identified in phase one, on importance and on difference. 3.2.2.1 Importance The importance score was a ranking of how important a respondent felt an attribute was when selecting a wood based panel for use in home construction. The importance score for each attribute was based on a five point scale ranging from 1 "no importance" to 5 "critical importance", see section one of the phase-two questionnaire in Appendix III. 3.2.2.2 Difference The difference scores for the attributes was a derived score. During the phase one questionnaire, respondents also identified all of the different types of structural panels that they considered during the purchasing process. Based on this information, each product was scored in phase two as to the level that it possessed each of the important attributes. The scores were based on a four point scale, ranging form 1 "not at all" to 4 "high degree". See section three of the phase-two questionnaire in Appendix III. The difference scores for each attribute was then derived by taking the best performing product relative to each attribute, subtracting that score from the worst performing attribute for each product and adding one. The result was a difference score for each attribute ranging from 1 "very similar" to 4 "very different". See section two of the phase-two questionnaire in Appendix III. 3.3 R E S U L T S A N A L Y S I S Upon the completion of the 19 surveys, the results were thoroughly analyzed. The phase-one questionnaires were summarized and compared to industry wide statistics giving a profile of the respondents and providing an understanding of how well the sample represented industry wide traits. The 36 phase-two questionnaires were statistically analyzed and the determinant attributes were identified using univariate techniques like significance tests and one-way ANOVA tests. 37 4 RESEARCH RESULTS 4.1 D E M O G R A P H I C AND DESCRIPTIVE STATISTICS The following sections are the summarized results of the respondent profiles gathered during the phase-one questionnaire, refer to Appendix IV to review the raw data. 4.1.1 Respondent Categor ies A n d Size Out of the 19 respondents the classifications were 26 percent distributors, 53 percent homebuilders and 21 percent trading houses, refer to Figure 16. During classification the respondents were grouped according to their function. If a trading house was manufacturing homes, they were classified as a builder. For example, Sekisui, a major trading house, was classified as a builder because they purchase panels exclusively for manufacturing homes. Figure 16. Respondents surveyed, by category. Qstributor Homebuilder Respondent Category Trading House The average size of the organization sun/eyed for each category based on panel consumption estimates for 1998, were distributors at 14,000 cubic meters, builders at 14,300 cubic meters and trading houses at 500,000 cubic meters. 4.1.2 Panel Type A n d Usage The respondents interviewed had an estimated cumulative 1998 panel consumption of 1.8 million cubic meters. Of the total volumes consumed by the respondents, trading houses represented the largest consumption volume followed by the builders and then distributors, see Figure 17. Figure 17. Total respondent panel volumes, by category. 1600 1400 ters 1200 5 1000 n 3 o 800 "O C ro 600 tn o .c 400 F 200 0 Distributor Homebuilder Trading House Respondent Category 38 After adjusting for any double counting between the categories and deducting any non-structural panel usage, the total net consumption of the respondents was estimated at 1.65 million cubic meters. In 1998, total Japanese structural panel demand was 7.3 million cubic meters (APA 2000) meaning that the respondents surveyed accounted for 23 percent of the total market. It is estimated that the total construction panel market in Japan is 3.8 million cubic meters, see Table 5 below. Table 5. Japanese construction usage of structural panels Panel Use Volume* Percent Roof Sheathing 1,190,000 31% Wall Sheathing 919,000 24% Floor Sheathing 1,710,720 45% Total Market 3,819,000 100% Source: Interex 2 0 0 1 . After deducting the respondent volumes to exclude any non-construction use (i.e. packaging and concrete forming panels) the total volume surveyed was 553,300 cubic meters, which results in a market coverage of an estimated 14 percent. The structural panels that respondents either use or consider using are shown in Table 6 below. Table 6. Structural panels purchased or considered by respondents Panels Used In Structural Applications Distributors Builders Trading Houses Russian Larch Plywood 13% 6% 0% Domestically (Japanese) Produced Lauan Plywood 12% 3% 16% Imported (Indonesia/Malaysia) Produced Lauan Plywood 4% 0% 64% Canadian Structural Plywood (CSP) 28% 25% 11% Southern Yellow Pine Plywood (SYP) 1% 0% 0% Oriented Strand Board (OSB) 30% 52% 5% Douglas Fir Plywood (DFP) 1 1 % 3% 0% Medium Density Fiberboard (MDF) or Particleboard 1% 1 1 % 4% Total 100% 100% 100% Of the panels consumed by the respondents 77 percent of the volume was imported, while domestic panel manufacturers supplied the remaining 23 percent. Relative to the entire Japanese structural panel market, the respondents had a much higher percentage of imports, as the total market is estimated to consume 55 39 percent imported and 45 percent domestically produced panels (APA 2000). Considering that the judgement sample was selected by Interex and Canfor, both of whom market imported wood products to Japan, it was anticipated that the sample would be skewed toward respondents that were focused on imported panels. From the companies surveyed, on a percentage basis per respondent, Canada, Malaysia and Indonesia dominated imports. On a volume basis, Malaysia and Indonesia were the largest panel importing countries as they were a significant percentage of the trading house volumes, which was the largest consuming group, refer to Figure 18 for panel import percentages by respondent category. Figure 18. Respondent import panel consumption by country. Distributors Builders Trading Houses The respondent's imports were weighted much heavier to Canada than the market as a whole. Southeast Asian regions, Indonesia and Malaysia, still dominate Japanese panel imports an estimated 89 percent of the market, refer to Figure 19. The dominant panel thicknesses were 9.5 mm and 12.0 mm, while the largest panel sizes consumed were 3'x8', 3'x9', 4'x8' and a broad classification category called "other". See Table 7 for an overview of the respondents panel thicknesses and Table 8 for respondent panel dimensions. Figure 19. Japanese panel imports by country. Canada 40 Table 7. Respondent primary panel thicknesses Thickness (mm) Distributors Builders Trading Houses Average 6.4 0% 0% 4% 1% 9.0 4% 1% 0% 2% 9.5 53% 24% 63% 40% 11.0 0% 12% 0% 6% 11.1 0% 7% 0% 3% 11.5 2% 0% 1% 1% 12.0 12% 16% 29% 18% 12.5 11% 10% 2% 8% 15.0 0% 3% 0% 2% 15.5 4% 2% 0% 2% 18.0 1% 4% 2% 3% 18.5 0% 4% 0% 2% 21.0 0% 10% 0% 5% 25.0 0% 7% 0% 3% 28.0 12% 0% 0% 3% 100% 100% 100% 100% Table 8. Respondent primary panel dimensions Dimensions (feet) Distributors Builders Trading Houses Average 3x6 29% 13% 54% 26% 3x8 11% ' 32% 16% 23% 3x9 9% 9% 16% 10% 3x10 8% 5% 3% 5% 4x8 17% 10% 12% 12% Other 26% 32% 0% 23% 100% 100% 100% 100% Panel usage for the respondents was predominately traditional, platform and prefabricated homes, making up 88 percent of total consumption, see Figure Figure 20. Respondent panel consumption by end-use. Concrete Form -\ Packaging 41 4.1.3 Panel Quality Difference Traditionally the Japanese mentality was highly protectionist and as a result their markets had a strong preference for domestically produced products. Japanese customers held the general view that domestically produced products were superior to imported goods. However, with changing demographics and fresh views on business, the protectionist market trait is beginning to diminish. The Japanese market is becoming increasingly westernized in culture and in taste (Cohen et. al 2001). Recent surveys have shown that country of origin has little bearing on Japanese purchasing decisions (Cohen and Gaston 1998). These trends are seen in increasing imports and the quicker adoption rate of new products. The respondents were each asked if they felt there was a difference between hardwood plywood made in Japan versus imported hardwood plywood. Out of the 19 respondents, 13 of them had an opinion on this topic, while 6 did not know and stated that they had never directly compared the two products. From the respondents that answered, 5 or 38 percent felt there was a difference while 8 or 62 percent felt there was no difference. Of the 5 that felt there was a difference, 4 felt that the Japanese produced hardwood plywood was superior and 1 respondent felt that imported hardwood was of higher quality. 4.1.4 Home Builders The final portion of the phase-one questionnaire profiled the companies classified as builders. From the 10 companies sun/eyed, 9 were actually homebuilders while 1 exclusively manufactured and installed wooden sub-floor systems predominantly for concrete or industrial buildings. Out of the remaining 9 homebuilders, 2 were traditional home builders, 4 built 2x4 platform homes, 2 built prefabricated hybrid homes and 1 built both 2x4 and traditional homes, refer to Figure 21. Figure 21. Builders surveyed, by home type. Of the 9 homebuilders surveyed, 2 would not reveal ' Hybrid how many homes they built during 1997 as they viewed this statistic as confidential, the remaining 7 homebuilders accounted for 43,845 homes. Wooden housing starts during 1997 were estimated at 611,000 (JAWIC 2001) making the respondents market share approximately 7 percent. The four 2x4 builders total housing starts for 1997 were 41,120. In P l a t f o r m (2x4) 1997 it was estimated that 2x4 housing starts were 45% 79,000 (JAWIC 2001) making the surveyed 2x4 builders market share approximately 52 percent. 42 The majority of the homes built from the companies surveyed ranged from 31 to 50 tsubo5, representing 70 percent, refer to Table 9. Table 9. Builders survey, average home sizes Home Size Percent < 30 tsubo 15% 31 to 40 tsubo 34% 41 to 50 tsubo 36% > 51 tsubo 15% Total 100% 4.2 DETERMINANT ATTRIBUTE IDENTIFICATION The results from the determinant attribute analysis, based on direct dual questioning, identified the panel attributes that had the most effect on Japanese customer purchasing decisions. The methods used to identify these attributes were described in Section 3. 4.2.1 Attribute Selection After a review of prior panel research and discussions with panel marketing and production staff from Ainsworth, Interex and Canfor, a comprehensive list of attributes was created. For the objective technical characteristics the following ten attributes were identified as having the potential to affect purchasing decisions, refer to Table 10 below. Table 10. Technical (objective) attributes 1 Linear Expansion 2 Panel Stiffness 3 Panel Strength 4 Internal Bond 5 Density of Panel 6 Nail Withdrawal 7 Uniform Panel Thickness 8 Physical Appearance 9 Thickness Swel 10 Formaldehyde Emissions The second group of attributes created was based on subjective marketing issues. The ten attributes that were identified are shown in Table 11. Table 11. Marketing (subjective) attributes 5 Tsubo - equivalent to 3.36 square meters. 43 1 Competitive Price 2 Long-term Supply 3 Shipment Arrives In Good Condition 4 Product Availability 5 Environmental Forestry Practices 6 Overal Quality 7 Flexibility of Panel Properties 8 Panel Size 9 Availability of Thicknesses 10 Merchandising Finally, the last group of attributes created was based on subjective service issues. There were 12 different service attributes that were considered to potentially influence purchasing decisions, as shown in Table 12. Table 12. Service (subjective) attributes 1 Manufacturer's Reputation 2 Accessibility of Manufacturer 3 High Level of Service 4 On-time Delivery 5 Manufacturer's Knowledge of Products 6 Manufacturer's Awareness of Customer Needs 7 Manufacturer's Flexibility 8 Personal Relationship With Manufacturer 9 Manufacturer's Market Knowledge 10 Ability to Fil  Rush Orders 11 Ability To Fil  Smal Orders 12 After Sales Service The attribute groups were presented in the phase one questionnaire and each respondent identified the five most influential ones from each category, as described in Section 3.2.1. In addition, as discussed in Section 2.4.2.4.1 and 3.2.1.1, there was an opportunity to add additional attributes to each category if the respondent felt that influential attributes were missing. The process resulted in the addition of three attributes to the technical attribute category and no additional attributes were added to the marketing and service categories. The attributes added to the technical category are listed below in Table 13. Table 13. Additional technical (objective) attributes identified by the respondents during the phase one questionnaire 1 2 3 Sound Panel Durability Edge Swel Penetration 4.2.2 Attribute Frequency The first step in identifying the determinant attributes was based on frequency of selection. With the substantial list of attributes and the open-ended attribute selection process for each respondent, it was felt that the attribute lists were exhaustive and did not exclude any potentially influential attributes. However, the challenges created by this method were an unequal number of observations per attribute which created a slightly more complicated statistical analysis and numerous attributes to consider. Including the three different attribute categories and the ones that were added by the respondents, there was a total of 35 44 attributes to include in the determinant analysis, refer to Figure 22 for a listing of attributes and associated response frequency. Figure 22. Attributes and respondent selection frequency. Frequency of Attribute Selection 20 18 16 14 12 10 8 6 4 2 0 n 2 5 _ 12 10 n 7 8 n 11 14 13 X L i l l n n n 5 S n n D CD E Z Cl X L I o 1 i § 5 -Attribute The attributes were chosen for further analysis based on respondent selection frequency. Any attributes that did not have a selection frequency greater than 10, exceeding 50 percent, were eliminated and considered as non-influential attributes. Referring back to Figure 22, any attribute that was below the dotted line was excluded leaving 14 different attributes to consider. The reduced set of attributes and selection frequencies now considered to be "determinant" attribute frequencies are shown Figure 23. The arbitrary way of reducing the set of attributes based on frequency was justified in three ways. First, common sense dictated that when considering an attribute's determinance, i.e. effect on customers' purchasing decisions, it is logical that the attribute should not be used as a guide for marketing decisions if at least 50 percent of the respondents did not feel that it had any influence on their decision making. Second, when using direct questioning it has been proven that an effective method, although not as effective as others, for identifying determinant attributes is selection frequency. When using direct questioning, attributes can be classed as determinant if they are among the most frequently chosen (Alpert 45 1971), refer to section 2.4.2.1. Finally, the frequency and determinance score for each of the attributes was tested for a relationship with correlation analysis. With direct dual questioning, the higher the determinance score the more likely the attribute will be deemed determinant. Thus, if frequency were also a predictor of determinance it would be expected that there would be a positive relationship (correlation) between frequency and the determinance score. To test this relationship a linear correlation coefficient (r) was calculated at .5113 and compared to a critical r value with n-2 degrees of freedom at a .10 level of significance. The critical r value was calculated to be .4577 (Kozak 1966) and thus the hypothesis that there was no significant relationship between selection frequency and determinance score was rejected and the results support the validity of utilizing an arbitrary frequency cutoff. The results lend credibility to the method used but are not conclusive given that the hypothesis is not rejected at a .05 level of significance. Figure 23. Attributes with determinant frequencies. Determinant Frequency of Attribute Selection 20 Each of the three attributes that were included by the respondents in the phase-one questionnaire, Sound Penetration, Panel Durability and Edge Swell, were only selected once by the respondents that included them. Thus, they were considered non-determinant by the arbitrary frequency cutoff. Given the above results, it was felt that the 14 selected attributes included all of the potentially influential and determinant attributes. 46 4.2.3 Importance Each of the 14 attributes were first considered based on their importance, which was derived by each of the 19 respondents' scores. Refer to Table 14 below to review the mean respondents' importance scores per attribute and corresponding standard deviation. Table 14. Mean importance scores for technical, marketing and service attributes for structural panels in Japan Attributes Standard Deviation Mean Importance1 Competitive Price .42 4.79 Thickness Swell . .62 4.50 Panel Size .53 4.50 Panel Strength .82 4.00 Shipment Arrives In Good Condition .70 3.91 Linear Expansion .65 . 3.73 On-time Delivery .47 3.71 Uniform Panel Thickness 1.34 3.70 Formaldehyde Emissions .72 3.63 Overall Quality .74 3.60 Long-term Supply .90 3.27 Ability To Fill Rush Orders .47 3.27 Environmental Forestry Practices .72 2.83 Manufacturer's Awareness of Customer Needs .63 2.80 1Scale 1 to 5:1=No/Little Importance; 5=Absolutely Critical 4.2.4 Difference Following importance, the 14 attributes were then considered on their difference scores. The difference scores were calculated by ranking each attribute's performance based on the structural panels that the respondent considered when making a purchase. The different panels considered when making a purchasing decision can be seen in Section 4.1.2, Table 6. The result was a derivation of a difference score for each respondent and attribute. A mean difference score for each attribute was then calculated and is listed in Table 15. 47 Table 15. Mean difference scores for technical, marketing and service attributes for structural panels in Japan Attributes Standard Deviation Mean Differnence 1 Long-term Supply .93 3.45 Environmental Forestry Practices .79 3.42 Thickness Swell .69 3.33 Shipment Arrives In Good Condition .60 3.18 Competitive Price .60 3.16 Panel Size .32 2.90 Formaldehyde Emissions .77 2.75 Uniform Panel Thickness .82 2.70 Overall Quality 1.06 2.60 Ability To Fill Rush Orders .93 2.55 On-time Delivery .87 2.41 Linear Expansion 1.29 2.36 Panel Strength .88 1.90 Manufacturer's Awareness of Customer Needs .70 1.40 'Scale 1 to 4:1=Very Similar; 4=Very Different 4.2.5 Determinance The importance scores, scale of 1 to 5, and difference scores, scale of 1 to 4, where combined with a multiplicative model deriving a determinance score ranging from 1 to 20, as described in Section 2.4.2.4.3. The resulting mean determinance scores are shown below in Table 16. Table 16. Mean determinance scores for technical, marketing and service attributes for structural panels in Japan Attributes Standard Deviation Mean Determinance 1 Competitive Price 3.54 15.21 Thickness Swell 4.22 15.17 Panel Size 2.33 13.10 Shipment Arrives In Good Condition 3.80 12.55 Long-term Supply 4.13 11.09 Formaldehyde Emissions 3.36 9.94 Uniform Panel Thickness 4.61 9.80 Environmental Forestry Practices 2.81 9.58 Linear Expansion 5.78 9.27 On-time Delivery 3.60 8.94 Overall Quality 3.17 8.93 Ability To Fill Rush Orders 4.11 8.55 Panel Strength 3.92 7.60 Manufacturer's Awareness of Customer Needs 1.93 3.80 1Scale 1 to 20:1=Little Determinance; 20=High Determinance 48 4.2.6 Response B ias After deriving each of the determinance scores an adjustment was made for potential response bias, as discussed in Section 2.4.2.4.4. The adjustment was made as there is a significant amount of evidence that suggests there is an improvement in the validity of the determinance results. However, for comparison purposes a full determinant attribute analysis was completed with and without a response bias adjustment and both methods gave identical results. As would be expected the determinance scores that were adjusted for response bias had a higher between groups source of variation in the one-way analysis of variance, leading to a stronger rejection for a difference in means. In order to complete the determinant attribute analysis a response bias adjustment was made. The response bias adjustment was made through row centering each of the respondent's attribute scores by subtracting the respondent's average score from each individual score. In addition a constant of 7 was added to the row centered score as illustrated below in equation (3). (3) DN^iD^-X^ + C DNij = row centered (normalized) determinance score Dij = determinance score for attribute i and respondent j Xj = mean value of Dij for all i of respondent j C = Constant, 7 The constant was added as a matter of convenience to avoid negative determinance scores and 7 was chosen, as it was the smallest number that could be added while maintaining positive scores. No information loss is associated with the response bias adjustment and the addition of a constant as the determinant analysis identifies the perceived relative determinacy of the product attributes as opposed to their absolute determinacy (Moriarty and Reibstein 1986). The resulting determinance scores are listed in Table 17 and are illustrated in Figure 24. 49 Table 17. Row centered mean determinance scores for technical, marketing and service attributes for structural panels in Japan Attributes Standard Deviation Mean Determinance 1 Competitive Price 2.79 11.61 Thickness Swell 3.76 11.52 Panel Size 2.89 9.31 Shipment Arrives In Good Condition 4.28 8.74 Long-term Supply 3.58 7.64 Formaldehyde Emissions 3.24 6.60 Environmental Forestry Practices 2.67 6.52 Uniform Panel Thickness 4.66 5.78 Linear Expansion 5.32 5.36 On-time Delivery 2.91 5.19 Overall Quality 3.10 5.11 Ability To Fill Rush Orders 3.16 5.10 Panel Strength 3.95 4.06 Manufacturer's Awareness of Customer Needs 2.20 .35 1Mean overall score is 7 (constant), the higher the score the more impact on purchasing decisions. Figure 24. Row centered mean determinance scores for technical, marketing and service attributes for structural panels in Japan. Ability To Fill R u sh Orders nufacturer 's Awa renes s of C . Needs O n-tim e D elivery Pane l S i ze O verall Q uality Env i ronmenta l Forestry Pract i ces S h ip m en t A rrives In Good Condit ion Long-term Supp ly C om petitive P rice Forma ldehyde Em i s s i ons Th i c knes s Swe l l Uniform P anel T h ickn ess Pane l Strength Lin ear Expans ion 1 5. 1 0 19 2 0.35 1 5 1 1 1 1 5 1 6.52  9 8.74 9 7.64 31 I 6.60 5.78 .36 I 1 4.06 1 5 2.0 4.0 6.0 8.0 10.0 Determinance Sco re J 1 1 .61 ] 1 1 .52 1 2.0 14.0 50 4.2.7 Determinant Attributes - All Attributes Once the determinance scores for each attribute had been calculated and adjustments made, they were tested to reveal which of the attributes were determinant. The first step was to find out if there were any significant differences between the 14 sample means. To accomplish this a one-way analysis of variance was completed (Bearden 1977 and Seward and Sinclair 1988). 4.2.7.1 Analysis of Variance Assumptions In order to utilize the analysis of variance technique four different assumptions must be considered (Kozak 1998): 1. The populations for each group from which the samples were obtained must be normally, or approximately normally, distributed (normality). 2. Statistically, the group to group variances should not be different (equal variances). 3. The samples from group to group must be independent (independence). 4. The samples must be taken randomly from each group (randomness). The population that the sample was drawn from was large, estimated to be over 5000 companies, and considered to be normally distributed. However, as described by Kleinbaum et al. (1988) the normality assumption can be relaxed if you are dealing with relatively large sample sizes (e.g., 20 or more). In this case with a sample size of 19 and the large population size, it was felt that the normality assumption was satisfied. Statistically there were no differences in the sample variances for each of the attributes. Although the test statistic for analysis of variance is insensitive to departures from the equal variance requirement it is recommended that a test for homogeneity of variances be performed, particularly for work that has an unequal number of observations per sample as was the case with this research (Walpole 1982). The equal variance assumption was tested with a method called Bartlett's test. The Bartlett's test found no differences in the sample variances for each of the attributes at a .05 level of significance. The test calculated a chi-square value of 15.491 and a chi-square critical value of 22.36, accepting the hypothesis of equal variances. The samples from group to group were independently taken in separate interviews. The interview guidelines where plainly set out by the phase-two questionnaire and the results from one interview had no effect on the results of another, satisfying the assumption. 51 The random sample assumption for the analysis of variance was clearly violated as the survey was based on a judgment sample. However, analysis of variance using fixed-effects for a one way ANOVA was still an appropriate and accurate method to use considering that all of the other assumptions had been met; the one-way analysis of variance test has proven to be robust, meaning that with modest deviations from the above assumptions the results will hold true (Kleinbaum et al. 1988). 4.2.7.2 Analysis of Variance A one way analysis of variance was performed on the 19 different respondent scores for the 14 different attributes, see the following one way analysis of variance summary and ANOVA in Table 18. Table 18. Single factor summary and ANOVA table for the 14 different attributes Summary Groups Count Sum Average Variance St. Dev. Linear Expansion 11 58.939 5.358 28.269 5.317 Panel Strength 10 40.569 4.057 15.587 3.948 Uniform Panel Thickness 10 57.814 5.781 21.759 4.665 Thickness Swell 18 207.386 11.521 14.129 3.759 Formaldehyde Emissions 16 105.595 6.600 10.479 3.237 Competitive Price 19 220.497 11.605 7.799 2.793 Long-term Supply 11 84.053 7.641 12.825 3.581 Shipment Arrives in Good Condition 11 96.178 8.743 18.329 4.281 Environmental Forestry Practices 12 78.289 6.524 7.133 2.671 Overall Quality 15 76.720 5.115 9.637 3.104 Panel Size 10 93.111 9.311 8.352 2.890 On-time Delivery 17 88.247 5.191 8.446 2.906 Manufacturer's Awareness of Customer Needs 10 3.547 0.355 4.838 2.200 Ability to Fill Rush Orders 11 56.053 5.096 9.965 3.157 A N O V A Source of Variation S S df MS F P-Value F crit Between Groups 1589.126 13 122.240 10.032 2.189E-15 1.779 Within Groups 2034.970 167 12.185 Total 3624.095 180 52 The one way analysis of variance tesfs the hypothesis that the sample means for each of the 14 different attributes are statistically not different. The test resulted in an F value of 10.03 compared to an F critical value of 1.78, rejecting the hypothesis and proving that there are significant differences between the sample means. 2.7.3 Attribute Identification The hypothesis for the analysis of variance was rejected denoting that the means were not all equal and that one or more of the attribute means were statistically more determinant than the rest. In order to decipher which means were the most determinant, a Scheffe test was used to separate the significantly different means into subsets of homogeneous means. The Scheffe test found three distinct different groups of means. Below in Table 19 each of the attributes is listed with their determinance mean and the dotted lines under the attributes indicate the three distinct groups of homogeneous means. Table 19. Homogeneous attribute groups and determinant attributes co O rer 4 — o 3 cn « fact <° CD CU C CD 3 CD Z C L_ CO CO 2 < CD C CD i CO ai c CO CL co or — CO LL CD i° < CO O CD > o cu > "53 Q CD c O c g CO c CO CL X LU L_ CO CD c CD 5 <° CD CO CL CD E 5 t O £ 'E I -=) co _ CD ™ .y CD CD e £• CD "> CD 2 T3 CO ID .<2 E E Q . CL CO cn c o CD O > :tj < o c O CD T3 E o f3 CO CD N CO ID c CD 0-CD CO co co CD c O 0.35 4.06 5.10 5.11 5.19 5.36 5.78 6.52 6.60 7.64 8.74 9.31 11.52 11.61 *Found to be significantly above the mean level of determinance. As per the determinance scores, the most determinant attributes, in order of determinance, are Competitive Price, Thickness Swell, Panel Size, Shipment Arrives in Good Condition, Long-term Supply, Formaldehyde Emissions, Environmental Forestry Practices, Uniform Panel Thickness and Linear Expansion. A second group of determinant attributes also includes On-time Delivery, Overall Quality, Ability To Fill Rush Orders and Panel Strength. Clearly, the above analysis indicates that Manufacturer's Awareness of C. Needs is at a lesser level of determinance than the other attributes. As illustrated in Figure 25, any attribute with a determinance score exceeding 5.36 was deemed determinant. 53 Figure 25. Determinant Attributes. Competitive Price Thickness Sw ell Panel Size 1 Shipment Arrives In Good Condition Long-term Supply Formaldehyde Emissions Environmental Forestry Practices I Uniform Panel Thickness G Linear Expansion On-time Delivery Overall Quality Ability To Fill Rush Orders Panel Strength anufacturer's Aw areness of C. Needs Q 0.35 11.61 11.52 • 9 3 1 2 8.74 | 7.64 6.60 1 6.52 • 5.78 | 5.36 ] 5.19 1 5111 10 : 5 ] 4.06 Most Determinant Determinant Non-Determinant 2.0 4.0 6.0 8.0 10.0 12.0 14.0 Determinance Score 4.2.7.4 Further Attribute Identification Out of the 14 attributes, the analysis of variance/ Scheffe method found 9 to be determinant. In an effort to investigate the determinant group and further refine the selection by identifying any attributes that were significantly more determinant than others, a one tailed Z-test method was utilized (Alpert 1971 and Lumpkin et al. 1985). From the most determinant group of homogeneous means, a grand determinant mean was calculated, this average was meant to indicate a typical level of determinance when making a purchasing decision. Any attribute that was statistically greater than the grand determinance mean was posited as determinant. The grand determinance mean was 8.12 with a standard deviation of 2.338. Each of the attributes were then tested against the grand mean, a one tailed T-test was used instead of a Z-test as the population variance was estimated with the sample variance. Competitive Price and Thickness Swell proved to be significantly above the mean level of determinance and thus were considered to be the most determinant attributes, refer back to Figure 25. 54 4.2.8 Determinant Attr ibutes - By Attribute Category After the identification of the determinant attributes was completed, the attributes were then considered via the same methods for each product category, technical, marketing and service, see Table 20. This was done to provide information that could be utilized for the formulation for an overall marketing strategy. For example, the service attributes have proven to be the least determinant; however, when formulating a marketing plan service will certainly be part of the overall strategy and thus it is important to know which of the service attributes are determinant. Table 20. Attributes by category Attributes and Category Mean Determinance Technical Linear Expansion 5.36 Panel Strength 4.06 Uniform Panel Thickness 5.78 Thickness Swell 11.52 Formaldehyde Emissions 6.60 Marketing Competitive Price 11.61 Long-term Supply 7.64 Shipment Arrives In Good Condition 8.74 Environmental Forestry Practices 6.52 Overall Quality 5.11 Panel Size 9.31 Service On-time Delivery 5.19 Manufacturer's Awareness of Customer Needs 0.35 Ability To Fill Rush Orders 5.10 For each of the categories the attributes were analyzed as done initially for the attributes as a whole. A one way analysis of variance was completed to determine if there were significant differences in the means. As well, a Scheffe test was performed in order to group the means and a T-test in order to identify any attributes that had a statistically higher determinance mean than the group average. 4.2.8.1 Technical Attribute Determinance The one way analysis of variance found that there were significant differences in the attribute means at a .05 level of significance, refer to Appendix V to review the ANOVA table. The analysis of variance 55 assumption for statistically similar variances between the sample means was checked with a Bartlett's . test and found that the variances were equal, satisfying the assumption. The Scheffe test found two " groups of homogenous means with Thickness Swell being the lone most determinant attribute, refer to the doted lines in Table 21. Table 21. Homogeneous attribute groups, technical attribute category CD C CO I 55 "55 c CD 0-c g CO c CD Q . X LU CD CD C CD C CO CD CO 0_ CD E - i fc O £ IE C H CD Is <5 2 "D CO 11 few CD CO CO CO CD c o 4.06 5.36 5.78 6.60 11.52 'Found to be significantly above the mean level of determinance. Thickness Swell also proved to be statistically higher than the mean determinance level using a one tailed T-test at a .05 level of significance. 4.2.8.2 Marketing Attribute Determinance The one way analysis of variance of the marketing attributes also found significant differences in the attribute means at a .05 level of significance, refer to Appendix VI to review the ANOVA table. Again the equal variance assumption was tested with a Bartlett's test and found the variances to be statistically not different. The Scheffe test found two groups of homogenous means with Competitive Price, Panel Size, Shipment Arrives In Good Condition, and Long-term Supply being the most determinant, refer to the doted lines in Table 22. Table 22. Homogeneous attribute groups, marketing attribute category CD 3 o CD > o CO _ CD CD . O c o CD CD i * > CO C CD LU * CL CL 3 CO CD C o CD N to a3 c CD 0. CD O CD > CD CL E o o 5.11 6.52 7.64 8.74 9.31 11.61 *Found to be significantly above the mean level of determinance. 56 The attributes were tested against the mean determinance level using a one tailed T-test at a .05 level of significance and found that only Competitive Price was significantly above the average. 4.2.8.3 Service Attribute Determinance The one way analysis of variance of the service attributes again found significant differences in the attribute means at a .05 level of significance, refer to Appendix VII to review the ANOVA table. The Bartlett's test was used to check the variances and found them to be statistically not different satisfying the analysis of variance assumption. The Scheffe test found two groups of homogenous means with On-time Delivery and Ability To Fill Rush Orders being the most determinant, refer to the doted lines in Table 23. Table 23. Homogeneous attribute groups, service attribute category er's of C. Rush ivery D m " — cn CD fact nes eed LL. CD Q fact nes eed To Ord. CD To Ord. E CD 5 >. Ord. 1 ^1 Abil On 0 . 3 5 5 . 1 0 5 . 1 9 The attributes were tested against the mean determinance level using a one tailed T-test at a .05 level of significance and found that none of the attributes'were statistically above the average. 4.2.8.4 Attribute Category Importance The final aspect that was considered in regards to the attribute categories was the weight each category had on the respondents when making a purchasing decision. Category importance was assessed with a direct questioning approach, during the phase-one questionnaire. The respondents were asked to rate, on a percentage basis, how much they felt the particular attribute category effected their panel purchases. The results for rating each of the attributes categories placed technical attributes to be the most important at 43.6 percent, followed by marketing attributes at 38.9 percent and lastly, service attributes at 17.5 percent. The results, although interesting given that they indicate that most customers felt that technical and marketing attributes were more influential than service attributes, were not statistically 57 significant. The percentage ranking provided by each of the respondents had such a high standard deviation that no reliable conclusions could be drawn. It would; however, appear that the service attributes were felt to be the least important, refer to Table 24. Table 24. Category Importance Attribute Average Ranking Standard Category Percentage Deviation Technical 44% 23% Marketing 39 % 22 % Service 17% 11% 100% 4.2.9 Determinant Attr ibutes - By Respondent Category The final consideration for the determinant attribute analysis was done by market segment. The attributes were evaluated for determinance for each respondent classification, distributors, builders and trading houses. The process for identifying the determinant attributes was as done in the previous analyses. 4.2.9.1 Distributors Out of the 19 respondents, 5 were classified as distributors (n=5). The distributors' responses for each attribute were considered and the attributes with over a 50 percent response rate (i.e. a frequency of 3 or more out of 5 distributors) were selected as potentially determinant. From the 35 attributes the following 14 were selected on frequency as potentially determinant for the distributor market segment, refer to Figure 26. 58 Figure 26. Determinant frequency selection for distributors. Determinant Frequency of Attribute Selection - Distributors 6.00 5.00 4.00 3.00 2.00 1.00 n n r n V ^ # ,0 * « ^ + # J* / f -co J T W © O _<JC Jf» * < J Attribute An analysis of variance was completed on the 14 different attributes and found that there was a significant difference in the attributes sample means at a .05 level of significance, refer to Appendix VIII for the analysis of variance ANOVA table. Bartlett's test was used to check the equal variance assumption for the analysis of variance and found that they were not statistically different, satisfying the assumption. Scheffe's test identified two groups of homogeneous means, see Table 25 and refer to the dotted lines. The attributes were also T-tested against the mean level of determinance. Thickness Swell and Competitive Price were found to be significantly more determinant than the other attributes in the most determinant group of means, as noted in Table 25. 59 Table 25. Homogeneous attribute groups and determinant attributes for distributors in M _ '<- o 2? CD co J§ -2 "5 3 5 £ CD l i C 0) CD 0) 5 <« CO CO 0. CD If 3 Q. a. =i CO c o CD > "55 Q CD E • c O CD a> Q T3 CO 15 .S2 E E few ca o CD > o CD o ' £ CD CO CO 0) co CO CD 5 CO 2 "a c o CO c CO a. X LU Lm CO CD CO g <D Q > 'E ^ CD -a E 9 CO CD N CO "53 c CO 0 . CD co CO CO CD c J £ o CD O CD > CD CL E o O 1.10 4.78 5.40 6.42 6.54 6.70 6.74 7.59 7.61 8.02 9.77 11.77 12.74 13.34 'Found to be significantly above the mean level of determinance. >.9.2 Builders The builder's most frequently chosen attributes were selected for determinant analysis. There were 10 respondents classified as builders and any attributes that were selected by 5 or more of the respondents were included in the determinant analysis, refer to Figure 27. Figure 27. Determinant frequency selection for builders. Determinant Frequency of Attribute Selection 12.00 10.00 ui S 8.00 c o j» 6.00 ct ° 4.00 o z 2.00 f .<S* o # Jfi Jp / / / / / J - -^ / ^ <? <t j r j r ©<» «^ <r o • -s° Attribute 60 An analysis of variance was completed on the 14 attributes and found that there were significant differences within the sample means, refer to Appendix IX for the analysis of variance ANOVA table. The assumption of equal variances was tested with Bartlett's test which found them to be equal. Finally, Scheffe's test was used to separate the means into homogeneous groups. Two groups of means were found to be different from each other, refer to dotted lines in Table 26 below. Table 26. Homogeneous attribute groups and determinant attributes for builders n p CD o o iS CD 3 CD c -o CO CD o c cn c CO 0) c CO 0. CD U e CD CO V) 0) CO C O 1_ CD CE < _cg 'co > < o 3 T3 O CD -a 1 o (/) 3 CC o 15 < g 'co c CO a. x LU L_ CO CD c CD > a5 Q CD E • c O CO 3 a CD > o in c o E LU CD TJ >, . C CD _g CO E CD 3 .9 | "> c LU Q. CL 3 C O cn o -a o o O CD Q > S ' E c < o V O CO CD o CD > CD CL O O CD CO co co CD C O 1.63 4.32 5.78 5.92 6.02 6.36 6.39 6.82 7.40 8.02 11.07 11.82 12.32 13.62 *Found to be significantly above the mean level of determinance. T-testing each mean against that average level of determinance found that only Thickness Swell was significantly above the mean determinance level. 4.2.9.3 Trading Houses Of the 19 respondents, 4 were classified as trading houses. Although this was a rather small sample size to consider (n=4) the trading houses represented such a large volume of panels it was felt that considering them as a separate segment was worthwhile, refer to Section 4.1.2. However, due to the small sample size it was recognized that the possibility of identifying differences would be limited. Again the group of 35 attributes were narrowed based on selection frequency. Out of the entire group of attributes, 18 of them were considered to be potentially determinant. Any of the attributes that were considered by 2 or more of the trading houses were included in the attribute analysis, refer to Figure 28 for the attributes and their corresponding selection frequencies. 61 Figure 28. Determinant frequency selection for trading houses. Determinant Frequency of Attribute Selection 4.50 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 n jJP & A e . 0 ^ afT xe> o 1 ^ # ^ j p v ^ Is* • / f yj ^ Attribute An analysis of variance was completed to identify if there were any differences in attributes means and the equal variance assumption was checked with Bartlett's test. The variances were found to be equal and the analysis of variance found significant differences in the sample means at a .05 level of significance, refer to Appendix X for the analysis of variance ANOVA table. However, with an average sample size of 2.72, any differences picked up by the statistical test would have to be extremely large. Although the analysis of variance test found significant differences in the sample means by rejecting the equality hypothesis, it was a relatively weak rejection with an F statistic of 2.96 and an F critical value of 1.96. Correspondingly there were not large enough differences in means for the Sheffe's test to identify the different groups of means and thus they were considered as one block. T-testing the means against the average level of determinance for all means found that Formaldehyde Emissions and Competitive Price were significantly more determinate than the average, refer to Table 27. 62 Table 27. Homogeneous attribute groups and determinant attributes for trading houses 3 3 CO "O " i - O i i " C o CO 3 CO c n co cu 5 1 o c cn cn CD C (0 < cn l .2 3 C CO cn cn cu c O CU c CO CL E £ c CD CO ^ s, co-g? a> a) 3 S o 2 >> CL CL 3 CO E I CD C o •o o o o cn S i < o CU E CL CO CO o LL CO 3 .2 2 > c LU e cu •p cn 3 CC o 15 < co E CO = a LL a p i I- C >l 15 < in c o "cn co 'E LU CD T3 >* J : CD co CU CL E o O * - LO CD LO LO CD CS1 CO CO O CO & -e? no icsl' «^  lis LO GO CO GO CO CO 00 CO S - N- CO CO 0 0 LO 0 0 CO CM co co a> ^ « 5 *Found to be significantly above the mean level of determinance. .2.9.4 Distributors And Trading Houses The last market segment to consider was a combined group of both distributors and trading houses. This analysis was done in order to raise the sample size to 9, enabling a more sensitive analysis and was justified because both trading houses and distributors chiefly perform the same function of purchasing and reselling panels. Again the 35 attributes were refined, based on respondent frequency. There were 14 attributes with a selection frequency exceeding 50 percent, refer to Figure 29. Figure 29. Determinant frequency selection for distributors and trading houses. Determinant F requency of Attribute Select ion 10 9 8 7 6 5 4 3 2 1 0 1 • _ • ' / / / / / / / / ^ V 4? J* A 4 W v C # » # Attribute 63 After identifying the most frequently selected attributes an analysis of variance was completed to determine if there were any differences in the attribute means. The means were found to be different at a .05 level of significance, refer to Appendix XI for the analysis of variance ANOVA table. The equal variance assumption was tested with a Bartlett's test and the variances were found to be not different. Table 28. Homogeneous attribute groups and determinant attributes for distributors and trading houses co o D co -o CD B Z o o i2 CD C " O TO CD 5 o c OA ts> CD C CJ CD c ro c *2 CD 11 jo ^ c CD CD o CD > o Q. a. D 00 C D C o t CD "55 a O ro c CD v 00 "53 c CO 0. if) 0) LL CO _ CD CO O C o | 2 o 1 > c LU ro fo CO c g CO CO I LU CD "D >. J 3 CD CO E o o to c CD O l l H O CD E Q. In 00 CD N 00 "CD c CD O-CD 00 CO CO CD c o CD o CD > CD Q. E o O 1.63 4.32 5.78 5.92 6.02 6.36 6.39 6.82 7.40 8.02 11.07 11.82 12.32 13.62 Tound to be significantly above the mean level of determinance. T-testing the attributes versus the mean determinance level, Competitive Price was identified as being significantly above the average. 64 5 DISCUSSION 5.1 DETERMINANT ATTRIBUTE C O M P A R I S O N BY R E S P O N D E N T C A T E G O R Y The following table, Table 29, compares the determinant attributes for all respondents to the determinant attributes that were identified for each of the four different respondent categories, distributors, builders, trading houses and distributors and trading houses. The shaded boxes represent attributes that were found to be determinant in Section 4.2. Table 29. Determinant attribute results, comparison by respondent category Rank All -£ Respondents d Distributors i Builders S Trading o_ cc Houses ^ Distributors § and Trading 0 1 Houses 1 "Competitive Price 1 'Competit ive Price 1 "Thickness Swell 1 "Competitive Price 1 "Competitive Price 2 'Thickness Swell 2 •Thickness Swell 2 Competitive Price 2 "Formaldehyde Emissions 2 Thickness Swell 3 Panel Size 3 Panel Size 3 Shipment Arrives In Good Condition 3 Thickness Swell 3 Panel Size 4 Shipment Arrives In Good Condition 4 Shipment Arrives In Good Condition 4 Long-term Supply 4 Panel Size 4 Shipment Arrives In Good Condition 5 Long-term Supply 5 Linear Expansion 5 Environmental Forestry Practices 5 Ability To Fill Small Orders 5 Formaldehyde Emissions 6 Formaldehyde Emissions 6 Nail Withdrawal 6 Formaldehyde Emissions 6 Ability To Fill Rush Orders 6 Nail Withdrawal 7 Environmental Forestry Practices 7 After Sales Service 7 Overall Quality 7 Environmental Forestry Practices 7 Environmental Forestry Practices 8 Uniform Panel Thickness 8 Overall Quality 8 On-time Delivery 8 Shipment Arrives In Good Condition 8 Panel Strength 9 Linear Expansion 9 Formaldehyde Emissions 9 Linear Expansion 9 On-time Delivery 9 On-time Delivery 10 On-time Delivery 10 On-time Delivery 10 Ability To Fill Rush Orders 10 Long-term Supply 10 Long-term Supply 11 Overall Quality 11 Long-term Supply 11 Product Availability 11 M.'s Market Knowledge 11 Overall Quality 12 Ability To Fill Rush Orders 12 Uniform Panel Thickness 12 After Sales Service 12 Nail Withdrawal 12 M's Market Knowledge 13 Panel Strength 13 M's Market Knowledge 13 Panel Strength 13 Panel Strength 13 Uniform Panel Thickness 14 M's Awareness of C. Needs 14 M's Knowledge of Products 14 M's Awareness of C. Needs 14 Uniform Panel Thickness 14 M's Knowledge of Products 15 Overall Quality 16 M.'s Awareness of C. Needs 17 Internal Bond 18 M's Knowledge of Products *Found to be statistically above the average determinance level. 65 The following is a discussion of the most determinant attributes based on the All Respondent category which has the strongest statistical results with the largest sample size, shown in the first column of Table 29. 5.1.1 Competi t ive Pr ice And Th ickness Swel l Competitive Price and Thickness Swell were the two most determinant attributes for every respondent category, with the exception of trading houses where Thickness Swell came in third. Competitive Price was the leading determinant attribute for all of the respondent categories excluding builders, where it came in second. Thickness Swell came in second for all categories not including builders where it was the leading attribute and for trading houses where it was the third. Refer back to Table 29 for a list of the respondent categories and determinant attributes. When comparing the respondent categories, the most notable difference is within the builder's category where Thickness Swell was identified as the most determinant attribute. Given how critical Competitive Price has been for the development of structural panel markets in both North American and Japan, it is interesting that a technical attribute was found to be the most determinant for the builder respondents. 5.1.1.1 Competitive Price Competitive Price has been the key attribute that has shaped the structural panel market and continues to be the most determinant. Competitive Price being identified as the highest determinant attribute was further confirmed by some of the respondent's comments made during the interviews. The majority of the respondents felt that OSB must have a competitive price to increase in sales confirming the results of the attribute analysis. Maruyoshi Co. Ltd. (in Ijtt., 7 Sept. 1998), a trading house, felt that "in the long-term OSB has to be a lower price for it to become a panel of choice. If the price is equivalent or higher, users will prefer plywood." This is consistent with the product life cycle which states that a new product must provide an incentive that overcomes any risk or hassle that is associated with changing products, also known as switching costs (Levitt 1965). According to Venichu (in Ijtt., 8 Sept. 1998), a distributor, "quality was the most important aspect when considering a building material; however, nowadays the most important issue to customers' is price." Mitsubishi (in Ijtt., 8 Sept. 1998), a trading house, felt that "OSB had acquired its current market share by offering a competitive price." Finally, Nichimen (in Ijtt., 9 Sept. 1998), a major trading house, "felt that good quality was required to get into the market, while competitive price will decide which product is purchased." 66 5.1.1.1.1 Changing From Plywood To OSB How big of a change is it for Japanese panel customers to switch from plywood to OSB? This is an important question, as it will ultimately determine how attractive OSB's group of attributes must be for customers to overcome switching costs. Iwatani, a Japanese homebuilder, has not used OSB because the price difference is not attractive enough to warrant a change. Iwatani (in Ijtt., Sept. 1998), explains the issues changing from OSB to plywood in the following. "The change from plywood to OSB in Japan is a significant one that goes right to the heart of the end customer, the homebuyer. A home purchase is a very big deal in Japan. Many of Iwatani's customers visit their homes during construction and it is difficult to explain OSB. Plywood has been used for years in Japan, traditionally for appearance grade in post and beam construction. Now that sheathing is starting to be used for structural purposes, the customers still view it as an appearance grade building material, via experience. However, they do understand that it is reasonable to use plywood in home construction. Thus, if OSB were to be used for sheathing, it would not only require the explanation of OSB it would also require a complete explanation as to the function of plywood. The customer would either need to be convinced that plywood is for structural purposes and that a change in methods is important, or convinced that OSB is beautiful. Both approaches would require a difficult educational process." 5.1.1.1.2 Product Differentiation Obviously, the most important issue for suppliers is price given that it determines profitability and overall competitiveness. The possibility of increasing panel prices through differentiation is a central issue for panel suppliers as they design marketing strategies. Historical panel product life cycles would indicate that significant growth in market share is unattainable without a competitive price. Nevertheless, one of the areas to explore, based on the results of the determinant attribute analysis is the builder market segment, as builders were the only group that did not find Competitive Price to be the most determinant attribute. Possibly this segment is different because builders directly use the structural panels that they purchase and thus have a different view of competitive price/cost. Total cost to a builder is the total cost of the home that they are constructing as opposed to just the cost of the panels. For instance, if a builder found that they had a problem with a panel that resulted in a claim or extra construction labor to replace the panel, it would cost the builder more than any savings that they would have incurred by obtaining a lower priced panel. The home builder Shokusan Jutaku (in Ijtt., 8 Sept. 1998) stated that "the most important thing is quality then 67 price, if a panel is priced low but the quality is not good the total cost will far surpass the cost of the panels." Based on the determinant attribute analysis it may be possible to create a relationship with builders that is less price sensitive than with trading houses and distributors. Although it should be noted that builders did find Competitive Price to be the second most determinant attribute as many structural panel producers are probably capable of supplying a high quality panel for a competitive price. 5.1.1.1.3 Exchange Rates A considerable challenge with supplying a competitive priced panel to the Japanese market is the influence of exchange rates. Japan's structural panel market is based almost entirely on imports as even domestic panels are produced with imported logs, and correspondingly price competitiveness for all panel products are dependent on exchange rates. Additionally, according to Chris Gaston (in Ijtt., 6 Mar. 2001), a researcher focused on the Japanese market, Japan is becoming much more of an open marketplace and often sources suppliers based on exchange rate advantages. Anticipating exchange rate changes is difficult because it takes into account world trade effected by foreign policy and the overall economic performance of different countries (Lipsey et al. 1988). Exchange rates are difficult if not impossible to predict, particularly in developing countries, which are often subjected to large changes in currency values, even in the short-run. Over the long-term the average value of exchange rates will depend on their purchasing power parity (PPP) exchange rate. The PPP exchange rate, is the level of exchange that holds two countries relative price levels the same when measured in a common currency. However, in practice, the fluctuation in exchange from the PPP rate has been wide and lasted for long periods of time causing substantial market discrepancies (Lipsey et al. 1988). However, over time the self-adjusting mechanisms, i.e. arbitrage opportunities and individual's actions, will adjust foreign exchange to eliminate these inefficiencies and producers should focus their marketing efforts where they can compete on a sustainable basis. 5.1.1.2 Thickness Swell Thickness Swell is a technical attribute that is a measure of the thickness alteration that occurs when a panel is subjected to moisture and was identified as the second most determinant attribute, refer back to Table 29. Thickness Swell was identified as a determinant attribute because of the difference in 68 performance between the competing panel products. "OSB performs the worst on thickness swell, Canadian Structural plywood and domestic softwood plywood are ok and hardwood plywood is the best", (Shokusan Jataku, in lift., 8 Sep. 1998). OSB's poor performance on thickness swell, when compared to plywood panels, is recognized in the market place and is perceived as a problem. Mitsubishi (in Ijtt., 8 Sep. 1998) feels that "OSB has a problem with thickness swell and that builders won't use it for floors." Shingu Shoko (in Ijtt., 10 Sep. 1998) stated in their interview that "thickness swell is the number one concern with OSB due to its ability to cause claims. Thickness Swell is the single biggest challenge that must be overcome before OSB will increase in market share." Thickness Swell is not just a perceived problem, as discussed in section 2.4.2.5, as it has been proven in the laboratory that OSB has a greater amount of swell relative to plywood. There are two potential solutions, one is to obviously change the manufacturing process to create a panel that swells less. Alternatively, it is interesting to consider that OSB has the same thickness swell disadvantage to plywood in North America as it does in Japan, but has still managed to claim the majority share of the total North American market. 5.1.2 Panel Size Panel Size was the third most determinant attribute for the all respondent, distributor and distributor and trading house categories, refer to Table 29. Panel Size was the fourth determinant attribute for trading houses and unexpectedly was non-determinant for the builder's category. Panel Size was non-determinant for the builder category as only 4 of the 10 builders surveyed chose Panel Size as an influential attribute and; therefore, was not considered in the determinant analysis. Panel Size's non-determinance for builders is possibly due to the fact that by the time the panels are offered to the builders, they are the sizes specified. For example, if a panel must to be cut to meet a builder's specification, it would have been considered prior to offering the panel to the builder. Panel Size was a determinant attribute, excluding the builder's category, because of the differences in sizing capabilities of plywood and OSB. Due to the manufacturing process, plywood is only offered in the standard 3 and 4 foot widths in a set number of lengths, while OSB is offered in a variety of sizes, all the way up to 8 by 24 foot sized panels. Accordingly, OSB has a much greater sizing flexibility than plywood which can be a market place advantage over other types of panels and accounts for the differences and determinance. Tostem (in Ijtt., 7 Sept. 1998) felt that OSB's larger size is its biggest strength. Similarly, Panel Size was found as a determinant attribute for the Japanese market by research that was carried out in 1990 (Ainsworth 1995). However, Panel Size was found to be determinant due to OSB's sizing limitations as 69 opposed to today's sizing flexibility. At the time, OSB was only produced in 4 foot widths and if it was supplied in 3 foot widths, the 1 foot off-cut was included in the price or in the shipment. Shimada (in Ijtt., 10 Sept. 1998) recalls that they originally were forced to sell 1 by 6 foot off cuts of OSB and they managed to create some 1 foot markets, but it was difficult as the customers had to be persuaded to take the product. In 1994, the first OSB facility capable of producing a 3 foot panel began production and the OSB sizing limitation was overcome. Currently, there are approximately 4 different OSB producers that can efficiently supply 3 foot wide panels to Japan. Maruyoshi (in litt., 7 Sept. 1998), recalling OSB's beginnings in Japan, remembers when OSB was only available in 4 by 8 foot sizes, the market was very limited and now with 3 by 6 foot sizes and a competitive price the potential market for OSB is large. 5.1.3 Shipment Arrives In Good Condition Shipment Arrives In Good Condition was the fourth most determinant attribute. It was ranked fourth when considering all of the respondents and also by distributors and trading house categories. The building respondents ranked it as the third most determinant attribute while trading houses found it to be fourth, refer to Table 29. Shipment Arrives In Good Condition was found to be a determinant attribute as panels that were imported were subject to damage during ocean transportation, which was not the case for domestically produced panels. In addition to incurring more transit damage it is more difficult to settle claims on imported panels as an overseas producer, opposed to a domestic supplier, must be dealt with. Venichu (in Ijtt., 8 Sept. 1998), commented that "damaged panels are a problem, on average, out of 70 pieces of OSB, 3 pieces [or 4.3 percent] are damaged. But Venichu can not afford the additional time and effort to claim for the damaged panels, so they are discounted and sold. The solution is to take this into consideration when calculating the price for OSB." In addition to providing less damaged panels, local producers have the ability to immediately replace any damaged panels, something that is more difficult for an overseas supplier. As clearly seen Shipment Arrives In Good Condition can effect the Competitive Price attribute. For example, Mitsubishi (in Ijtt., 8 Sept. 1998) commented that if the price of OSB is the same as plywood, OSB is more expensive because of additional fall down from transit damage. During the research in Japan, the respondents concerns for OSB's damage in transit was over and above both domestic and imported plywood. As such, some additional investigation was done in an attempt to quantify the differences between plywood and OSB. A visit to the Kawasaki port was made to view panel imports of plywood and OSB. A third party called the All Nippon Checkers Committee (ANCC) audits 70 imports of structural panels and records any damage. From ANCC's historical records, approximately 10 percent of all structural panels delivered to the port are damaged in transit. The breakdown of the damage is shown below in Table 31. Table 30. Breakdown of in transit panel damage Source Damage Percent Forklift (Scratching) 40% One Bundle Hitting Another (Broken Corners) 50% Moving Straps (Loose Panels) 10% 100% Source: ANCC Both plywood and OSB's bundles are, on average, damaged 10 percent, but it is estimated that the hardwood plywood panels only receive 1 percent damage where OSB receives substantially more, possibly even the full 10 percent. The difference is in product packaging. As illustrated in Figure 30, Indonesian and Malaysian hardwood plywood's packaging is vastly superior to OSB's. Hardwood plywood packaging includes 1-inch boards for the bottom base, length and width strapping, metal corner protectors, plastic covers and veneer coverings on all of the five exposed sides. Figure 30. Imported hardwood plywood packaging. Kawasaki port, Japan, 1998. North American OSB producers must consider how to compete with this high quality packaging for other imported panels. Packaging OSB on an equivalent basis and incurring the additional labor and material costs would almost certainly make equivalent packing unfeasible for North American producers. However, one of the weaknesses which was noted by the Japanese structural panel customers is the difficulty and 71 environmental concerns of disposing the extra packaging. The environmental concerns to the additional packaging could possibly be an avenue to investigate when developing a competitive strategy for OSB. 5.1.4 Long-term Supply Long-term Supply was identified as a determinant attribute in all five of the respondent groups. It was ranked as high as third for builders and as low as eleventh for distributors, refer to Table 29. The general feeling for most producers is that long-term fiber supply is an issue for Indonesia and Malaysia and that alternative fiber supplies must be considered. OSB was viewed as the best product for long-term as the fiber supply supporting production is viewed as environmentally responsible and managed on a sustainable basis. One customer, Bankyo, named one of the producer's OSB panel as N21, standing for Natural 21s t Century, conveying the meaning that OSB will be available on a long-term basis throughout the century (in U., 9 Sep 1998). See Figure 31, which is a copy of Bankyo's N21 OSB brochure. Figure 31. Bankyo floor, N21 OSB brochure. Source: Bankyo Floor, 1998. 5.1.5 Formaldehyde Emiss ions Formaldehyde Emissions is an attribute that has become important to Japanese panel customers relatively recently. Room contamination, in reference to indoor air quality, has become a social issue in Japan and as a result, so has formaldehyde emissions from building products. Problems with indoor air quality has become known as sick house syndrome and the Japanese Ministry of Health and Welfare has announced guidelines on indoor air contamination in regards to toxic building materials emissions (Anonymous 2000). 72 Thus, Formaldehyde Emissions, which has been named as one of the concerns for building products, has become an important issue when selecting housing materials as home owners are concerned with avoiding sick house syndrome. 5.1.6 Environmental Forestry Pract ices Environmental Forestry Practices was the seventh most determinant attribute for the all respondent, trading house and distributor and trading house categories, refer to Table 29. The concern with environmental forestry practices is tied in with the Long-term Supply attribute. The Japanese panel customer's feel that if producer's do not have sustainable forestry practices, their ability to supply panels in the long-term will be limited and correspondingly are apprehensive about basing their home design on that panel. 5.1.7 Other Determinant Attr ibutes Other attributes that were found determinant for the all respondent category were Uniform Panel Thickness, Linear Expansion, On Time Delivery, Overall Quality, Ability To Fill Rush Orders, Panel Strength and Manufacturer's Awareness of Customer Needs. Refer to Table 29 for additional attributes that were found to be determinant for the other respondent categories. 5.2 DETERMINANT ATTRIBUTE C O M P A R I S O N BY ATTRIBUTE C A T E G O R Y To provide additional insight, determinant attributes were identified for each of the different product attribute categories, technical, marketing and service. When developing an overall marketing strategy each of the attribute categories should be considered. For instance, sen/ice attributes appear to be the least important of the determinant attributes; however, a comprehensive marketing plan will certainly provide a strategy for service. Additionally, service attributes are often the ones that individual producers can have the most effect on, as opposed to attributes like Competitive Price which is highly dependent on exchange rates. Therefore, for market strategy development, it is worthwhile to know which attributes are determinant for each attribute category. The shaded boxes in Table 30 represent the attributes that were found to be determinant. 73 Table 31. Determinant attribute results, comparison by attribute category Rank Technical Rank Marketing Rank Service 1 "Thickness Swell 1 "Competitive Price 1 "On-time Delivery 2 Formaldehyde Emissions 2 Panel Size 2 "Ability To Fill Rush Orders 3 Uniform Panel Thickness 3 Shipment Arrives In Good Condition 3 Manufacturer's Awareness of C. Needs 4 Linear Expansion 4 Long-term Supply 5 Panel Strength 5 Environmental Forestry Practices 6 Overall Quality *Found to be statistically above the average determinance level. 5.3 R E S E A R C H LIMITATIONS AND S U G G E S T I O N S 5.3.1 Judgement Sample The preceding research has provided insight into the attributes of a structural panel that determine purchasing behavior of Japanese customers. A potential weakness of the research is its inability to make statistically sound inferences about the total population of Japanese structural panel customers. The research, as described in Section 3.1.1, selected the companies for interviewing by taking a judgment sample and, correspondingly, there was no element of randomness in the sampling. Therefore, on a statistical basis the results cannot be used to infer back to the entire population of structural panel customers because the data collection was not based on a statistically sound sampling method. Other simple sampling methods that could have been considered are simple random sampling, unequal probability sampling and stratified random sampling. By introducing randomness to the sampling technique it would be possible to make statistically sound inferences back to the total population of Japanese structural panel customers. Nevertheless, the benefits gained by using a random sampling method would likely not offset the benefits lost by not utilizing a judgment sample. Without a judgment sampling there would be no personal relationships with the companies being interviewed and potentially the validity of the data could be affected. Conceivably, respondent companies would have been reluctant to share the information requested during the interviews with an unknown entity. Additionally, without a personal relationship it would be difficult to gain access to the senior personal responsible for panel purchasing. Lastly, by focusing on the Interex and Canfor contacts, the respondents 74 were biased towards the purchasing of imported panels. This point was actually a positive aspect for North American producers that consider Japan an export market, as the results would be more pertinent to panel imports. 5.3.2 Sample Size The sample size was limited by budgetary and personnel constraints which limited the interview time available in Japan, refer to Section 3.1.1. Obviously, this was not a statistically sound method for determining sample size, however; the constraints were real and binding. For further research it would be strongly suggested that a larger sample be taken. A larger sample size would allow the individual respondent categories and their differences to be thoroughly investigated. A large sample size would also provide the options of using different statistical methods for interpreting the data, such as multidimensional scaling and cluster analysis. 75 6 CONCLUSION The comparison of the product life cycles of the structural panel markets in North America and Japan has demonstrated that OSB has the potential for significant growth. OSB was first introduced to Japan in the early 1990's and since that time has been going through the development stage of the product life cycle. The development phase of OSB in North America lasted approximately 12 years before it began rapid expansion. In Japan, OSB has been in the market for approximately 11 years and could be poised to enter the growth phase of the product life cycle. In addition, OSB's growth in Japan was considered to have the potential to exceed that of North America, as many of the early difficulties experienced with OSB should have been overcome in the North American development stage. To reach the growth stage of the PLC, OSB in Japan will have to perform well on the determinant attributes for Japanese structural panels. The determinant attributes for structural panels in Japan have been identified as, in order of determinance, Competitive Price, Thickness Swell, Panel Size, Shipment Arrives In Good Condition, Long-term Supply, Formaldehyde Emissions, Environmental Forestry Practices, Uniform Panel Thickness and Linear Expansion. The determinant attribute analysis and the historical development of structural panel markets in both North America and Japan, strongly support Competitive Price as the most determinant attribute. Whether North American produced OSB can out perform competing panels on the determinant attributes is unclear. Unfortunately, the most unclear attribute is also the most determinant, Competitive Price. The biggest uncertainty for Competitive price is foreign exchange. For example, in January of 1996 12mm thick, 3 by 6 foot OSB was selling for US $320 versus the competing hardwood plywood panel at US $443, creating a $123 advantage for OSB. It is expected that a price advantage of this magnitude would push OSB into the growth stage. However, with changing market forces and a large swing in the Indonesian exchange rate the OSB pricing advantage was changed into a US $17 disadvantage, refer to Table 31. Table 32. Example of plywood and OSB price differentials US$CIF Hardwood Plywood OSB Difference 11.4mm 3x6 12.0mm $443.48 $320.00 $123.48 $434.70 $315.00 $119.70 $232.69 $250.00 ($17.31) $304.15 $256.00 $48.15 Source: Interex 1999. January 1996 April 1997 August 1998 December 1998 76 The second most determinant attribute for structural panels in Japan is Thickness Swell. OSB's performance on Thickness Swell relative to other structural panels is clearly inferior. It would also appear that OSB's performance on the other technical determinant attributes are either not as good or are unclear. The other technical attributes include Formaldehyde Emissions, Uniform Panel Thickness and Linear Expansion. Furthermore, OSB does not perform as well on the attribute of Shipment Arrives In Good Condition. However, OSB has superior performance on Panel Size, Long-term Supply and Environmental Forest Practices. Table 33 reviews an estimate of OSB's performance relative to other competing panel products. OSB's performance estimates are based oh the phase-two questionnaires and the respondents comments made during the personal interviews. Table 33. OSB's estimated performance relative to other panel products Determinant Attribute OSB's Performance Relative To Other Panel Products (In Order of Determinance) (Superior, Unclear, Poorer) Competitive Price Unclear Thickness Swell Poorer Panel Size Superior Shipment Arrives In Good Condition Poorer Long-term Supply Superior Formaldehyde Emissions Unclear Environmental Forest Practices Superior Uniform Panel Thickness Poorer Linear Expansion Poorer Overall, based on the attribute analysis it would appear that OSB will continue in its current development stage until its relative performance on some of the determinant attributes is enhanced. Based on the product life cycle analysis Japan represents an excellent potential for OSB, but substantial growth is not imminent like it was in the North American market. An important factor that could enhance OSB and other substitute panel's ability to penetrate the Japanese market would be a substantial increase in the demand for structural panels in Japan. This increase in Japanese demand could become a reality as the enhanced performance of homes constructed with structural panels is becoming recognized. Finally, OSB producers have the ability to adapt their product and optimize marketing strategies to effect OSB's progression through the PLC in Japan. North American OSB producers should adapt their marketing and development strategies to enhance the performance on the determinant attributes. 77 7 LITERATURE CITED Ainsworth, M.D. 1995. Marketing Oriented Strand Board In Japan. Masters. Thesis, University of British Columbia, Vancouver, B.C. Alpert, A.I. 1971. Identification of Determinant Attributes: A Comparison of Methods. J. Mar. Res. Vol. Vlll: 184-91. Alpert, A.I. and J.H. Myers. 1968. Determinant Buying Attitudes: Meaning and Measurement. J. Mar. Vol. 32:13-20. Anderson, Jr, T.W., E.P. Cox III, and D.G. Fulcher. 1976. Bank Selection Decisions and Market Segmentation. J. of Mar. (40): 40-45. Anonymous. 2000. The Ministry of Welfare Setting Guideline on Indoor Air Contamination. Jpn. Lumber Report. Nov. 10., No. 334:1. APA The Engineered Wood Association. 2000. Japan Market Report - 2000. Sept.22. APKINDO. 1991. Strategies to Control Japanese Plywood Markets. Jpn. Lumber J. Vol 32, No. 14:12. Armacost, R.L. and J.C. Hosseini. 1994. Identification of Determinant Attributes Using the Analytic Hierarchy Process. J. Academy Mar. Sci. Vol 22, No. 4: 383-392. Bass, F.M., and W. Wilkie. 1973. A comparative analysis of attitudinal predictions or brand preference. J. Mar. Research 10(3): 262-269. Bearden, W.O. 1977. Determinant attributes of store patronage: Downtown versus outlying shopping centers. J. Retailing 53 (2): 15-22. Beckman, M.D., D.L.Kurtz and L.E. Boone. 1982. Foundations of Marketing, Second Canadian Edition. Holt, Rinehart and Winston of Canada, Limited, Toronto. 262-269. Cohen, D. H. and C. Gaston. 1998. Japan's Value added Market: Wood product attributes and competition - attribute analysis,: engineered structural members. Forintek Canada Corp Report. Prepared for Forest Renewal BC (Victoria, BC) 27 pgs. Cohen, D. H. and C. Gaston. 2001. Trends and Changes in Japanese Building Requirements, Export Opportunities. CMHC. Government of Canada. 7 8 Cohen, D. H., I. Macdonald and R. Kozak. 2001. The Japanese distribution system for finished building products: in press with Forest Renewal BC. Cour. R.M. 1955. Plywood Age. Binford's and Mort, Portland Oregon. Dhalla, N. K. and S. Yuspeh. 1976. Forget the Product Life Cycle Concept! Harvard Business Review. Jan/Feb. No. 76104. Doyle, P. and J. Saunders. Spring 1985. Market Segmentation and Positioning in Specialized Industrial Markets. J. Mar. Vol. 49: 24-32. Elias, E.G. 2000. Indonesian Wood Based Panel Industry. APA - The Engineered Wood Association: 25 pgs. Fassnacht, D.L. 1964. The new southern pine plywood industry. For. Prod. J. (1):21-22. Food and Agriculture Organization (FAO). 2001. FAO Stat Forestry Data, Forestry Trade Data. United Nations. Forebes, C L , S.A. Sinclair, R.J. Bush, and P.A. Araman. 1994. Influence of product and supplier attributes on hardwood lumber purchase decisions in the furniture business. For. Prod. J. 44(2): 51-56. Franke, G.R. 1985. Evaluating measures through data quantification: Applying dual scaling to an advertising copy test. J. Bus. Research 13 (1): 61 -69. Green, P.E. and F.J. Carmone. 1978. Some methodological alternatives in the analysis of life style data. J. Econ. And Bus. 30(2):158-161. Heeler, R.M., C. Okechuku, and S. Reid. 1979. Attribute Importance: Contrasting Measurement. J. Mar. Research. 16(2):60-63. Howell, D. C. 1987. Statistical methods for psychology. 2 n d ed. Duxbury Press, Boston, MA. Interex Forest Products Ltd. 1999. Japan OSB Market Analysis. April. Vancouver, B.C. Interex Forest Products Ltd. 2000. OSB Market Report. Vancouver, B.C. Interex Forest Products Ltd. 2001. Japan Monthly Market Report. January. Vancouver, B.C. Japan Wood-Products Information And Research Center (JAWIC). January, 2001. Wood Supply And Demand Information Service. Kleinbaum, D.G., L.L. Kupper, and K.E. Muller. 1988. Applied Regression Analysis and Other Multivariable Methods. 2 n d ed. PWS-Kent Publishing Company Boston, pp. 344-347. 79 Kozak, A. 1966. Research Notes, Multiple Correlation Coefficient Tables Up To 100 Independent Variables. No. 57. Kozak, A. 1998. Lecture in Advanced Biometics, Forestry 430. Oct 7, Nov 17. University of British Columbia, Vancouver, B.C. Lehmann, D.R., and J. Hulbert. 1972. Are Three-Point Scales Always Good Enough? J. Mar. Research 9: 444-446. Levitt, T. 1965. Exploit the Product Life Cycle. Harvard Business Review. Nov. Dec. 81-94. Lipsy, R.G., D.D. Purvis and P.O. Steiner. 1988. Macroeconomics, Sixth Edition. Harper & Row, Publishers, New York. 876-895. Lumpkin, J. R., B. A. Greenberg, and J. L. Goldstucker. Summer 1985. Marketplace Needs of the Elderly: Determinant Attributes and Store Choice, J. of Retailing. (61) 75-105. Marshall, P. 1998. Lecture in Sampling Methods, Forestry 431. 7 Jan. University of British Columbia, Vancouver, B.C. Moriarty, R.T., and D.J. Reibstein. 1986. Benefit segmentation in industrial markets. J. Bus. Res. 14 (6) 463-486. Parasuraman, A. 1983. Debranding: a product strategy with profit potential. J. Business Strategy 4(1):82-87. Porter, M.E., 1980. Competitive Strategy. The Free Press., New York. 156-184. Resource Information Systems Inc. (RISI). 2001. Wood Products Year Book 2000. Rosenberg, N., P. Ince, K. Skog, and A. Plantinga. 1990. Understanding the adoption of new technology in the forest products industry. For. Prod. J. 40(10):15-22. Schaninger, CM., and W.C. Buss. 1986. Removing response-style effects in attribute-derminance ratings to identify market segments. J. Bus. Res. 14 (3): 237-252. Seward, K.E., and S.A. Sinclair. 1988. Retailers' perceptions of structural panel attributes and market segments. For. Prod. J. 38(4):25-31. Shook, R. 1999. Forecasting adoption and substitution of successive generations of structural wood panel products in the United States. For. Sci. 45 (2). Shook, R., W.R. Turner and I.L. Eastin. 1998. Adoption, Diffusion, And Substitution Of Structural Wood Panels. CINTRAFOR Working Paper 65, University of Washington. 80 Sinclair, A. and B.G. Hansen. 1993. The relationship between purchase decisions and quality assessment of office furniture. Wood and Fiber Sci. 25(2): 142-152. Sinclair, A. and E.C. Stalling. 1990. How to Identify Differences Between Market Segments With Attribute Analysis. Ind. Mar. Management. (19) 31-40. Sinclair, S.A., 1991. Forest Products Marketing. McGraw-Hill, Inc., New York. 37-57,127-139. Spelter, H. 1994. Capacity, Production, and Manufacturing of Wood-Based Panel in North America. USDA For. Service., For. Prod. Lab. General Tech. Rep., Madison, Wisconsin. Spelter, H. N., D.B. McKeever and I. Durbak. 1997. Review of wood-based panel sector in United States and Canada. Gen. Tech. Tep FPL-GTR-99. Wadison, Wl: U.S. Department of Agriculture, Forest Service, Forest Products Laboratory. Statistics Canada. 2001. Trade Statistics, Database. Sweitzer, R.W. 1975. Determinant Attribute Segmentation. J. of the Academy of Marketing Sci. Vol 3. No. 1.85-98. Thomas, J.P. and R.H. Waterman Jr. 1982. In Search of Excellence. New York. Harper & Row. Trinka, M.W., S.A. Sinclair, and T.C. Marcin. 1992. Determinant Attribute Analysis: A Tool For New Wood Product Development. Wood and Sci. 24(4): 385-391. Walpole, R.W., 1982. Introduction to Statistics. Macmillan Publishing Co., Inc., New York. 396-397. 81 A P P E N D I X I: Interview Int roduct ion Letter AINSWORTH LETTERHEAD September X, 1998 Dear Structural Panel Customer: (Letters to be personalized) We are currently conducting market research on the wood products industry in Japan and we would appreciate your help. We are participating in a study on behalf of Ainsworth Lumber directed at the structural panel market. Please note, any information that you provide will be held in strict confidence within the participating organizations. The organizations conducting this market research are Ainsworth Lumber Co. Ltd. and the University of British Columbia (UBC). UBC is one of Canada's major university's and has one of North America's premier forestry faculties which conducts a significant amount of research in the forest products field. The individuals carrying out the research from UBC are Dr. David Cohen and Derek Welbourn. Dr. Cohen is a professor in the Wood Science Department specializing in forest products marketing and the Japanese market. Derek Welbourn is a graduate student in the faculty of forestry in the Wood Science Department and is supervised by Dr. Cohen. Previously, Mr. Welbourn was in the Marketing & Business Development department at Ainsworth and is currently on leave to conduct his studies. In order to gain insight into the structural panel market, we have elected to study all competing structural panels, including plywood and Oriented Strand Board. The goal of the research is to understand the trends and changing conditions in the Japanese structural panel market. This information will allow Ainsworth to focus on the most important issues when servicing the Japanese panel market. We feel that by conducting market research studies we will be in a better position to understand and service your needs. Please fill out the attached questionnaire and we will have further discussions upon our visit. Thank-you for participation in our market research. Sincerely, Michael Ainsworth Vice President Ainsworth Lumber Co. Ltd. Interex/Canfor Dr. David Cohen Associate Professor University of British Columbia 82 APPENDIX II: Phase-one Questionnaire 83 Ainsworth Lumber Co. Ltd / University of British Columbia Market Research - Japan UNIVERSITY OF BRITISH COLUMBIA AND AINSWORTH LUMBER CO. LTD. STRUCTURAL PANEL MARKET STUDY 1. 3. Overall, how important do you feel each specific attributes group (Technical, Marketing, Service) is when making a structural panel purchasing decision? • Technical Attributes % • Marketing Attributes % • Supplier Service Attributes % Estimate the total annual volume of structural panels purchased by your company in (cubic meters): • 1996 • 1997 • 1998 (anticipated) Estimate the percentage of your wood panels that are imported and domestically produced: • Imported % • Domestic % • Total = 100 % Estimate what percent of your imported wood panels are from the following countries: • Canada % United States % Malaysia Indonesia % % Others (specify) % % % % % % % % % % % Total = 100 Which structural panels do you currently use or consider using (i.e. OSB, C S P etc.) A. % I B. % I C. % I D. % I E. % I f. % I G. 100 % % Please indicate all primary thicknesses of wood panels your company utilizes: 6.4mm % 8.5 mm • % 9.0 mm % 9.5 mm % 11.5mm T8T6rnrrT % 12.0mm % 12.5mm % 15.0mm % 24.0mm % 27.0mm % 28.0mm Other % Other % Other % Other % Other Other % Other % Other Total = 100 _% % Please estimate what percent of the following panel sizes your company purchases: 3 x 6 % 3 x 8 % ( 3 x 9 % 3 x 9 % 3 x 10 % 4 x 8 % j Other % Other % 84 Ainsworth Lumber Co. Ltd / University of British Columbia Market Research - Japan 6. Of the total volume of panel products handled by your firm, what percent are for: • Packaging & Crating % • Traditional Home Construction % • Roof Sheathing % • Floor Sheathing % • Wall Sheathing % 100 % • 2 x 4 Platform Construction % • Roof Sheathing % • Floor Sheathing % . Wall Sheathing % 100 % • Prefabrication Panels % • Pourform % 100 % 7. What do you feel some of the overall strengths and weaknesses with your main type of structural panels? • Strengths • Weaknesses 8. Are there any differences between Structural Panels made in Indonesia and Malaysia? Yes No (check S one) Please Describe: 9. Are there any differences between Structural Panels made in Indonesia or Malaysia and Panels made in Japan with Indonesian or Malaysian logs? Yes No (check S one) Please Describe: 10. What type of housing does your company produce? Please indicate the proportions. Traditional post and beam % 2 x 4 % Hybrid (mix of post and beam and 2 x 4 ) % 100% 11. What was the number of houses your company built in: 1995 1996 1997 85 Ainsworth Lumber Co. Ltd / University of British Columbia Market Research - Japan 12. Of the houses that your company built in 1997, what percentages were wood structural material, steel structural material, concrete or masonry? Wood % Steel % Concrete % Masonry % Other (please specify) % % 13. What size are the houses your company builds? Please indicate the proportions. < 30 tsubo % 3 1 - 4 0 tsubo % 41 - 50 tsubo % > 50 tsubo % 86 Ainsworth Lumber Co. Ltd / University of British Columbia Market Research - Japan Attribute Definit ions: CLASS OF PRODUCT ATTRIBUTES: • Technical Attributes • Marketing Attributes • Supplier Service Attributes SPECIFIC ATTRIBUTES: Technical Attributes • Linear Expansion - panel linear size change • Panel Stiffness - (Modulus of Elasticity - MOE) parallel and perpendicular, measured wet and dry in kgf/cm 2 • Panel Strength - (Modulus of Rupture - MOR) parallel and perpendicular, measured wet and dry in kgf/cm 2 • Internal Bond - pounds per square inch, or kgf/cm 2 • Density of Panel - pounds per cubic foot • Nail Withdrawal - kgf • Uniform Panel Thickness - JAS , plus 5% to minus 3% • Physical Appearance - physical appearance of the panel • Formaldehyde Emissions - Formaldehyde that is emitted from the panel • Thickness Swell - % swelling, as per 3 r d party of company standards Marketing Attributes • Competitive Price - what does it cost compared to other products. • Long-term Supply - ability for the panel product to be supplied now and into the future, how secure is it's fiber supply. • Shipment arrives in good condition - the concern of damage in transit. • Product availability - are there a choice of panel suppliers and is the product easy to access • Environmental Forestry Practices - environmentally sound, aware of issues that are becoming international • Overall Quality - overall view of the panel product, given combination of attributes. • Flexibility of panel properties - engineering capabilities to meet different applications • Panel Size - 3 , 4 foot widths or other specialty size supply capabilities • Availability of Thicknesses - is a variety of thicknesses available • Merchandising - appearance from packaging point of view, logo, straps, covers etc. • Presence of company logo and customer logo I name - printed on the panel • Stability of Quality- same quality panel each order Supplier Service Attributes • Manufacturer's Reputation - how the manufacturer is viewed • Accessibility of manufacturer - how easy is it to access the manufacturer • High level of service - prompt service of price quotes and dealings with customer complaints • On-time delivery - lead-time from order-time to delivery, is this time stable and reliable • Manufacturer's knowledge of their products - product knowledge and support material (i.e. providing samples, technical information) • Manufacturer's awareness of customers needs - understanding customer requirements via technical and service issues • Manufacturer's Flexibility - ability to help distributor with special requests • Personal relationship with manufacturer - how important is it be familiar with the supplier and their business • Manufacturers market knowledge - knowledge of markets and trends • Ability to fill rush orders - relatively short lead-time • Ability to fill small orders - manufacturer is willing to deliver small orders • After sales service - Manufacturer follows up to see if expectations are met • Credit terms - flexibility of credit terms • Manufacturer provides product training - does the manufacturer provide training services that are useful 87 Ainsworth Lumber Co. Ltd / University of British Columbia Market Research - Japan P h a s e O n e Quest ionna i re : For each category indicate THE FIVE MOST IMPORTANT ATTRIBUTES. Please add in any important attributes that you feel are missing in the spaces provided at the end of each section. Technical Attributes: Attribute: Linear Expansion Importance Panel Stiffness Panel Strength Internal Bond Density of Panel Nail Withdrawal Uniform panel thickness Physical Appearance Formaldehyde Emissions Thickness Swell Other (please specify). Other (please specify). Other (please specify). Other (please specify) Marketing Attributes: Attribute: Competitive Price Importance Long-term Supply Shipment Arrives In Good Condition Product Availability Environmental Forest Practices Overall Quality Flexibility Of Panel Properties Panel Size Availability of Thicknesses Merchandising Presence of Company Logo and Customer Logo / Name Other (please specify) ] Other (please specify). Other (please specify) Other (please specify). Supplier Service Attributes: Attribute: Manufacturer's Reputation Importance Accessibility of Manufacturer High Level of Service . . . On-time delivery Manufacturer's Knowledge of Their Products Manufacturer's Awareness of Customers Needs Manufacturer's Flexibility Personal Relationship With Manufacturer Manufacturer's Market Knowledge Ability To Fill Rush Orders . . . . Ability To Fill Small Orders After Sales Service Other (please specify). Other (please specify). Other (please specify). Other (please specify), APPENDIX III: Phase-two Questionnaire 89 Date of Interview Company Interviewed Interview Number UNIVERSITY OF BRITISH COLUMBIA AND AINSWORTH LUMBER CO. LTD. STRUCTURAL PANEL MARKET STUDY When selecting a wood based panel product for use in home construction, which product attributes do you feel are the most important? (Please rank each characteristic on a scale of 1 = of no importance to 5 = critically important.) Attributes Top 5 Technical Attributes: 1. No Some Very Importance Importance Important Important Cri (1) (2) (3) (4) (5 I 2. ( ) ( ) ( ) ( ) ( ) ! 3. ( ) ( ) ( ) ( ) ( ) I 4. ( ) ( ) ( ) ( ) ( ) I ical Top 5 Marketing Attributes: 1. 2. 3. 4. 5. Top 5 Supplier Service Attributes: 1. 2. 3. 4. 5. 90 2. How much difference do you feel there is between panel product A, B, C, D, E, F & G (specified in phase 1), for each of the product attributes? (Please check the one space that you feel best describes the differences between the products.) Attributes Very Very Similar Similar Different Different (1) (2) (3) (4) Top 5 Technical Attributes: 1. ( ) ( ) ( ) ( ) I 2. ( ) ( ) ( ) ( ) I 3. ( ) ( ) ( ) ( ) I 4. ( ) ( ) ( ) ( ) I 5. ( ) ( ) ( ) ( ) Top 5 Marketing Attributes: 1. ( ) ( ) ( ) ( ) I 2. ( ) ( ) ( ) ( ) I 3. ( ) ( ) ( ) ( ) I 4. ( ) ( ) ( ) M I 5. ( ) ( ) ( ) ( ) Top 5 Supplier Service Attributes: 1. ( ) ( ) ( ) ( ) I 2. ( ) ( ) ( ) ( ) I 3. ( ) ( ) ( ) ( ) I 4. ( ) ( ) M • M I 5. ( ) ( ) ( ) ( ) 91 How do each of the panel products you consider when purchasing rate in terms of your top 5 product attributes? For each product, indicate whether the product possesses that attribute to a High degree (4), Considerable degree (3), Limi ed degree (2), or Not At All (1). Top 5 Technical Attributes For Each Product (A to G The Product Possess The Technical Attribute To What Degree 1. Product A Not All Limited Considerable High Degree Degree Degree (2) (3) (4) Product B Product C Product D Product E Product F Product G Product A Product B Product C Product D Product E Product F Product G Product A Product B Product C Product D Product E Product F Product G Product A Product B Product C Product D Product E Product F Product G 92 5 Product A ( ) ( ) ( ) ( ) Product B ( ) _-_( ) ( ) ( ) Product C ( ) ( ) ( ) ( ) Product D ( ) ( ) ( ) ( ) | Product E ( ) ( ) ( ) ( ) Product F ( ) ( ) ( ) ( ) I Product G ( ) ( ) o o Top 5 Marketing Attributes For Each Product (A to G): The Product Possess The Technical Attribute To What Degree ^ 1. Not At All (D Limited Degree (2) Considerable High Degree Degree (3) (4) Product A ( ) ( ) ( ) ( ) Product B ( ) ( ) ( ) ( ) | Product C ( ) ( ) ( ) (") Product D ( ) ( ) ( ) ( ) | Product E ( ) ( ) ( ) ( ) Product F ( ) ( ) ( ) ( ) | Product G ( ) ( ) ( ) ( ) 2. Product A ( ) ( ) ( ) ( ) Product B. ( ) ( ) ( ) ( ) ] Product C ( ) ( ) ( ) ( ) Product D ( ) ( ) ( ) ( ) | Product E ( ) ( ) ( ) ( ) Product F ( ) ( ) ( ) ( ) I Product G ( ) ( ) ( ) ( ) 3. Product A ( ) ( ) -.<) O Product B ( ) ( ) ( ) ( ) | Product C ( ) ( ) ( ) ( ) Product D ( ) ( ) ( ) ( ) I Product E ( ) ( ) ( ) ( ) L Product F ( ) : ( ) ( ) ( ) I Product G ( ) ( ) 93 4 Product A ( ) ( ) ( ) ( ) 1 Product B ( ) ( ) ( ) ( ) I Product C ( ) ( ) ( ) ( ) [ Product D ( ) ( ) ( ) ( ) I Product E ( ) ( ) ( ) ( ) 1 Product F ( ) ( ) ( ) ( ) I Product G ( ) ( ) ( ) ( ) 5. Product A ( ) ( ) ( ) ( ) | Product B ( ) ( ) ( ) ( ) I Product C ( ) ( ) ( ) ( ) 1 Product D ( ) ( ) ( ) ( ) I Product E ( ) ( ) ( ) • ( ) | Product F ( ) ( ) ( ) ( ) I Product G ( » ( ) o o Top 5 Supplier Service Attributes For Each Produc t (A to G) The Product Possess The Technical Attribute To What Degree • / 1. Not W All Limited Degree (2) Considerable High Degree Degree (3) (4) Product A ) ( ) ( ) ( ) I Product B ) ( ) ( ) ( ) I Product C ) ( ) ( ) ( ) | Product D ) ( ) ( ) ( ) I Product E ) ( ) ( ) ( ) Product F ) ( ) ( ) ( ) I Product G ) ( ) ( ) ( ) 2. Product A ) ( ) ( ) ( ) . Product B ) ( ) ( ) ( ) I Product C ) ( ) ( ) ( ) Product D ) ( ) ( ) ( ) I Product E ) ( ) ( ) ( ) Product F ) ( ) ( ) ( ) I Product G ( ) ( ) ( ) ( ) 94 3 Product A ( ) ( ) ( ) ( ) 1 Product B • ( ) ( ) ( ) ( ) I Product C ( ) ( ) ( ) O | Product D ( ) ( ) ( ) ( ) 1 Product E O O O O 1 Product F ( ) ( ) ( ) ( ) I Product G ( ) ( ) ( ) ( ) 4. Product A ( ) ( ) ( ) ( ) I Product B ( ) ( ) ( ) ( ) 1 Product C ( ) ( ) ( ) ( ) | Product D ( ) ( ) ( ) ( ) 1 Product E ( ) ( ) ( ) ( ) I Product F ( ) ( ) ( ) ( ) 1 Product G ( ) ( ) ( ) ( ) 5. Product A ( ) O O O I Product B ( ) ( ) ( ) ( ) 1 Product C ( ) ( ) ( ) H | Product D ( ) ( ) ( ) ( ) I Product E ( ) ( ) ( ) ( ) i Product F ( ) ( ) ( ) ( ) I Product G ( ) ( ) ( ) ( ) 95 APPENDIX IV: Phase-one Results (Raw Data) Interview Number 10 2.0 3.0 4.0 5.0 Company Classification (Trading House, Distributor, Builder) Distributor Builder T.H. Distributor Builder 1) Which structural panels do you currently use or consider using? Russian Larch 39% 30% 0% 25% 0% Domestic Lauan 30% 0% 55% 20% 0% Imported Lauan 20% 0% 0% 0% 0% CSP 5% 0% 38% 30% 80% SYP 5% 0% 0% 0% 0% OSB 1% .70% 7% 20% 20% DFP 0% 0% 0% 0% 0% MDF/PB 0% 0% 0% 5% 0% 100% 100% 100% 100% 100% 2.) Annual Volume Purchased (M ')? 1996 1,000 16,000 N/A 50,000 N/A 1997 950 6,000 46,000 45,000 N/A 1998 720 11,000 15,500 40,000 13,479 3.) Percent Imported? Imported 50% 70% 45% 50% 100% Domestic 50% 30% 55% 50% 0% 100% 100% 100% 100% 100% 4.) Origin of Volume Imported? Canada 10% 100% 100% 83% 100% United States 10% 0% 0% 0% 0% Malaysia 30% 0% 0% 0% 0% Indonesia 50% 0% 0% 17% 0% Other 0% 0% 0% 0% 0% 100% 100% 100% 100% 100% 5.) Primary Thicknesses (mm)? 6.4 0% 0% 0% 0% 0% 8.5 0% 0% 0% 0% 0% 9.0 0% 2.5% 0% 20% 0% 9.5 50% 55% 50% 35% 11% 11.0 0% 0% 0% 0% 0% 11.1 0% 35% 0% 0% 0% 11.5 0% 0% 0% 0% 0% 12.0 50% 2.5% 50% 0% 0% 12.5 0% 0% 0% 30% 0% 15.0 0% 2.5% 0% 0% 0% 15.5 0% 0% 0% 0% 0% 18.0 0% 0% 0% 5% 0% 18.5 0% 0% 0% 0% 23% 21.0 0% 0% 0% 0% 0% 24.0 0% 0% 0% 0% 0% 25.0 0% 0% 0% 0% 66% 27.0 0% 0% 0% 0% 0% 28.0 0% 2.5% 0% 10% 0% Other n% 0% 0% 0% 0% 100% 100% 100% 100% 100% 6.) Primary Dimensions (feet)? 3 x 6 50% 30% 50% 20% 0% 3 x 8 25% 20% 30% 0% 0% 3 x 9 15% 0% 10% 0% 11% 3x10 10% 30% 10% 5% 0% 4 x 8 0% 0% 0% 0% 66% Other 0% 20% 0% 75% 23% 100% 100% 100% 100% 100% 7.) Panel Usage? Packaging & Crating N/A 0%' N/A 10% 0% Traditional 70% 90% 100% Roof 30% 35% 35% Wall 20% 35% 35% Floor 50% 30% 30% Platform (2x4) 30% 0% 0% Roof 30% 0% 0% Wall 20% 0% 0% Floor 50% 0% 0% Prefabrication 0% 0% 0% Pourform 0_%_ 0%. 0% 96 6J3 7 J 8.0 9.0 10.0 \ Interview Number Company Classification (Trading House, Distributor, Builder) TH. Builder Distributor Distributor Builder 1.) Which structural panels do you currently use or consider using? Russian Larch 0% 0% 0% 0% 0% Domestic Lauan 0% 0% 0% 10% 0% Imported Lauan 97% 0% 0% 0% 0% CSP 1% 0% 65% 30% 100% SYP 0% 0% 0% 0% 0% OSB 2% 0% 30% 60% 0% DFP 0% 0% 5% 0% 0% MDF/PB 0% 100% 0% 0% 0% 100% 100% 100% 100% 100% 2.) Annual Volume Purchased (MJ)? 1996 241,200 N/A 5,000 19,647 6,000 1997 248,000 N/A 5,000 19,211 6,000 1998 247,000 111,789 5,000 23,164 6,000 3.) Percent Imported? Imported 100% 30% 100% 90% 30% Domestic 0% 70% 0% 10% 70% 100% 100% 100% 100% 100% 4.) Origin of Volume Imported? Canada 4% 0% 100% 100% 100% United States 0% 0% 0% 0% 0% Malaysia 66% 16% 0% 0% 0% Indonesia 30% 16% 0% 0% 0% Other 0% 68% 0% 0% 0% 100% 100% 100% 100% 100% 5.) Primary Thicknesses (mm)? 6.4 6% 0% 0% 0% 0% 8.5 0% 0% 0% 0% 0% 9.0 0% 0% 0% 0% 0% 9.5 80% 0% 60% 80% 0% 1.1.0 0% 0% 0% 0% 0% 11.1 0% 0% 0% 0% 0% 11.5 0% 0% 0% 0% 0% •12.0 0% 0% 0% 10% 0% 12.5 7% 0% 18% 5% 100% 15.0 0% 0% 0% 0% 0% 15.5 0% 0% 17% 5% 0% 18.0 7% 0% 0% 0% 0% 18.5 0% 0% 0% 0% • 0% 21.0 0% 100% 0% 0% 0% 24.0 0% 0% 0% 0% 0% 25.0 0% 0% 0% 0% 0% 27.0 0% 0% 0% 0% 0% 28.0 0% 0% 0% 0% 0% Other 0% 0% 5% 0% 0% 100% 100% 100% 100% 100% 6.) Primary Dimensions (feet)? 3x6 27% 0% 10% 60% 35% 3 x 8 10% 0% 5% 8% . 35% 3 x 9 27% 0% 5% 20% 0% 3x10 0% 0% 0% 2% 0% 4 x 8 36% 0% 75% 10% 0% Other 0% 100% 5% 0% 30% 100% 100% 100% 100% 100% 7.) Panel Usage? Packaging & Crating 0% N/A . o% 0% 0% Traditional 3% 10% 60% 0% Roof 0% 50% 40% 0% Wall 0% 0% 50% 0% Floor 0% 50% 10% 0% Platform (2x4) 0% 80% 10% 100% Roof 0% 25% 30% 30% Wall 0% 35% 30% 35% Floor 0% 40% 40% 35% Prefabrication 0% 10% 0% 0% Pourform 97% 0% 30% 0% 97 11.0 12.0 13.0 14.0 15.0 | Interview Number Company Classification (Trading House, Distributor, Builder) Builder TH . T.H. Builder Builder 1.) Which structural panels do you currently use or consider using? Russian Larch 0% 0% 0% 0% 25% Domestic Lauan 0% 0.5% 10% 0% 25% Imported Lauan 0% 86% 72% 0% 0% CSP 0% 5% 0% 33% 35% SYP 0% 0% 0% 0% 0% OSB 100% 8% 3% 33% 15% DFP 0% 0% 0% 33% 0% MDF/PB 0% 1.5% 16% 0% 0% 100% 100% 100% 100% 100% 2.) Annual Volume Purchased (MJ)? 1996 - 433,250 - N/A 75,000 1997 - 256,000 1,175,000 - 56,000 1998 500 200,500 1,035,000 - 50,000 3.) Percent Imported? Imported 100% 70% 90% 100% 50% Domestic 0% 30% 10% 0%. 50% 100% 100% 100% 100% 100% 4.) Origin of Volume Imported? Canada 100% 12% 5% 100% 100% United States 0% 1% 0% 0% 0% Malaysia 0% 38% 80% 0% 0% Indonesia 0% 47% 10% 0% 0% Other 0% 2% 5% 0% 0% 100% 100% 100% 100% 100% 5.) Primary Thicknesses (mm)? 6.4 0% 9.5% 0% 0% 0% 8.5 0% 0% 0% 0% 0% 9.0 0% 0% 0% 0% 10% 9.5 0% 70.2% 50% 0% 35% 11.0 100% 0% 0% 0% 15% 11.1 0% 0% 0% 30% 0% 11.5 0% 3.1% 0% 0% 0% 12.0 0% 16.8% 50% 30% 10% 12.5 0% 0% 0% 0% 0% 15.0 0% 0% 0% 0% 30% 15.5 0% 0% 0% 20% 0% 18.0 0% 0% 0% 0% 0% 18.5 0% 0% 0% 20% 0% 21.0 0% 0% 0% 0% 0% 24.0 0% 0% 0% 0% 0% 25.0 0% 0% 0% 0% 0% 27.0 0% 0% 0% 0% 0% 28.0 0% 0% 0% 0% 0% Other 0%. 0%. 0% 0% 0% 100% 100% 100% 100% 100% 6.) Primary Dimensions (feet)? 3x6 0% 90% 50% 0% 45% 3x8 50% 0% 25% 88% 5% 3x9 50% 0% 25% 12% 5% 3 x 10 0% 0% 0% 0% 0% 4x8 0% 10% 0% 0% 35% Other 0% 0% 0%. 0%. 10% 100% 100% 100% 100% 100% 7.) Panel Usage? Packaging & Crating 0% 20% N/A 0% 0% Traditional 0% 17% 0% 0% Roof 0% 0% 0% 0% Wall 0% 0% 0% 0% Floor 0% 0% 0% 0% Platform (2x4) 0% 17% 100% 100% Roof 0% 0% 33% 30% Wall 0% 0% 33% 35% Floor 0% 0% 33% 35% Prefabrication 100% 17% 0% 0% Pourform 0% 30% 0% 0% 98 16.0 17.0 18.0 19.0 | [Interview Number Company Classification (Trading House, Distributor, Builder) Builder Builder Distributor Builder 1.) Which structural panels do you currently use or consider using? Average Russian Larch 0% 0% 0% 0% 6% Domestic Lauan 0% 0% 0% 0% 8% Imported Lauan 0% 0% 0% 0% 14% CSP 0% 0% 10% 0% 23% SYP 0% 0% 0% 0% 0% OSB 85% 100% 40% 100% 36% DFP 0% 0% 50% 0% 5% MDF/PB 15% 0% 0% 0% 7% 100% 100% 100% 100% 100% 2.) Annual Volume Purchased (MJ)? Totals 1996 - N/A - 2,400 855,485 1997 19,000 - - 2,400 1,890,552 1998 17,400 - 1,500 1,800 1,786,346 3.) Percent Imported? Average Imported 100% 100% 95% 100% 77% Domestic 0% 0% 5% 0% 23% 100% 100% 100% 100% 100% 4.) Origin of Volume Imported? Average Canada 100% 100% 50% 100% 72% United States 0% 0% 50% 0% 3% Malaysia 0% 0% 0% 0% 12% Indonesia 0% 0% 0% 0% 9% Other 0% 0% 0% 0% 4% 100% 100% 100% 100% 100% 5.) Primary Thicknesses (mm)? Average 6.4 0% 0% 0% 0% 88% 8.5 0% 0% 0% 0% 116% 9.0 0% 0% 0% 0% 124% 9.5 30% 60% 40% 50% 164% 11.0 0% 0% 0% 0% 155% 11.1 0% 0% 0% 0% 154% 11.5 0% 0% 10% 0% 157% 12.0 70% 0% 0% 50% 179% 12.5 0% 0% 0% 0% 178% 15.0 0% 0% 0% 0% 206% 15.5 0% 0% 0% 0% 213% 18.0 0% 40% 0% 0% 248% 18.5 0% 0% 0% 0% 254% 21.0 . 0% 0% 0% 0% 291% 24.0 0% 0% • 0% 0% 327% 25.0 0% 0% 0% 0% 344% 27.0 0% 0% 0% 0% 368% 28.0 0% 0% 50% 0% 385% Other 0% 0% 0% 0% 0% 100% 100% 100% 100% 3952% 6.) Primary Dimensions (feet) ? Average 3 x 6 0% 0% 5% 20% 26% 3 x 8 68% 0% 15% 50% 23% 3 x 9 0% 0% 5% 10% 10% 3x10 0% 0% 25% 20% 5% 4 x 8 0% 0% 0% 0% 12% Other 32% 100% 50% 0% 23% 100% 100% 100% 100% 100% 7.) Panel Usage? Average Packaging & Crating 0% 0% 0% 0% 2% Traditional 0% 0% 90% 100% 36% Roof 0% 0% 20% 0% 14% Wall 0% 0% 30% 0% 11% Floor 0% 0% 50% 0% 15% Platform (2x4) 100% 0% 0% 0% 36% Roof 100% 0% 0% 0% 19% Wall 0% 0% 0% 0% 13% Floor 0% 0% 0% 0% 16% Prefabrication 0% 100% 10% 0% 16% Pourform 0% 0% 0% 0% 10% 99 APPENDIX V: ANOVA - Technical Attributes T e c h n i c a l A t t r ibutes Anova: Single Factor SUMMARY Groups Count Sum Average Variance Linear Expansion 11 55.83 5.08 21.93 Panel Strength 10 43.58 4.36 14.56 Uniform Thickness 9 48.75 5.42 19.24 Thickness Swell 18 201.67 11.20 8.82 Formaldehyde Emissions 16 98.17 6.14 15.18 ANOVA | Source of Variation S S df_ MS F P-value F crit | Between Groups 463.12 4.00 115.78 7.75 4.3641E-05 2.53 Within Groups 881.88 59.00 14.95 Total 1,345.00 63.00 100 APPENDIX VI: ANOVA - Marketing Attributes Marketing Attributes Anova: Single Factor SUMMARY Groups Count Sum Average Variance \ Competitive Price 19 195.90 10.31 8.16 Long-term Supply 10 63.55 6.36 11.41 Shipment Arrives In Good Condition 11 82.75 7.52 17.42 Environmental Forestry Practices 12 62.00 5.17 4.82 Overall Quality 15 55.90 3.73 8.57 Panel Size 10 78.90 7.89 6.83 ANOVA -" Source of Variation SS df MS F P-value F crit Between Groups 424.37 5 84.87 9.16 9.33748E-07 2.34 Within Groups 658.23 71 9.27 Total 1,082.60 76 101 A P P E N D I X VII: A N O V A - S e r v i c e A t t r i b u t e s Service Attributes Anova: Single Factor SUMMARY Groups Count Sum Average Variance | On-time Delivery 9 78.50 8.72 3.83 Manufacturer's Awareness of C. Needs 9 30.50 3.39 6.21 Ability To Fill Rush Orders 9 80.00 8.89 4.15 ANOVA Source of Variation SS df MS F P-value F crit Between Groups 176.17 2.00 88.08 18.63 1.31E-05 3.40 Within Groups 113.50 24.00 4.73 Total 289.67 26.00 102 APPENDIX VIII: ANOVA - Distributor Attributes Distr ibutor Attr ibutes Anova: Single Factor SUMMARY Groups Count Sum Average Variance | Linear Expansion 3 24.05 8.02 28.35 Nail Withdrawal 4 30.42 7.61 1.54 Uniform Panel Thickness 4 21.60 5.40 25.52 Thickness Swell 5 63.72 12.74 1.13 Formaldehyde Emissions 4 26.80 6.70 3.79 Competitive Price 5 66.72 13.34 4.00 Long-term Supply 3 19.26 6.42 3.07 Shipment Arrives In Good Condition 3 29.30 9.77 7.53 Overall Quality 5 33.72 6.74 15.87 Panel Size 3 35.30 11.77 8.17 On-time Delivery 5 32.72 6.54 11.36 Manufacturer's Knowledge of Products 3 3.30 1.10 1.86 Manufacturer's Market Knowledge 3 14.34 4.78 3.77 After Sales Service 3 22.76 7.59 22.34 ANOVA Source of Variation SS df MS F P-value F crit Between'Groups 542.12 13.00 41.70 4.37 0.000167 1.98 Within Groups 372.16 39.00 9.54 Total 914.29 52.00 103 A P P E N D I X IX: A N O V A - B u i l d e r A t t r i b u t e s Bui lder Attributes Anova: Single Factor S U M M A R Y Groups Count Sum Average Variance Linear Expansion 8 50.90 6.36 30.23 Panel Strength 5 21.62 4.32 8.94 Thickness Swell 10 136.23 13.62 17.39 Formaldehyde Emissions 9 66.63 7.40 9.02 Competitive Price 10 123.23 12.32 10.82 Long-term Supply 6 66.40 11.07 11.19 Shipment Arrives In Good Condition 5 59.10 11.82 25.97 Product Availability 6 35.51 5.92 29.99 Environmental Forestry Practices 7 56.13 8.02 10.75 Overall Quality 8 54.52 6.82 11.49 On-time Delivery 8 51.15 6.39 6.26 Manufacturer's Awareness of C. Needs 6 9.80 1.63 8.37 Ability To Fill Rush Orders 7 42.11 6.02 10.61 After Sales Service 6 34.70 5.78 17.47 A N O V A Source of Variation S S df MS F P-value F crit Between Groups 1,082.78 13.00 83.29 5.73 1.8E-07 1.83 Within Groups 1,264.79 87.00 14.54 Total 2,347.57 100.00 104 APPENDIX X: ANOVA - Trading House Attributes T r a d i n g H o u s e At t r ibutes Anova: Single Factor SUMMARY Groups Count Sum Average Variance Panel Strength 3 21.15 7.05 42.10 Internal Bond 2 2.53 1.27 4.11 Nail Withdrawal 2 15.29 7.65 6.58 Uniform Panel Thickness 3 19.39 6.46 22.96 Thickness Swell 3 35.15 11.72 19.91 Formaldehyde Emissions 3 37.02 12.34 9.48 Competitive Price 4 59.85 14.96 7.07 Long-term Supply 2 15.56 7.78 17.07 Shipment Arrives In Good Condition 3 25.15 8.38 14.94 Environmental Forestry Practices 4 33.85 8.46 6.38 Overall Quality 2 11.29 5.65 1.33 Panel Size 3 28.02 9.34 6.66 On-time Delivery 4 31.85 7.96 9.77 Manufacturer's Knowledge of Products 2 0.32 0.16 3.39 Manufacturer's Awareness of C. Needs 3 5.02 1.67 0.88 Manufacturer's Market Knowledge 2 15.29 7.65 35.04 Ability To Fill Rush Orders 2 17.56 8.78 33.27 Ability To Fill Small Orders 2 17.69 8.85 0.00 ANOVA | Source of Variation S S <# MS F P-value F crit | Between Groups 656.39 17.00 38.61 2.96 0.004238 1.96 Within Groups 404.32 31.00 13.04 Total 1,060.71 48.00 105 APPENDIX XI: ANOVA - Distributors and Trading House Attributes Trading House and Distributor Attributes Anova: Single Factor SUMMARY Groups Count Sum Average Variance] Panel Strength 5 34.59 6.92 20.03 Nail Withdrawal 6 45.13 7.52 2.54 Uniform Panel Thickness 7 37.98 5.43 20.54 Thickness Swell 8 97.49 12.19 7.56 Formaldehyde Emissions 7 60.56 8.65 11.44 Competitive Price 9 123.20 13.69 4.92 Long-term Supply 5 32.64 6.53 2.71 Shipment Arrives In Good Condition 6 53.56 8.93 10.47 Environmental Forestry Practices 5 36.76 7.35 3.21 Overall Quality 7 43.76 6.25 14.43 Panel Size 6 61.18 10.20 10.97 On-time Delivery 9 61.20 6.80 8.00 Manufacturer's Knowledge of Product 5 3.46 0.69 1.63 Manufacturer's Market Knowledge 5 28.49 5.70 12.59 ANOVA Source of Variation SS df MS F P-value F crit Between Groups 862.92 13.00 66.38 7.05 9E-09 1.85 Within Groups 715.40 76.00 9.41 Total 1,578.31 89.00 106 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0075351/manifest

Comment

Related Items