The Importance of the Brand
A study into the importance of the Brand under different levels
involvement
The Importance of the Brand
A study into the importance of the Brand under different levels
involvement
Peter van Es
October 2008
University of Groningen
Faculty of economics and business
Master Thesis Business Administration
Marketing Research and Marketing Management
Author:
Supervisors:
Peter van Es
dr. J.E. Wieringa
Saffierstraat 148
dr. J.E.M. van Nierop
9743 LM Groningen
N: s1660764
T: 06 55 89 6516
Management summary
There are different theories which discuss the level of involvement of a consumer. Within these theories there is some vagueness under what conditions the brand is a decisive attribute for the purchase decision. The problem statement for this study is: Is there a difference in the importance of brand equity, at the point of purchase, between high and low involvement products?
Based on the literature overview it is clear that both price and brand name are important determinants of product quality perceptions. The product quality is assumed to be the most important attribute which consumers base their decision on. In the literature there was no research on involvement at the point of purchase found. The level of involvement at the point of purchase (situational involvement) will play a role in how consumers will judge product characteristics and information about the product. Within the literature, reflecting theory on involvement to the point of purchase, it is however not apparent when and if involvement plays a decisive role on the purchase decision.
A conceptual model is developed based on the findings in the literature. This model shows the available information for making a purchase decision. The model illustrates that the product information (i.e. price, proportion), brand equity (i.e. brand knowledge, brand loyalty) and the situational involvement (high, low) have an influence on the purchase decision. This study will focus on the relationship between the level of involvement and brand equity.
The conjoint methodology in the present study will be Choice-Based Conjoint Analysis. The main reason to use this method of analysis is the possibility to create real life sets of stimuli. The goal of this study evolves around choice making at the point of purchase. For that reason it is realistic to give respondents a full range of products with price and other attributes as given. The products used in this study are macaroni (low involvement situation) and a notebook (high involvement).
attribute on which the consumers make their purchase. The findings indicate that the brand name (and thus the brand equity) is less important under conditions of high involvement. Under high involvement all attributes – that are relevant to the consumer – are considered and the brand name plays just a minor role in that. This does not exclude the possibility that the brand name of a product under high involvement is not decisive. As soon as all attributes are more or less the same brand equity will play a decisive role.
Preface
Since the start of the idea for a thesis, in the beginning of March of this year, a lot has happened that influenced the process and the results. During this final phase of your study career you learn that doing research, without preset boundaries, contains a lot of surprising elements and overlooked opportunities and threats. It is an experience similar to be dropped somewhere in Europe without a map. You know the basic elements and places of the continent, but have to figure out on your own how to get back home.
During this struggle of finding my way back home, the whole process of writing this thesis, a lot of people were willing to help me find my way. First of all, my thanks go out to all those respondents who filled out the questionnaire, without their help the research could not be done. Second of all I like to thank my supervisor dr. J.E. Wieringa. Although we did not meet often you always gave interesting insights en helpful comments to improve or adjust the research. Especially in the final part of the thesis you were very supportive. Third of all I like to thank my fellow Marketing students who gave helpful comments at the start, during and at the end of this process. Special thanks case goes to Riko Ooijevaar who helped during the whole process with supportive comments and lively discussions and to Linda Postma who kindly corrected my first draft on basic grammar. Last but not least I like to thank my family who always were very interested in my research and the progress I made. Although the people above all helped me writing this thesis, I am still responsible for all the mistakes and errors.
My final thanks are for my fiancé, Marijke. During my journey the last seven months I never walked alone.
Peter van Es
Content
Management summary ...3
Preface ...5
Chapter 1 Introduction...7
Chapter 2 Theoretical Framework ...10
2.1 Introduction ...10
2.2 Brand Equity...10
2.2.1 Before Brand Equity...10
2.2.2 The emergence of Brand Equity...13
2.2.3 Brand Equity in the present study...16
2.2.4 The link between involvement and brand equity...16
2.3 Involvement ...17
2.3.1 Previous findings on Involvement...17
2.3.2 Involvement in the recent literature...20
2.3.3 Involvement at the point of purchase...21
2.3.4 Involvement in the present study...22
2.4 Conceptual Model...23
2.5 Hypotheses ...25
Chapter 3 Research Design...27
3.1 Technique ...27
3.2 Conjoint Analysis Design...27
3.3 Data Collection Method ...30
3.3.1 Evaluation Method...32
3.4 Data Collection Sample...32
3.5 Analysis: Importance ...33
Chapter 4 Results ...35
4.1 General statistics ...35
4.2 Results ...36
4.2.1 Low involvement without brand...36
4.2.2 Low involvement with brand...37
4.2.3 High involvement without brand...38
4.2.4 High involvement with brand...39
Chapter 5 Conclusions and Recommendations ...42
5.1 Discussion...42
5.2 Limitations...44
5.3 Further research on this topic ...45
Reference List ...47
Chapter 1 Introduction
The importance of the brand is indisputable. Building awareness for a brand, by means of advertising, builds brand equity (Aaker 1991, 1996, Keller 1993, 2007). When two cars are equal on all attributes, but with two different brand names, the brand equity of the brands will be decisive. The knowledge of consumers of a product category, a brand and a store will always be part of the Consumer Decision Process (CDP, Blackwell et al., 2006). However, for different purchases and different consumers the level of involvement will not always be the same (Percy and Elliot, 2005). In marketing books, when discussing the role of the brand, advertising or the purchase decision, there is often a distinction in high and low involvement. Percy and Elliot (2005) underline the importance of knowing if the target audience is high or low involved with the product.
Petty and Cacioppo (1983) developed a model for the likelihood of elaboration of an advertisement, the elaboration likelihood model (ELM). One of the main findings in that article is that under conditions of high involvement the consumer is persuaded by messages content and argument quality, whereas under conditions of low involvement the consumer is persuaded by certain cues and surface characteristics (i.e. emotions, colours). Based on this it is already questionable if the brand of a product has the same importance under different conditions of involvement.
Blackwell et al. (2006) discuss in their book Consumer Behaviour the consumer decision process by means of the CDP-model. Before the pre-purchase phase the consumer searches his or her memory internally for information about the products and brands he or she knows (the search phase). For more complicated purchases with higher financial and perceived risk (high involvement decision according to Percy and Elliot (2005)) the larger the possibility the consumer will also search externally for information. After the search phase, the pre-purchase phase starts. In that phase the consumer will make a decision between different alternatives based on (for that particular consumer) important attributes of the product.
of low involvement the consumer is more likely to trust on their own knowledge (internal search) and choice a brand and product they know. Although it is a given fact that investing in the brand awareness pays of, is it under certain conditions of high involvement not better to invest more in product quality and development? Based on the information above producers might agree on that statement. Under conditions of high involvement the consumer should be persuaded by quality arguments and at the point of use the consumer should have a satisfying experience. The focus for high involvement products should thus be more on development than on brand building.
In the literature there are some indications that under low involvement the focus is on the experience. Ehrenberg (1974) distinguishes between high and low involvement by means of attitude change. In that article Ehrenberg discussed the path to forming an attitude. Under conditions of high involvement the consumer first has certain believes about a brand, then forms an attitude and then tries the brand. Under conditions of low involvement the consumer starts with certain believes, then tries a brand and after trial the consumer forms an attitude towards the brand. Based on this article the brand is not so decisive at the point of purchase under conditions of low involvement, the products are homogeneous to a high extent. During the purchase phase the consumer will get a brand and after trial its attitude is formed. This means that managers of low involvement products should invest more in the product and the experience of the usages of the product. Whereas the managers of high involvement products should focus more on creating believes and attitudes for there brand.
In summary, there are different theories which discuss the level of involvement of a consumer. Within these theories there is some vagueness under what conditions the brand is a decisive attribute for the purchase decision. This is the starting point of this thesis. The problem statement for this study will be: Is there a difference in the importance of brand equity, at the point of purchase, between high and low involvement products? The questions that this thesis tries to answer are:
• Does involvement and brand equity separately influence the purchase decision? • What is the decisive attribute (within the product information available) for a
purchase decision under high or low involvement conditions?
• What is the impact of the brand name (and thus brand equity) on the other attributes?
Chapter 2 Theoretical Framework
2.1 IntroductionThis chapter is a literature overview containing the important topics of the problem statement. In addition, a conceptual design for the research is developed and the hypotheses for the research are defined. The chapter starts with the introduction and discussion of Brand Equity and Involvement. In these parts, essential elements of the conceptual model will be discussed.
2.2 Brand Equity
One of the most popular and potentially important marketing concepts to arise in the 1980s is brand equity. Its emergence, however, has meant both good news and bad news to marketers. The good news is that brand equity has elevated the importance of the brand in marketing strategy and provided focus for managerial interest and research activity. The bad news is that confusingly, the concept has been defined a number of different ways for a number of different purposes. No common viewpoint has emerged about how to conceptualize and measure brand equity (Keller, 2007).
2.2.1 Before Brand Equity
Before the 1980s authors believed that several cues play a role in choice of products. The cue that got the most attention was perceived quality. Scitovszky (1945) observes that the use of price as an indicator of product quality is not unreasonable. Scitovszky explains that the occurrence of using price as an indicator, represents a belief that price in the marketplace is determined by the interplay of the power of competitive supply and demand. Such power would lead to a "natural" ordering of competing products on a price scale for the consumer, which results in a strong positive relationship between price and product quality. Consequently, given the belief that price and quality are positively related, it is a natural internal process that consumers would use price as an indicator of quality. Subsequently, other economic and marketing theorists expanded the argument to include other signals of product quality such as brand and store names and advertising expenditures (Dodds, Monroe and Grewal, 1991).
satisfied with their purchase decision and tended to select fewer information dimensions. Jacoby and Olsen conclude by stating: These data appear to suggest that brand name does indeed serve at least some information chunking function (perhaps in terms of facilitating information retrieval from long term memory) in consumer decision-making.
So before the term brand equity was introduced multiple authors discovered that decisions were not solely based on price alone. Clearly, perceptions of the same price stimulus may vary across consumers and, for one consumer, across products, purchase situations, and time (Cooper 1969b). Price can be both an indicator of the amount of sacrifice needed to purchase a product and an indicator of the level of quality. Higher prices lead to higher perceived quality and consequently to a greater willingness to buy. At the same time, the higher price represents a monetary measure of what must be sacrificed to purchase the good, leading to a reduced willingness to buy. The cognitive trade-off between perceptions of quality and sacrifice results in perceptions of value (Dodds, Monroe and Grewal, 1991).
That trade-off was already observed by Scitovszky (1945) as a paradoxical situation in which a product, service or good offered at a lower price than competing commodities would be both more attractive to the consumer because it is cheaper and less attractive because of its suspected inferior quality. Monroe and Krishnan (1985), using Monroe's (1979) conceptualization of perceived value, provided a model relating price, perceived quality, perceived sacrifice, perceived value, and willingness to buy (Figure l.1). In that model, actual price is an objective external attribute of a commodity that consumers perceive as a stimulus. That part of the model is based on earlier research by Jacoby and Olsen (1977) who concluded that price has both objective external properties and subjective internal representations that are derived from the perceptions of price, thus resulting in some meaning to consumers. The perception of value in turn directly influences willingness to buy. Perceptions of value will increase as price increases from below the buyers’ lower acceptable price limit to some acceptable price within their acceptable price range. However, as price increases beyond the acceptable range, perceptions of value would decline. This can vary across consumers, due to use of own “objective” characteristics.
quality and value, and hence willingness to buy (Figure 1.2). Based on previous research the authors expect that the primary effect of the additional cues of brand and store name is to enhance the effect of price on buyers’ quality perception.
Objective Price Perception of Price Perceived Quality Perceived Sacrifice Perceived Value Willingness to buy - + + + + Figure 1.1
Conceptual Relationship of Price Effect (Monroe and Krinsha 1985)
Objective Price Perception of Price Perceived Quality Perceived Sacrifice Perceived Value Willingness to buy - + + + + Figure 1.2
Extended Conceptualization to include Brand and Store Name (Dodds, Monroe and Grewal 1991)
However, the study shows that the price – perceived quality relationship was not enhanced, but rather diminished by the cues of brand and store name. This suggests that when the brand name is known, the importance of price in a consumer decision will decline.
Dodds, Monroe and Grewal also find that when price is the only extrinsic cue available, the subjects clearly perceived quality to be positively related to price. In their study they used two different products (stereo and calculator) and per product two different brand names. In their results there were some differences in the findings across the two products. According to the authors it is most likely that the difference in brand favourability accounts for these dissimilarities. Brand and store name had a positive effect on perceptions of quality, value, and willingness to buy. The results also show a large price-perceived quality effect. The effect of brand name information on quality perceptions is found to be large in all conditions. The brand name effect increases in the presence of price and store information. However, store name has a moderate effect on quality perceptions when presented alone, and in effect diminishes as price and brand information are provided. In conclusion, it is clear that both price and brand name are important determinants of product quality perceptions, and that the effect of store name, though positive, is small. Dodds, Monroe and Grewal conclude with the statement that the relative effect of these three variables on perceptions of value and purchase intentions clearly warrants additional research, as does their effect on actual choice.
2.2.2 The emergence of Brand Equity
In 1991 Aaker publishes Managing Brand Equity and becomes a guru in the field of this concept. Moreover, his work is becoming the starting point and the foundation of most of the academic research into Brand Equity after 1991. As a concept, Brand Equity is the incremental utility or value added to a product by its brand name, such as Coca-cola, Levi’s and Nike. As a definition, brand equity is the difference in consumer choice between the focal branded product and an unbranded product given the same level of product features (Yoo, Donthu and Lee 2000).
and reducing reliance on promotions. For the consumer brand equity creates value differently. First, brand equity makes products recognizable and saves the consumer time searching. Second, it creates a level of reliability to the brand. The brand becomes an indicator for a certain level of quality. Third and last, it gives the consumer a possibility to differentiate them as a person. Brand equity creates a certain identity for a brand and associations for the consumer. This proposition of Aaker has been well supported by additional research by many authors. In the article of Yoo, Donthu and Lee (2000) the findings of previous researchers are listed. Brand equity
• affects merger and acquisition decision making (Mahajan, Rao, and Srivastava 1994) • affects stock market responses (Lane and Jacobson 1995; Simon and Sullivan 1993) • determines the extendibility of a brand name (Rangaswamy et al. 1993).
• increases the probability of brand choice, willingness to pay premium prices, marketing communication effectiveness, brand licensing opportunities, and decreases vulnerability to competitive marketing actions and elastic responses to price increases (Barwise 1993; Farquhar et al. 1991; Keller 1993; Simon and Sullivan 1993; Smith and Park 1992).
• provides sustainable competitive advantages to the firm from a managerial perspective (Bharadwaj, Varadarajan, and Fahy 1993).
brand. The image comprises all thoughts and opinions of the consumers about a certain brand. As soon as these thoughts and opinions are shared by the target audience it becomes difficult to change. For that reason it is an important part of the brand equity that is difficult to control and influence by the firm. Brand associations include all associations that consumers think of when they see or hear about the brand. These associations can be created by the firm by (i.e.) using famous endorsers with shared values or goals (being the best). These associations create value for the consumers because it represents the identity, values and believes of a brand. A consumer can identify oneself with these associations.
Yoo, Donthu and Lee (2000) consider the suggestions of Keller as common dimensions of brand equity. In summary, high brand equity implies that customers have a lot of positive and strong associations related to the brand, perceive the brands of high quality, and are loyal to the brand. The authors suggest that brand equity can be created, maintained, and expanded by strengthening the dimensions of brand equity. Any marketing action has the potential to affect brand equity because it represents the effect of accumulated marketing investments into the brand. Brand name recognition with strong associations, perceived quality of product, and brand loyalty can be developed through careful long-term investment (Yoo, Donthu, Lee, 2000). Thus, brand equity should be managed over time by maintaining the brand consistency, protecting the sources of brand equity, making appropriate decisions between fortifying and leveraging the brand, and fine-tuning the supporting marketing program (Keller 1998). Researchers also suggest that marketing decisions and market conditions affect brand equity. For example, Simon and Sullivan (1993) list advertising expenditures, sales force and marketing research expenditures, age of the brand, advertising share, order of entry, and product portfolio as sources of brand equity. Other marketing activities such as the use of public relations (Aaker 1991); warranties (Boulding and Kirmani 1993); slogans or jingles, symbols, and packages (Aaker 1991); company image, country of origin, and promotional events (Keller 1993); and brand-naming strategy (Keller, Heckler, and Houston 1998) have also been proposed.
brand as a results of their experiences over time. Thus, although strongly influenced by the marketing activity of the firm, brand equity ultimately depends on what resides in the minds of the consumers. Consumers’ differential responses, which make up brand equity, are reflected in perceptions, preferences, and behaviour related to all aspects of brand marketing, including their choice of a brand, recall of copy points from an ad, response to a sales promotion, and evaluation of a proposed brand extension.
2.2.3 Brand Equity in the present study
Multiple authors have defined brand equity in slightly different ways. This study is following the definition of brand equity is used as proposed by Yoo, Donthu and Lee (2000). They define brand equity as the difference in consumer choice between the focal branded product and an unbranded product given the same level of product features. This definition deals with the comparison of two products that are identical in all aspects except brand name (e.g., Samsung product versus no-name product). All consumers have an impression of what Samsung conveys about a product, but they do not have a similar impression about what no-name conveys. Samsung’s brand equity is the extra value embedded in its no-name, as perceived by the consumer, compared with an otherwise equal product without the name. The difference in consumer choice between these two products can be assessed by measuring the intention to buy or a preference for the focal brand in comparison with the no-name counterpart. This definition is commonly used in research as it gives the researcher the opportunity to test the effect or the size of brand equity by just using only the brand name in his research.
2.2.4 The link between involvement and brand equity
2.3 Involvement
In a recent article of Park et al (2007) involvement is defined as the perceived personal relevance of a product based on the individual consumer’s needs, interest and values (Krugman, 1966, Zaichkowsky, 1985). Involvement is important for multiple disciplines of academic research, such as psychology (e.g. social, consumer) and economics. A meta-analysis on involvement across all disciplines would find many definitions, interpretations and models to implement. This article will not be such an analysis. However, it is still relevant for the objective of this study to give a brief overview of different definitions and an explanation of why this article will view involvement as it does. The distinction of involvement that can be made is between situational and enduring involvement, within those definition a person can be high or low involved.
2.3.1 Previous findings on Involvement
In one of the most widely accepted frameworks for consumer involvement according to different articles (Richinis et al. 1992 a.o.), Houston and Rothschild (1978) distinguish between situational involvement and enduring involvement. Situational involvement is a temporary elevation of interest that fluctuates, usually within the time frame of a purchase decision, while enduring product involvement is a stable phenomenon that represents the consumer’s personal interest in the product over a long period. A consumer that is into automobiles will be high involved in this product. This is a form of high enduring product involvement. A person that is not into the world of cars will have a low enduring product involvement. Nonetheless, that latter person can be high involved whilst purchasing a car for his own usage. This is a form of high situational involvement. After the purchase the situational involvement is gone and what stays is the low enduring product involvement.
probability of being rejected because they are more likely to fall within the unacceptable range of a person’s implicit attitude continuum. The sizes of the different parts of latitude can be narrow or wide, a summary of this continuum in table 2.1.
LOA LOR LON
High Involved Narrow Wide Narrow
Low Involved Wide Narrow Wide
Kruggman (1965) has proposed an alternative view that has achieved considerable recognition among consumer researchers. According to this view, increasing involvement does not increase resistance to persuasion, but instead shifts the sequence of communication impact. Following to that research Ehrenberg (1974) distinguishes between high and low involvement by means of attitude change. In his article Ehrenberg discusses the path to forming an attitude. Under conditions of high involvement the consumer first has certain believes about a brand, then forms an attitude and then tries the brand. Under conditions of low involvement the consumer starts with certain believes, then tries a brand and after trial the consumer forms an attitude towards the brand.
Consumer researchers developed their notion on involvement by building on Krugman’s and Eherenberg’s view. Petty and Cacioppo (1981) developed a new view on the effects of
Table 2.1
Sherifs’ Theorem: Continuum on involvement (Sherif et al. 1965)
Low Involvement High Involvement
Believe Believe
Behavior Attitude
Attitude Behavior
Figure 2.3
involvement on consumer response to advertisements. The authors created the Elaboration Likelihood Model (ELM) of attitude change.
The basic idea of the ELM is that different methods of inducing persuasion may work best depending on whether the elaboration likelihood of the communication situation (i.e. the probability of message- or issue-relevant thought occurring) is high or low. Under conditions of high involvement and where elaboration is likely, attitude change follows the ‘central route’. This means that a person exercises conscientious and thorough consideration of the information that he or she feels is central to the true values and advantages of a particular position (Petty et al.1983). In other words, due to high involvement there exists a high need for information. This in turn means that consumers are persuaded by message content and argument quality. The result of this route is enduring attitudes, attitudes that predict behaviour and shape attitudes that are resistant to change.
Contrary to the central route, the peripheral route is followed under conditions of low involvement, and therefore low elaboration likelihood. In this route, the need for cognition is low, which means that influence is exerted by various cues associated with the issue or by surface characteristics (e.g. source, colours and emotions). The result of this route is a more temporarily attitude, which is less resistant to change. In table 2.2 an overview of the discussed ELM. In figure 2.4 the elaboration continuum is shown. A consumer can be placed anywhere on this continuum based on the results on the different concepts in table 2.2
Central Route Peripheral Route
Motivation High Involvement
High need for Recognition
Low Involvement Low need for recognition
Ability No distraction
No time pressure
Distraction Time pressure
Persuasion by Message content
Argument Quality
Cues
Surface characteristics (source, colours, emotions)
Result Enduring attitude
Attitudes predict behaviour Attitudes resistant to change
Less enduring attitude To al lesser extent Less resistant to change Table 2.2
Research in social psychology has supported the view that different variables affect persuasion under high and low involvement conditions. For example, the quality of the arguments contained in a message had a greater impact on persuasion under conditions of high rather than low involvement (Petty and Cacioppo 1979; Petty, Cacioppo and Heesacker 1981). On the other hand, peripheral cues such as the expertise or attractiveness of a message source (Chaiken 1980; Petty, Cacioppo and Goldman 1981; Rhine and Severance 1970) had a greater impact on persuasion under conditions of low rather than high involvement.
2.3.2 Involvement in the recent literature
In today’s literature on involvement, the research of Petty and Cacioppo (1983) is acknowledged. Percy and Elliot (2005) embrace the ELM, but have some additions. The authors make a distinction between high and low involvement and the effect on the buying process and do not consider attitude change. They define involvement in terms of perceived risk. When there is perceived risk the decision is described as high involvement and when there is little perceived risk, the decision is low involvement. Risk is defined by the authors in terms of fiscal risk (e.g. a lot of money is involved) or psychological risk because of personal of social involvement associated with the product. Percy and Elliot also state that perceived risk is dependent upon the target audience. In the end, perceived risk is different for every consumer due to i.e. the different social groups, stages in life or financial health.
Blackwell et al. (2006) argue that the degree of personal involvement is a key factor in determining the type of decision process that a consumer will follow. The degree of involvement is determined by how important consumers perceive the product or service to be.
not much thought many thoughts
Peripheral route
peripheral cues (rules of thumb)Central route
argument qualityElaboration Continuum
Figure 2.4Simply stated, the more important the product or the service to a consumer, the more motivated he or she is to search and be involved in the decision. Involvement becomes activated and felt when intrinsic personal characteristics (needs, values, self-concept) are confronted with appropriate marketing stimuli within a given situation.
In a recent study by Park, Han and Lee (2007) the authors chose for shampoo as a low involved product and a notebook as a high involved product. The respondents of this research had to buy the product. Thus in this research the authors made use of situational involvement. According to the authors there were three reasons for using situational involvement instead of enduring involvement. First, individuals’ involvement for the same product can be different depending on their personal characteristics. Enduring involvement is difficult to control in research because the researchers have to find respondents with high and low enduring involvement. Situational involvement is easier to manipulate by making extreme difference in a homogeneous and heterogeneous product. Second, it is difficult to control the compounding effects resulting from the difference between two products used in a laboratory experiment (e.g., shampoo vs. notebook). The method of giving subjects different goals for a single product can increase the internal validity of this study. Third, as Mittal (1995) suggests, the situational importance of a purchasing decision is likely to be most representative of the variance in the consumer’s involvement, even more than product-class involvement. And thus, the benefits for using situational involvement instead of enduring involvement are high. Situational will be the most representative at the point of purchase and is easier to manipulate while enduring is harder to control.
Park, Han and Lee (2007) do distinguish between consumers that for the first time independently buy a product and consumers that buy a product regularly or not for the first time. A first time buyer has always a higher level of involvement. Consequently, while using situational involvement in a research to manipulate consumers into conditions of low involvement, the research had to conduct a product that consumers buy on a regular basis.
2.3.3 Involvement at the point of purchase
Petty et al. (1983) under conditions of high involvement, consumers will be persuaded by messages content and argument quality. Under conditions of low involvement consumers are more sensitive for cues and surface characteristics. When the theory of Petty et al. (1983) is reflected on how people act at the point of purchase the consumer will be sensitive for argument quality of a product under conditions of high involvement and sensitive for product cues like brand name and associations under conditions of low involvement. Following this view, the brand equity seems to play a more decisive role under low involvement because a low involved consumer will judge a product on (i.e.) the colour, the size and/or the associations with the brand. Under conditions of high involvement the consumer will try to find reasonable arguments and brand equity is probably not decisive. The judgement will be made on other product characteristics and will leave brand name with relative less importance.
According to Ehrenberg (1974), high involved consumers form already an attitude towards a brand or product before they behave in a certain way. Low involved consumers form their attitude based on behaviour. So reflecting this theory to the point of purchase it is reasonable to assume that brand equity plays a more important role under conditions of high involvement because attitudes have to be formed and after that the first trial is made, under low involvement people have some beliefs but first try the product before the attitudes are formed. Consequently, under low involvement investments should be made in the product quality, performance and the product experience, under conditions of high involvement investments should be done in building the brand and shape consumers attitudes.
In summary, the level of involvement at the point of purchase (situational involvement) will play a role in how consumers will judge product characteristics and information about the product. Within the literature, reflecting theory on involvement to the point of purchase, it is however not apparent when and if involvement plays a decisive role on the purchase decision.
2.3.4 Involvement in the present study
consumer goods (FMCG) (i.e. shampoo) and consumers that buy durable goods (i.e. a notebook). The goal of this study is to measure the impact of brand equity on choice and tries to argue if there is a different impact on choice because of involvement. To do so involvement is defined in this article as the perceived risk of a consumer when purchasing the product (situational involvement). This definition is based on Percy and Elliot (2007) and the ELM. To also bear in mind the defining concepts of that model, the respondents will be asked to imagine that for the high involvement product they really need a new version of that product to stimulate them to be high involved.
2.4 Conceptual Model
Situational
Involvement
Product
Information
Brand
Equity
Point of
Purchase
Purchase
Decision
• Brand Loyalty • Brand Associations• Brand Knowledge (awerness and image)
2.5 Hypotheses
Based on the literature overview and the conceptual model the hypotheses for this research are developed. Noteworthy is that in the literature it was not noticeably conceptualized if there is an interaction effect between involvement and brand equity on the purchase decision. Therefore, the hypotheses are based on the knowledge described above and on the reflected theories of involvement. It is argued above that price is an excellent indicator for quality. It is assumed that consumers always try to buy the product with the highest perceived quality within their range of monetary power. Price also plays a role as indicator for how much money has to be sacrificed. Thus quality is the most important attribute in a purchase decision. However, quality of a product is often revealed after the purchase, hence the consumers have to use other attributes to decide which product has the highest perceived quality within the range of money they can sacrifice.
Using the research of Dodds, Monroe and Grewal (1991), it is expected that when the brand name is not mentioned, price is the most important attribute for the consumers’ purchase decision. It is also expected that under high involvement other attributes will be important indicators as well, due to the “argument quality” of the other indicators. In the ELM the high involved consumers take the central route in which they be persuaded by argument quality. Within the peripheral route low involved consumers are persuaded by cues and surface characteristics like source, colours and emotions. It is assumed that for the low involvement consumers price is the cue on which most decisions are based, whilst high involvement consumers also base their decision on the price but to a larger extent also on the argument quality of the other indicators. Following:
H1: Under conditions of high and low involvement, without knowledge of the brand name,
price will be the most important attribute for the purchase decision.
H2: Without knowledge of the brand name, price will be relatively more important under low
involvement conditions rather than under high involvement.
brand will play a relative important role in the purchase decision. Founded in the literature on involvement, it is expected that under low involvement brand will have a relatively higher importance than under high involvement, due to the fact that when consumers are low involved they base their decisions on cues rather than on argument quality and other information. It is also expected that brand will not have the highest relative importance of all attributes under both conditions of involvement. Following the literature discussed above, brand is an important attribute and will play a decisive role, however price is still an important factor on which consumers’ base their decisions. Price will play a double role, as sacrifice indicator and as quality indicator, for that reason it is expected that price will be the most important attribute. Following:
H3: Brand equity has a negative impact on the relative importance of all other attributes under
both conditions of involvement.
H4: Under low involvement conditions, brand equity will have a higher impact than under
high involvement and thus is more important under those conditions.
H5: Brand equity will be, under conditions of high and low involvement, not the decisive
argument for consumers’ purchase decision.
Chapter 3 Research Design
The primary goal of this study is to find if there is a difference of the importance of the brand name (equity) under different conditions of involvement. To come to such a conclusion the research should contain a range of products under low and high involvement. Unfortunately the conditions of this study are such that it is not possible to contain more than one product under both circumstances. The research will focus on two products with extreme levels of involvement. For low involvement this product will be macaroni and for high involvement it will be a notebook, following previous research in which FMCG are seen as products with a low perceived risk and electronics are seen with a high perceived risk and thus high involvement (Park, Han and Lee, 2007)
3.1 Technique
Conjoint Analysis is a multivariate technique developed specifically to understand how respondents develop preferences for any type of object (products, services or ideas). It is based on the simple premise that consumers evaluate the value of an object (real or hypothetical) by combining the separate amounts of value provided by each attribute. Moreover, consumers can best provide their estimates of preferences by judging objects formed by combinations of attributes. Utility, a subjective judgement of preference unique to each individual, is the most fundamental concept in conjoint analysis and the conceptual basis for measuring value (Hair et al 2006). With conjoint analysis, the importance of each attribute can be defined and the utility of specific levels within an attribute can be discovered. With this technique, it is possible to create a product that completely corresponds with the wishes of a consumer. It is achievable to view different utilities, preferences per level per attribute.
3.2 Conjoint Analysis Design
attributes. With aid of shopping assistants and comparison sites on the internet1 this resulted in three levels per attribute for both products. In table 3.1 an overview of the attributes and levels is presented.
Table 3.1 Attributes and levels
Macaroni (LI-product) Notebook (HI-product) Attributes Levels Attributes Levels
Brand Honig Gran'Italia Private Label Brand HP Fujitsu-Siemens Asus Price (per 500 gram) € 0.49 € 0.89 € 1.29 Price € 899 € 1199 € 1499 Speed (processor) 1.7 GHz 2.1 GHz 2.4 GHz Packaging Normal plastic
Box
Quality plastic
(re-closable) Size 100 GB
300 GB 500 GB
The intervals within the levels, that are quantifiable, are always of the same size within an attribute. One reason is that the preliminary research shows that these are the important and frequently used levels. Another reason is that the validity of the results will increase with intervals of the same size, so that differences in choice among consumers cannot be assigned to dissimilarity in interval size. For the packaging of macaroni a quality plastic re-closable bag is presented. This type of bag is not for sale in the Dutch supermarkets. This particular type of packaging was invented to create a third level for this attribute. Although it is not a central part of the study, it will be interesting to see if the Dutch consumer has any interest in this type of packaging.
After defining the attributes and levels the following step will be the choice of the conjoint methodology. This choice evolves around the basic characteristics of the proposed research. Examples of these characteristics are the number of attributes examined the level of analysis, the type of choice task and the data collection format. The number of attributes is specified above. In the case of low involvement their will be three attributes and four with high involvement. The level of the analysis can be individual, segment or aggregate. Because in this particular study the focus is on an effect on the whole market, the level of analysis will be on aggregated level. The choice task that respondents have to complete will be one of a
choice between sets of stimuli. An important advantage of this approach is the fact that the products are presented as real life options, therefore a respondent is forced into a decision as if he or she would make in the store at the point of purchase. The respondent will have to choose between three full sets of attributes and they will have an escape: a “none of these” option. The escape option is particular important as it allows estimation of the minimum requirements for a product to become considered (Vriens and Frazier, 2003). In this study, however, the expectation is that one of the three options will always have at least favourable options to the consumers.
In essence there are three types of conjoint methodologies (Hair at al. 2006):
• Traditional Conjoint Analysis; characterized by a simple additive model, generally containing up to nine factors estimated for each individual. A respondent evaluates stimuli constructed with selected levels from each attribute (known as full profiles). Although this format has been the foundation of conjoint studies for many years, two additional methodologies have been developed in an attempt to deal with certain design issues.
• Adaptive Conjoint Method; developed to accommodate a large number of factors (up to 30) that would not be feasible in traditional conjoint analysis. It employs a computerized process that adapts the stimuli shown to a respondent as the choice task proceeds. Moreover, the stimuli can be composed of subsets of attributes, thus allowing for many more attributes.
• Choice-based approach; employs a unique form of presenting stimuli in sets (choose one from a set of stimuli) rather than one by one. Due to the more complicated task, the number of factors included is more limited, but the approach does allow for inclusion of interactions and can be estimated at the aggregate or individual level.
3.3 Data Collection Method
The primary focus of this study will be on consumers that have to make their decisions in the store. The sample however, due to the circumstances (i.e. budget and time) of the writing process of this thesis, will consist of only students. Nevertheless this will probably not influence the results. For theory-testing research, a student sample has been deemed acceptable and even desirable. A maximally homogeneous sample (e.g., a student sample) has important advantages for theory validation research (Calder, Philips, and Tybout 1981). In addition, students are famous for their cheap diners (macaroni and cheese) and their use (college and private) of notebooks.
These students will be split up into two groups that are more or less similar to each other. The first group will complete ten choice tasks for the low involvement product (macaroni) without the attribute brand and following this group will complete ten choice tasks for the high involvement product (notebooks) with the attribute brand. The second group will complete ten choice tasks for the low involvement product with the attribute brand and following this group will complete ten choice tasks for the high involvement product without the attribute brand. So each respondent will have to evaluate 20 choice tasks. Due to the many combinations of attributes and levels that can be constructed, not every respondent will have to evaluate the same set of stimuli. For this particular part of the design, a software program (Sawtooth Software) is used. This program forms a questionnaire based on the attributes and levels and the design the user specifies.
In the program four designs can be specified. These are: 1. Complete Enumeration
2. Shortcut Method 3. Random Method
4. Balanced Overlap Method
Characteristic Traditional Conjoint Adaptive / Hybride Conjoint
Choice-Based Conjoint
Upper Limit Attributes 9 30 6
Level of Analysis Individual Individual Aggregate or Individual
Model Form Additive Additive Additive + Interaction
Choice Task Evaluating Full-Profile Stimuli One at a Time
Rating Stimuli Containing Subsets of Attributes
Choice between Sets of Stimuli
Data Collection Format Any Format Generally Computer-Based
Any Format Table 3.2
In short an explanation of the four techniques (Sawtooth Manual, 2007) following by the justification of the one chosen in this study.
The complete enumeration strategy considers all possible concepts, except those indicated as prohibited, and chooses each one so as to produce the most nearly orthogonal design for each respondent, in terms of main effects. The concepts within each task are also kept as different as possible, this is called minimal overlap. If an attribute has at least as many levels as the number of concepts in a task, then it is unlikely that any of its levels will appear more than once in any task.
The second method is a faster “shortcut” strategy; it makes a much simpler computation. It attempts to build each concept by choosing attribute levels used least frequently in previous concepts for that respondent. If two or more levels of an attribute are tied for the smallest number of previous occurrences, a selection is made at random. With the shortcut method, as well as with complete enumeration, an attempt is made to keep the concepts in any task as different from one another as possible (minimal overlap).
The third method is the random method and it employs random sampling with replacement for choosing concepts. Sampling with replacement permits level overlap within tasks. The random method permits an attribute to have identical levels across all concepts, but it does not permit two identical concepts (on all attributes) to appear within the same task.
3.3.1 Evaluation Method
One of the main advantages of the Sawtooth software is that the researcher does not have to come up with his own orthogonal design. Another main advantage is that the program can make ten different choice tasks for each respondent. On the downside, the program can do this only when the respondents will fill in the questionnaire on a computer that has the software on it. Due to the circumstances as a student, the author was forced to use a different approach. In this study the Sawtooth software designed three versions for each type of questionnaire (so 12 in total, three LI brand, three LI no brand, and the same for HI). These questionnaires were distributed online. The answers and the questionnaires were processed in the program SPSS and then analysed with another program: Latent Gold. The analysis will concentrate on relative importance of the attribute brand. For both high and low involvement there will follow a breakdown of the impact that the appearance of the brand makes on the relative importance of price. The results of two extreme cases will not give a significant answer on the research question: is there a difference of the importance on the brand between high and low involvement products? Nevertheless, the results may give an indication to what extend the importance of (knowing) the brand differs.
3.4 Data Collection Sample
The population of this research consist of all consumers in the Dutch market. The operational population consist of all consumers that this study can reach. At this moment, a certain discrepancy arises between whom the study wants to investigate and whom the study can investigate. The university provided the author with different mailing lists of students of economics and business at the University of Groningen. These lists were provided by teachers, one within the field of the master marketing and another teacher that had access to the lists of all Business and Economics students participating in a first and second year course of Statistics. The total number of students that fall into the sampling frame are 1400. Although it is mentioned above that a sample consisting of only students is not incorrect, and sometimes even desirable, the difference between the whole Dutch market and the students in Groningen is large. The technique of sampling that is used in this study is one of using all possible information. The sampling frame consisted of 1400 students and with the assumption that only ten percent will evaluate the questionnaire, the questionnaire was send to the whole mailing list. The three different versions of the questionnaire can be found in the appendix.
the low involvement product and ten in relation to the high involvement product. The computer program of Sawtooth software generated the choice tasks with a balanced overlap to get a realistic market simulation. Per choice tasks the respondent chooses out of three full sets of stimuli and a “none of these” option.
3.5 Analysis: Importance
The results of the study will be based on the analysis of importance in Latent Gold. For each part of the study (High/low involvement vs. brand/no brand) the program will calculate the importance per attribute (Latent Gold Choice Guide, 2006), to characterize the relative importance of each attribute. This is done by considering how much difference each attribute could make in the total utility of a product. That difference is the range in the attribute’s utility values. The maximum effect of attribute p is defined as:
(1) Maxeffp = max(Utilityp) – min(Utilityp)
These maximum effects can be compared across attributes. Often it is more relevant to compare the relative importance. The relative importances are obtained as follows:
(2) Releffp =
An example given; for a research with three attributes (brand, colour and price):
Range Relative Importance
Brand (B – C) 60-20 = 40 40/150 = 26,7
Colour (red – pink) 20 - 0 = 20 20/150 = 13,3
Price ($50 – $100) 90 – 0 = 90 90/150 = 60,0
Total 150 100 %
For this respondent price is relatively the most important attribute with 60 percent. When summarizing attribute importance for groups, as done is this study, it is best to compute importance individually and then average them. This is an automatic feature of the software program used in this study, Latent Gold.
The relative importance creates the opportunity to compare the results of the four parts. However, there is an effect in adding or subtracting an attribute. When all attributes are
Maxeffp Σp maxeffp
Table 3.3
Chapter 4 Results
In this part of the study, the results are presented. This chapter starts with some general statistics of the sample, after that the most important findings will be presented.
4.1 General statistics
This study consists of four different parts concerning the combination of brand and involvement, and each part consisted of three versions. The three versions within a part of the study measured more or less the same only with different choice tasks, these versions are taken together in the results. A schematic overview of the respondents:
Table number
The total number of respondents is 250 (18%). This is a good response compared to the expected response (ten percent). Not all of the respondents completed the whole questionnaire, which explains the discrepancy between the total number of respondents for LI no brand and HI brand (185 vs. 165) and for LI brand and HI no-brand (53 vs. 50). In addition a dozen respondents only completed one or two choice tasks or gave only the “none of these” option. This occurred most often with the notebook where the name of the brand was mentioned. A plausible explanation can be that the favourable brand of that consumer was not mentioned, for example Apple. One respondent mentioned that she did not make any choices because the colour of the notebook was not given, and that was the only attribute at which she made her decision. Because the goal of the study is not to test these options, but to test the impact of the brand (any brand) on the importance of the attributes, the respondents that did not make a choice at all, were removed from the dataset. The respondents can be put into two different groups, the group that evaluated LI brand and HI no brand (group 1) and the group that evaluated the HI brand and LI no brand (group 2). Consequently it can be tested if these groups are significantly different on age. As a result (with t = -5,400 and sig. 0,000) it
LI Brand (Group 1) LI no-brand (Group 2) HI Brand (Group 2) HI no-brand (Group 1) Male 30 (57%) 106 (57%) 98 (59%) 28 (56%) Female 23 (43%) 79 (43%) 67 (41%) 22 (44%) Total 53 (100%) 185 (100%) 165 (100%) 50 (100%) Male age 23,1 21,8 21,8 23,1 Female age 22,8 20,5 20,6 22,8 Mean age 22,9 21,2 21,3 23 Table 4.1
showed that the groups actually do differ on age. The second group is significantly younger than the first group, therefore the groups are not totally representative. Nonetheless the groups have a high level of similarity, both groups are students within the same field of study (economics) and all native Dutch speakers.
4.2 Results
The results of the analysis are shown in four components. The results will give the maximum and relative importance of the attributes and the utilities of every level of the attribute. Everything is derived from an analysis with Latent Gold using a one class solution. In all cases the first level of a specified attribute will be set to zero to clarify the difference in utility.
4.2.1 Low involvement without brand
Table 4.2a shows the maximum and relative importance of price in contrast with the importance of packaging. Price is the main factor why consumers chose the macaroni without any knowledge of the brand. An examination of the utility of the levels shows that most respondents chose the cheapest level more often than the middle or high priced version. Packaging is relatively not very important; nonetheless there is a preference for the quality re-closable plastic bag. The differences in utilities for the different levels are all significant (p<0,000).
Attributes Levels Class 1 Wald p-value
€ 0,49 0,000
€ 0,89 -1,8283
Price
€ 1,29 -3,6325
1256,394 0,000
Normal plastic bag 0,000
Box 1,0328
Packaging
Quality re-closable plastic bag 1,2772
181,5647 0,000
Attribute Maximum Relative Price 3,6325 0,7399
Packaging 1,2772 0,2601 Table 4.2a
Importance per attribute
4.2.2 Low involvement with brand
In Table 4.3a it is clear that the maximum effect of price has declined. Price loses 9 percent point of is importance to brand, that is an impact of 12 percent. Packaging suffers a harder impact; it loses 18 percent point, which comes to an impact of 72 percent. Brand has a relative importance on this low involvement product of 28 percent. The effect of brand equity is visible in Table 4.3b. Evaluating the utility it becomes clear that the two branded products have a higher utility than the private label product. This was expected, just like the higher utility of Gran’Italia which is a top shelf brand. The utility of the price has shifted. Naturally the introduction of the brand has effect on the utility of the price. Most remarkable is the shift of the middle price. The middle price version of the product has come notable closer to the low priced version of the product. This can be explained by the interaction effect of price and brand. When a preferred brand has not the preferred price (low), the preference in brand can be decisive in choosing a higher priced version of the product. People become less price-sensitive in this case. The shift in utility relative to the low priced version is presented in the next graph.
Attributes Maximum Relative Brand 1,2313 0,2780
Price 2,8758 0,6492
Packaging 0,3228 0,0729
Attributes Levels Class Wald p-value
Honig 0,000 Gran'Italia 0,1985 Brand Private Label -1,0329 65,5199 0,000 € 0,49 0,000 € 0,89 -0,813 Price € 1,29 -2,8758 200,1216 0,000
Normal plastic bag 0,000
Box 0,2905
Packaging
Quality re-closable plastic bag -0,0324
6,1048 0,047
Table 4.3a Importance per attribute
-4 -3,5 -3 -2,5 -2 -1,5 -1 -0,5 0 No-Brand Brand € 0,49 € 0,89 € 1,29
4.2.3 High involvement without brand
Attributes Levels Class1 Wald p-value
€ 899 0,000 € 1.199 -1,5815 Price € 1.499 -3,3585 223,478 0.000 1.7 GHz 0,000 2.1 GHz 1,122 Speed 2.4 GHz 1,9408 103,5596 0.000 100 Gb 0,000 300 Gb 1,2526 Size 500 Gb 2,0985 125,6163 0.000
As expected attribute price is the most important attribute for choosing a notebook. The outcome that a notebook with the lowest price was preferred more than a middle priced or a
Attribute Maximum Relative Price 3,3585 0,4540
Speed 1,9408 0,2624
Size 2,0985 0,2837 Figure 4.1
Shift in price between branded and none branded choice
Table 4.4a Importance per attribute
high priced notebook was against expectations. Anticipated was that when information of the brand and the quality be lacking the price was going to be a substitute for quality. There is a plausible explanation for this based on the literature. Price can also be a sacrifice indicator, the amount of money consumers have to sacrifice and can not spend at something else. Two attributes that cover an element of quality are the speed of the processor and the size of the hard drive. Both attributes show that the notebook is more preferred when the size is large and the speed is high. Consequently one might say that the respondents chose the option that is best suitable for them: lowest price, highest speed and largest memory. In this case speed and size, with respectively 26 and 28 percent relative importance, were important attributes but price was decisive.
4.2.4 High involvement with brand
Attributes Maximum Relative Brand 0,7555 0,1403 Price 2,7452 0,5096 Speed 1,0209 0,1895 Size 0,8651 0,1606
Attributes Levels Class1 Wald p-value
HP 0,0000 Fujitsu-Siemens -0,1223 Brand Asus -0,7555 70,0791 0,000 € 899 0,0000 € 1.199 -0,7045 Price € 1.499 -2,7452 585,8905 0,000 1.7 GHz 0,0000 2.1 GHz 0,9331 Speed 2.4 GHz 1,0209 133,8109 0,000 100 Gb 0,0000 300 Gb 0,8651 Size 500 Gb 0,6997 92,1858 0,000
The findings for the high involved product are extraordinary and against expectations. Expected was that the importance of price would decline as soon as the brand name was introduced. However, the results show an increase in relative importance of price of more than five and a half percent points, which is a positive impact of 12 percent. A decrease is
Table 4.5a Importance per attribute
visible in the maximum effect of the attribute. Although the expectation was that both the relative and the maximum importance would decline, the increase in relative importance is not significant. The attribute size lost the most of the attributes: 12 percent point, which is an impact of more than 43 percent. Speed only lost seven percent point and that comes to a total impact of almost 28 percent. Brand itself has an importance for the notebook of 14 percent. It is remarkable that brand has the lowest importance of all attributes by deciding to buy a notebook. Even though the price became more important the middle priced option was relatively more preferred to the lower price option in contrast to the decision process where the brand was not known.
Similar to the low involvement product there is possibly an interaction effect between price and brand. The middle priced option becomes more attractive in combination with the preferred brand. The consumer becomes more indifferent. Where speed and size become less important to the consumer, the consumer has also different preferences to the levels of the attribute. The relative distance between the middle and lowest option stay somewhat similar, however the distance between the middle option and the highest solution is almost disappeared. In the case of the size the middle option is more preferred than the highest option. The highest speed for the processor is still preferred, but with a minor difference with the middle speed. Most preferred is the notebook of HP of 899 euros with a processor speed of 2,4 GHz and an hard drive of 500 Gb.
Figures 4.2 till 4.4 show the path-worth’s of the following attributes: Size, Speed and Price
0 0,5 1 1,5 2 2,5 1.7 GHz 2.1 GHz 2.4 GHz No-Brand Brand
In figure 4.2 and 4.3 it is visible that the consumer indeed becomes indifferent between the middle and high option for speed and price. In this case the brand or price was the decisive attribute. It is noticeable that the notebooks with the smallest hard drive memory and/or the lowest speed of the processor are not preferred by the consumers.
