The Importance of Brands in the Age of
Transparency: A Choice-Based Conjoint Analysis
Master ThesisMSc in Business Administration – Marketing
Valerie Blank – 10279652 University of Amsterdam
Master Thesis
MSc in Business Administration – Marketing Track Supervisor: J. Demmers MSc.
Statement of originality
This document is written by Valerie Blank, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of Contents
Abstract ... 4 Introduction ... 4 Literature Review ... 8 Transparency ... 8 Benefits of Transparency ... 9Transparency and Consumer Choice Behavior ... 10
Brands ... 13 Methodology ... 16 Conjoint Design ... 16 Pre-test ... 17 Attributes ... 17 Data Collection ... 20
Statistical Model and Procedure ... 20
Results ... 22
Discussion ... 30
General Conclusion ... 30
Theoretical and Practical Implications ... 32
Limitations and Future Research ... 34
Abstract
!
Nowadays consumers are concerned about their environment and society, demanding more and more transparent products. This demand explains the rise of brands such as Tony Chocolonely and Clipper Teas. These brands communicate openly about their business practices. This research examines how transparency influences consumers’ choice for brands. Now that more brands are becoming transparent, a question that arises is: to what extent transparency is important in day-to-day choices among consumers? This is done by comparing the relative importances of different product attributes. A Choice-Based Conjoint Analysis was used to establish utilities for and relative importances of different product attributes. The product attributes were transparency, price, quality and brand. The Choice-Based Conjoint Analysis led to the findings that transparent companies are preferred over translucent and opaque
companies. Moreover, transparency through information disclosed on the packaging was preferred to transparency through information disclosed on the website’s homepage. Transparency was found to be more important than a product’s quality perceived by previous consumers. No significant difference was found between the relative importance of transparency and price. The main result of this research shows that there are two distinct groups in the age of transparency, that is, consumers who consider transparency as relatively more important than brands and consumers who consider brands as relatively more important than transparency. Overall, this research shows that in the transparent age, brands are less used for risk reduction but still function as social demonstrance. !
Introduction
Companies have a natural interest to keep their business opaque to the outside world (Sinha, 2000). However, nowadays companies are under immense pressure to be more transparent (Carter & Rogers, 2008). By 2004, approximately 68% of the Global 250 businesses generated a separate report, which included information on the
environmental, social, and economic impact of their businesses (KMPG, 2005). In 2005, 80% of the reports included this information (KMPG, 2005). Today’s
consumers are concerned about their environment and society, demanding more and more transparent products and services (Bhaduri & Ha-Brookshire, 2011). This demand explains the rise of companies such as Tony Chocolonely and Clipper Teas. These companies are openly communicating about their business practices. Taking a glance at the websites of these companies, information can be found on production processes, employee conditions and products. The market for such brands is expected to further increase (Slavin, 2009), as consumers are looking for ethical and
environmentally driven companies (Singh, 2015).
The growing demand for transparency among consumers comes from the growing need for information. The availability of and convenient easy access to information increased as Internet usage increased (Fournier & Avery, 2011). This has led to a more critical consumer with a higher information need (Rezabakhsh et al., 2006). On the Internet, information travels at high speed. Consumers can easily find critical reviews, reports and blogs. As a result, it is difficult for companies to keep information from consumers. Everything that can be exposed will be exposed (Fournier & Avery, 2011).
Companies have also become aware of positive effects of transparency and use transparency as a marketing tool. Previous research has showed that transparency enhances purchase intention (Kang & Hustvedt, 2014; Bhaduri & Ha-Brookshire, 2011), consumer satisfaction (Simintiras et al., 2015; Matzler et al., 2006) and
consumer loyalty (Singh, 2015). Now that more companies are becoming transparent, a question that arises is: to what extent is transparency relevant in day-to-day choices amongst consumers?
Cohn and Wolf (2013) argue that transparency and honesty are important in consumer decision-making, even more so than brand names. They found that between
60% and 80% of the consumers take a brand’s transparency and honesty into
consideration when purchasing a product. These results are obtained through a survey. However, one of the drawbacks of a survey is that the data collected are unlikely to be as detailed as those collected by other research methods (Lewis & Saunders, 2012), such as a Choice-Based Conjoint Analysis. This analysis involves choice tasks that are seen as realistic and comparable to the way consumers make choices in real world situations (Alves et al., 2008).
The Choice-Based Conjoint Analysis can be used to weigh the relative importance of transparency and brands in consumer decision-making. Transparency can be perceived differently depending on the brand, or can become less important if a brand is used as, for example, a status symbol (Fischer, Völckner & Sattler, 2010). Brand equity may thus have an impact on transparency, but transparency can also influence brand equity. On the one hand, brands can benefit from transparency. Brands can incorporate transparency in their marketing communications to
differentiate themselves from competitors (Bhaduri & Ha-Brookshire, 2015). On the other hand, while the importance of transparency is increasing, Cohn and Wolfe (2013) discuss that the importance of brands is decreasing. Consumers were
depending on brands to reduce risk, because a lack of information about the quality of a product before purchase made consumers use brand names as a quality guarantee (Fischer, Völckner & Sattler, 2010). Brand equity helped the consumers mitigate this risk by letting brands serve as a quality signal. With brands using transparency, more information is available. Therefore, reliance on brand names as a form of insurance is decreasing, which may lead to a decline in the importance of brands. Thus, whilst transparency is increasing brand equity, it seems to be making brands less important, creating a paradox.
However, previous research has not provided sufficient empirical proof to support this theory, so far. Therefore, this research addresses this gap in the literature by answering the following research question: How does transparency influence
consumers’ choice for brands?
Although transparency has gained some attention in the existing literature, there is still much to explore about the effect of transparency on consumer behavior (Granados, Gupta & Kaufmann, 2012). This research expands the transparency literature and adds to the theory of transparency on consumer behavior. More
compared to brand, taking price and quality into account. With the use of a Choice-Based Conjoint Analysis, this study tries to provide support for the findings of Cohn and Wolfe (2013).
A better understanding of the relative importance of transparency can have an impact on the positioning of products. If it turns out that brands are still more
important than transparency, despite the increasing demand of transparency by consumers, then brands can keep investing in marketing campaigns to enrich brand equity. However, if results show that brands are becoming less relevant as a result of the increase of transparency, consumers might switch more easily to alternative brands. New brands might, in that case, derive more benefits from focusing on the distribution of product information instead of spending the majority of resources on building brand equity.
First of all, this research will provide an overview of the relevant literature. Subsequently, the research design and other methodology issues will be explained, and where after, the results will be discussed. Finally, there will be a discussion, which will include conclusions, implications and suggestions for future research.
Literature Review
This part provides an overview of existing literature on transparency, consumer behavior and brands. First, there will be a review of the concept ‘transparency’. Subsequently, the benefits of transparency and the effect of transparency on
consumers’ choice behavior will be discussed. Finally, a reflection on brands in the age of transparency is presented.
Transparency
In the existing literature transparency has numerous definitions, of which a majority encompasses the concept of companies sharing information with consumers (Singh, 2015; Carter & Curry, 2010; Vishwanath & Kaufmann, 2001; Lamming, 1993; Lamming et al., 2001). For instance, Carter and Curry (2010) define transparency as “information revealing the allocation among agents in a supply-chain of proceeds from the sale of a product or service” (Carter & Curry, 2010, p. 760). In their research, Carter and Curry (2010) refer to pricing transparency. Besides pricing transparency, transparency can take other forms such as cost transparency (Sinha, 2000; Simintiras et al., 2015) or supply chain transparency (Lamming et al., 2001). Hultman and Axelsson (2007) distinguish four types of transparency: cost/price, organization, technology, and supply chain. They also state that three facets are relevant for all types of transparency: degree, direction and distribution. Firstly, the degree of transparency refers to the extent to which information is shared between companies and consumers. Companies can be transparent, translucent or opaque (Lamming, et al., 2001). Secondly, direction of transparency includes the flow of information. The flow of information may move in a single direction or two
directions. Thirdly, the distribution of transparency can be direct or indirect, meaning that transparency may be present either in a specific relationship or in a relationship that is connected to the central relationship.
In this research transparency is defined as “the visibility and accessibility of information especially concerning business practices” (Bhaduri & Ha-Brookshire, 2011, p. 136). Hence, companies who are not transparent will neglect buyers’ access to information, and/or they will provide irrelevant information or information that is
misrepresented, inaccurate or not well timed. According to Vishwanath and Kaufmann (2001), transparency should encompass dimensions such as access, comprehensiveness, relevance, quality and reliability. First of all, transparent
companies should provide available and accessible information. Second, transparency should involve relevant information. Ensuring relevance is challenging because relevance may depend on the needs of the user. Finally, transparent companies should provide qualitative and reliable information, meaning fair, reliable, timely, complete, consistent and presented in clear and easy terms.
Benefits of Transparency
Companies are under a lot of pressure to adopt a transparent way of doing business through the rise of the Internet and growing concern about the manner in which companies perform (Carter & Rogers, 2008). However, when companies are operating transparently, both the company and consumer benefit. The increased availability of and easy access to this information is one of the more empowering forces of the Internet (Fournier & Avery, 2011). Transparency is reducing information asymmetries between consumers and companies (Grewal et al., 2003), and leads to increased consumer power (Simintiras et al., 2015). Consumer empowerment refers to “the subjective state that is evoked by perceptions of greater personal control”
(Simintiras et al., 2015, p. 1971). Through transparency consumers can share their personal experiences concerning products and services (Simintiras et al., 2015). Additionally, transparency leads to an overall improvement in consumer welfare (Carter & Curry, 2010; Gu & Wenzel, 2011), and deals with uncertainty avoidance (Vishwanath, 2003). Transparency can also help to make easier and better choices (Scheibehenne, Greifender & Todd, 2010).
The availability of information can also benefit companies (Simintiras et al., 2015). For example, transparent companies tend to have more loyal customers than opaque companies (Singh, 2015; Simintiras et al., 2015). Moreover, Strutnin (2008) found that maintaining a transparent supply chain is relevant for building consumers’ brand loyalty. However, Boniface et al. (2012) did not found a relationship between transparency and consumers’ brand loyalty. Nonetheless, transparency helps build relationships between consumers and companies (Lowe, 2015), because transparency
is one of the basic conditions in establishing positive relationships between consumers and companies (Kang & Hustvedt, 2014).
Beside increasing brand loyalty and establishing relationships between consumers and companies, transparency could also create differentiation for companies and have an impact on profitability (Singh, 2015). More specifically, Margolis, Elfenbein and Walsh (2007) showed that transparency has a positive effect on a firm’s corporate financial performance. Another reason for companies to be transparent is that it enhances the value proposition (Kumar, 2006), so transparency can create a competitive advantage for a company.
In the age of transparency, companies seem to have no choice but to be transparent and provide full disclosure (Fournier & Avery, 2011). A company that is less open signals it has something to hide (Bertini & Gourville, 2012). Additionally, companies that see transparency as a threat may eventually find themselves being steered toward adopting more transparency (Singh, 2015). Perhaps for the best, because products of a transparent company tend to be chosen more often than products of a company that is operating opaquely (Carter & Curry, 2010).
Transparency and Consumer Choice Behavior
So far several scholars made an attempt to better understand the effects of
transparency on consumer choice behavior. Cohn and Wolfe (2013) found that about 70% of the consumers take transparency into consideration when buying a product. When transparency is included in the consumer’s consideration set, transparency is found to have a significant impact on choice behavior. For example, Carter and Curry’s (2010) results demonstrate that, on average, consumers value transparency over opaqueness. Information about transparent pricing is valued among the consumers, and may lead consumers to select the more expensive of two identical products. This runs counter to expectations, because Sinha (2000) states that transparency in the Internet age will cause lower sales, turn products into
commodities, lower consumer loyalty and increase perceptions of unfairness. Sinha (2000) views transparency from an economic perspective, which presumes that consumers make choices based upon price, when the product’s other attributes are held constant. However, Carter and Curry (2010) argue that transparency encourages
consumers to expand his or her utility function to also involve social aspects, resulting in high consumers’ preferences for transparent products.
In the existing literature, studies have shown that consumers have positive intentions towards transparent companies. According to Lafferty and Goldsmith (1991), consumers are more willing to purchase a product from a transparent company than from an opaque company. Other scholars have come to the same conclusion (Maxwell, 2002; De Pelsmacker et al., 2005). In addition, De Pelsmacker et al. (2006) found that a reason not to buy a product was that a company was not disclosing enough information. Therefore, it seems that consumers will give preference to high levels of transparency. Transparency is expected to be relatively important in consumers’ choice behavior. This leads to the following hypothesis:
H1: Consumers prefer transparent companies to translucent and opaque companies. Internet has become a major source of information (Peterson & Merino, 2003), but other sources of information continue to be favored over the Internet by many
consumers and in many situations (Ratchford et al., 2001). For instance, McNeill and Wyeth (2011) indicate driving forces in the decision-making process of purchasing. For instance, all respondents mentioned packaging as driving product choices, so packaging is a relevant factor in the decision-making process. The importance of packaging design and the use of packaging as a vehicle for communication and branding has increased (Rettie & Brewer, 2000), as it takes on a similar role to other marketing communications channels. This role is being reinforced by the fact that consumers may not think deeply about purchasing products before they go into the store. Indeed, 73% of purchase decisions are made at the point of sale (Connolly & Davidson, 1996). POPAI study (2014) confirms this high rate. This study showed that the in-store decision rate was 76% in 2012 and that this in-store decision rate
continued to grow to 82% in 2014. Moreover, 90% of the consumers make a purchase only after examining the package (Urbany, Dickson & Kalapurakal, 1996).
In addition, Behaeghel (1991) and Peters (1994) consider that packaging could be the most important communication tool, because it reaches almost all buyers, and it is present at the crucial purchasing moment. A product’s website is less expected to reach all buyers and may not be present at the crucial moment of purchasing. Therefore, it can be considered that information disclosed on the packaging is
preferred over information disclosed on the website’s homepage. This leads to the following hypothesis:
H2: Consumers prefer transparency through information disclosed on the packaging to transparency through information disclosed on the website.
Choice behavior is also dependent on several other product attributes. To make a decision, consumers require, at least, information about the price and the quality of choice alternatives they consider purchasing (Simintiras, et al., 2015). Decisions made in absence of this adequate and relevant information may result in the use of
alternative cues, shortcuts and coping strategies (Simintiras et al., 2015). Previous research suggests that price and quality are important determinants of shopping behavior and product choice (Bishop, 1984; Dickson & Sawyer, 1986).
Previous research mainly shows that consumers consider transparency and openness after price and quality. In Bhaduri and Ha-Brookshire’s research (2011), most participants ranked price and quality higher than transparency. Bhaduri and Ha-Brookshire (2011) also show that consumers who perceived a higher value in buying products from transparent businesses were willing to pay a higher price or
compromise a little on quality. Furthermore, in their study among consumers from the United Kingdom, United States of America and China, Cohn and Wolf (2013) found that the three most important aspects for purchase intentions are price, quality and transparency. About 70% of the consumers take transparency into account, but only after reviewing the price and the quality of the product. Only Chinese consumers think that transparency and honesty are more relevant than the price of the product. Thus, it is likely that price and quality are more important for consumers than transparency. Therefore, the following is hypothesized:
H3: Price and quality are relatively more important for consumers than transparency. In addition to price and quality, the brand of the product can also play a role in
consumers’ choice behavior. Brand equity is thought to drive brand purchase (Nenycz-Thiel & Romaniuk, 2009).
Brands
A brand is here defined as “a name, term, sign, symbol, or design, or combination of them which is intended to identify the goods and services of one sellers of group of sellers and to differentiate them from those of competitors” (Kotler, 1991, p. 442). The impact of a brand on consumer choice can be measured through brand equity. Brand equity refers to “the customer’s subjective and intangible assessment of the brand, above and beyond its objectively perceived value” (Lemon, Rust & Zeithaml, 2001, p. 22). This subjective assessment is determined by consumers’ brand
knowledge. Brand knowledge is built through brand awareness and brand image. Brand awareness concerns the consumer’s ability to identify a brand, and the ease with which it does so. Brand image concerns the brand perception that a consumer holds in memory (Keller, 1993).
Brands have been an important factor in the decision-making process when purchasing a product or service (Fischer, Völckner & Sattler, 2010). Cobb-Walgren, Ruble and Donthu (1995) showed in their study that greater brand equity results in a significantly greater consumer preference and purchase intention. This is applicable for both the service and product category. This positive impact of brand equity may result in a better firm performance (Kim, Kim & An, 2003).
An explanation for this result is that brands help consumers make purchase decisions when hardly any information is provided. When consumers purchase a product, they take on risk. A way in which consumers mitigate this risk is through their reliance on brand names as heuristics (Delvecchio, 2001). Consumers’ quality perceptions are based on extrinsic cues, such as price and brand names (Sprott & Shrimp, 2004). Consumers perceive brands with high equity, which is largely built through (expensive) marketing campaigns, to be of superior quality (Aaker, 1996). Consumers depend on this brand equity to serve as a quality signal. Therefore, when confronted with risk,consumers often rely on this signal and are more motivated to choose a brand as a form of insurance (Richardson et al., 1996). Indeed, brands become more important when less information about different product attributes is provided (Degeratu, Rangaswamy & Wu, 2000). More specifically, in an offline-shopping environment more risk is experienced due to unavailability of information. Furthermore, Fischer, Völckner and Sattler (2010) indicated two factors determining the need for a brand, that is, social demonstrance and risk reduction. However, in the
same study, risk reduction is found to explain the importance of brands more than social demonstrance.
With information, the risk of purchasing is reducing. As a consumer’s
knowledge of a product increases, dependence on heuristics, such as brand names and price, decreases (Delvecchio, 2001). On the one hand, consumers with relatively high levels of product information are able to evaluate brand on an objective level based on the product’s attributes (Rao & Monroe, 1988; Venkataraman, 1981). On the other hand, consumers lacking such information will be likely to depend on a heuristic when assessing product quality.
Rather than by interest in a brand or in information, transparency is caused by the role of risk reduction (Christensen, 2002). Information matters only for highly devoted consumers of a certain brand. Transparency is the means by which the risk associated by the purchase is mitigated, rather than the goal itself (Morgan, 2009). Thus, transparency and brands are to some extent functioning in the same way, to reduce risks.
Apart from the fact that we live in a transparent age, consumers may be still in the need for brands. For example, a specific brand may have superior quality. In Cohn & Wolfe’s (2013) research, consumers from US, UK and China prefer quality to transparency, so if a brand excels on quality consumers may still prefer that brand. Moreover, Fischer, Völckner and Sattler (2010) expected that, beside the functional benefit of risk reduction, a brand can function as social demonstrance. A brand can serve as a symbolic device that allows consumers to project their self-image (Levy, 1959). People try to keep and enhance their self-image by buying certain products (Shrauger, 1975). Consumers can also use brands to communicate to others what kind of person they are or would like to be (Escalas & Bettman, 2005). More specifically, they employ brands as status symbol or as a means to signal group membership. Despite the fact that risk reduction is a more important factor for brand relevance, social demonstrance was found to be a factor for brand relevance as well.
Therefore, it can be expected that for some consumers brands will preserve their importance in spite of increased transparency. In addition, consumers may also be extremely loyal to a brand. A reasonable question is why consumers keep being loyal to brands when they can switch to better alternatives due to transparency. Researchers have noted a decline of loyal segments (East & Hammond, 1996). What this means is that, for consumers to become and remain loyal, they must believe a
certain company’s product continues to offer the best choice alternative. They also must do this while naively avoiding communications from competitors that argue that the loyalist’s consumable is no longer the most efficient, lowest priced, of the highest quality, and so forth (Oliver, 1999). Hence, it can be expected that consumers might still be in the need for brands, despite the increased availability and access to
information.
Altogether, consumers were depending on brands to reduce risks, because a lack of information about the product made consumers use brand names as quality signals. However, as consumers possess more information and rely less on heuristics, consumers are less in need for these quality signals. In that case, brands become less important in the eyes of the consumers. The disclosure of information reduces the perceived risk of purchase. Nevertheless, consumers may still need brands, because brands also function as social demonstrance. Thus, brands will preserve their importance despite increased transparency. Moreover, consumers may also be extremely loyal to a brand, and therefore, ignore communications from competitors that imply that they are the best choice alternative. Thus, it can be expected that when the relative importance of transparency increases, the relative importance of brands decreases and vice versa. Hence:
H4: There is a negative correlation between the relative importance of transparency on one hand and brands on the other.
The relative importance of transparency does not exclude the relative importance of brands and the relative importance of brands does not exclude the relative importance of transparency. Therefore, it can be argued that there are two distinct groups, that is, consumers who prefer high levels of transparency to premium brands and consumers who prefer premium brands to high levels of transparency. Thus, there are different consumer segments in the age of transparency. Specifically:
H5: There are two distinct groups of consumers, one that considers transparency as relatively more important than brands and one that considers brands as relatively more important than transparency.
Methodology
For understanding how important certain factors are in decision-making processes of consumers, a Choice-Based Conjoint Analysis (CBC) is conducted. A CBC is different from other methods used for assessing relationships or generating importances in that the judgments are not constrained to individual attributes in insolation but are made on the basis of a range of coexisting attributes (Aspinall et al., 2005). Before the CBC is conducted, a pre-test is done to determine the levels of the attributes. Furthermore, the data collection and the statistical model are discussed.
Conjoint Design
A CBC is used to measure consumers’ preferences. CBC is a trade-off analysis (Green, Krieger & Wind, 2001), which involves choice tasks that offer greater confidence that the preferences elicited from consumers reflect the kinds of choices people make in their daily lives (Alves et al., 2008). Consumers evaluate a set of attributes. More specifically, consumers are asked to choose an option that they prefer the most based on different attributes. Letting consumers determine their preference multiple times, CBC generates utilities, which are values placed on the different attributes.
CBC was carefully chosen for this research, although different types of conjoint analysis exist. First, CBC is a relatively fast method. A data collecting method that is less time consuming is more likely to get a high response rate (Lewis & Saunders, 2012). Furthermore, data collection includes simulated purchase
decisions, which is a more realistic and relatively easy task for respondents (Desarbo et al., 1995). It represents market behavior (Huber et al., 1992). Third, product-specific attributes or levels (e.g., unique features of a brand or different price levels) can be easily accommodated and brand-specific utilities can be estimated (Desarbo et al., 1995). Finally, CBC analysis is thought to be a reliable tool for measuring brand utility (Aaker, 1996; Keller, 1993). It is seen as the best tool for marketers to uncover consumers’ trade-offs between different brands and attributes (Green, Krieger & Wind, 2001). CBC is previously used to measure the effect of transparency (Carter & Curry, 2010). In this study, there are four attributes for a chocolate bar: transparency, price, quality and brand. The levels of these attributes are determined through a pre-test.
Pre-test
A pre-test was done among 30 respondents to determine the levels of the attributes. Another purpose of the pre-test was to assure that the distances between the levels of the attributes are perceived as similar. The data is collected through an online survey with the use of Qualtrics. Non-probability sampling techniques are used to reach respondents and potential respondents are approached through e-mail and WhatsApp. The survey was formulated in Dutch and only Dutch-speaking people have
participated.
For every attribute, multiple levels are tested in the pre-test. Respondents had to indicate to what extent they found the multiple levels open and transparent, cheap or expensive, and of low or high quality on a 7-point Likert scale. In order to
determine brands’ equities, the scale of Yoo and Donthu (2001) was used. They adjusted Aaker’s (1991, 1996) and Keller’s (1993) brand equity scale through combining brand awareness and brand associations and including overall brand equity. Altogether, respondents evaluated four brands and two private labels via Yoo and Donthu’s scale (2001).
Brand equity was measured for Tony Chocolonely, Milka, Cote d’Or, Verkade, Albert Heijn, and Jumbo. Cronbach’s Alpha is calculated for the multidimensional consumer-based brand equity scales, and all scales had a Cronbach’s Alpha above .7 (see table 1). Next, a Repeated Measures Anova is conducted to see if the levels and distances between these levels were significant. This was also done for the price levels, quality levels and transparency levels. All levels were significant, but more important, the differences between the levels were significant. Therefore, the levels for the four attributes can be determined. Moreover, it can be concluded that distances between the levels of one attribute are similar to distances between the levels of another attribute.
Attributes
Through the Repeated Measures Anova levels within and between attributes are compared and determined. For price three different levels are selected: low (€ 1,49), medium (€ 1,99) and high (€2,49). Usually chocolate bars are valued around these
prices. There are two possible options for quality: low (three stars) and high (four stars). Notice have been given in the survey that this quality is perceived by previous consumers. As to brand, there are three different options, two brands and one private label brand. Tony Chocolonely, Cote d’Or and Albert Heijn’s private label are selected because pre-test results showed that these brands significantly differed from each other. Additionally, they were given the greatest brand equity within their branch. The transparency attribute has four levels. Firstly, an opaque option, which states that information about the production process and cacao farmers is not provided to consumers. Secondly, a translucent option, which reflects information about the production process and cacao farmers on a back-page of the product’s website. Finally, there were two transparency options. One included easy accessible information on the homepage of the product’s website and the other included
information on the packaging of the product. An overview of the attributes and levels is presented in table 2.
Table 1
Cronbach’s Alpha of brand equity scales Variables Cronbach’s Alpha (α) Tony Chocolonely .89 Milka .78 Cote d’Or .88 Verkade .87 Albert Heijn .91 Jumbo .91
Table 2
Levels of the attributes
Attributes Level 1 Level 2 Level 3 Level 4
(1) Transparency Information about production process and cacao farmers is not provided by the chocolate bar’s company (opaque) Information about production process and cacao farmers is disclosed on the back-page of the chocolate bar’s website (translucent) Information about production process and cacao farmers is disclosed on the homepage of the chocolate bar’s website (transparent) Information about production process and cacao farmers is disclosed on the packaging of the chocolate bar (transparent) (2) Price € 1,49 € 1,99 € 2,49 (3) Quality ★★★ ★★★★ (4) Brand Tony Chocolonely
Cote d’Or Albert Heijn’s chocolate bar
Thus, there are 54 different options (3 x 2 x 3 x 3) in total. Respondents did not had to compare all of these options separately. Despite the fact that the conjoint analysis did not include a full-profile design, it can predict a consumer’s choice with a high accuracy (Omre, 1998). Sawtooth Software recommends a range of 8 to 15 choice tasks, while Johnson and Omre (1996) did not found a loss of reliability of the first 20 tasks within a CBC. However, it is likely that respondents are less patient and conscientious with long CBC surveys than respondents of 20 years ago. Therefore, each respondent was confronted with 15 choice tasks from which she or he had to choose one of the four alternatives. Within the four alternatives a ‘none-option’ is included because, as in the real world, consumers can decline to purchase (Omre, 2009). Additionally, Johnson and Omre (1996) found that respondents did not choose the none-option to, for example, avoid difficult tasks, but rather chose the none-option because the alternatives were not attractive. Beside the three random combinations of the attribute’s levels and the none-option, there is a fixed combination added in the
middle of the survey. Fixed tasks permit identification and removal of inconsistent respondents. This is especially relevant when using Hierarchical Bayesian Estimation (Omre, 2014). The fixed task consisted of the lowest price, highest quality level, highest level of transparency and the brand with the greatest brand equity (i.e., Tony Chocolonely). After finishing all these tasks, questions about the frequency of eating chocolate and demographic characteristics are asked. Demographic questions
included sex, age, education, and occupation. According to Johnson and Omre (1996), the more tasks are added in the survey, the fewer respondents are needed and vice versa. Nonetheless, sample sizes for conjoint studies generally range from about 150 to 1.200 respondents (Omre, 2010).
Data Collection
The data was collected through an online survey with the use of Sawtooth Discover Software. To reach respondents non-probability sampling techniques, that is, self-selection sampling and snowball sampling, have been used. The URL-link to the survey was posted on social media and distributed through e-mail. The survey was written in Dutch and solely distributed to Dutch-speaking people. A total of 195 respondents have fulfilled the survey. Most people were younger than 25 years old (30.8%) and the division of men and women was, respectively, 64.5% and 35.5%. Finally, the majority of the respondents enjoyed a university education (38.5%). This study is done according to the code of ethics of the University of Amsterdam.
Personal data was not collected or utilized in this research.
Statistical Model and Procedure
The analysis of conjoint data from choice experiments typically employs a
multinomial logit model to estimate the part-worths (Allenby, Rossi & McCulloch, 2005):
(1) Pr ! ! = ! exp[!!′ !!]
!"# ! [!!!!!]
vector of attribute-levels that describes the choice alternative and !! is a vector of regression coefficients that indicate the part-worths of attribute levels.
The deterministic utility, !!!!
!, is an additive part-worth function of the price levels, quality levels, transparency levels and brand names:
(2) !!!!
! = ! !!!,!!!!,! +!!!!,!!!!,!+!!!!,!!!!,!+!!!!,!!!!,!+!!!!,!!!!,!+!!!!,!!!!,! +!!!!,!!!!,! +!!!!,!!!!,! +!!!!,!!!!,!+!!!!,!!!!,!+!!!!,!!!!,! +!!!!,!!!!,!
The subscripts B1, B2 and B3 indicate the brands. Tony Chocolonely and Cote d’Or are the two brands and Albert Heijn’s chocolate bar is the private label brand. The quality levels (i.e., three and four stars) are represented by the subscripts Q3 and Q4. Furthermore, subscripts P1, P2 and P3 denote the price levels (i.e., € 1.49, € 1.99 and € 2.49). Finally, transparency levels are indicated by the subscripts T1, T2, T3 and T4. T1 and T2 denote transparency by means of, respectively, information published on the company’s homepage and through information provided on the packaging. T3 denotes translucent business practices and T4 denotes opaque business practices.
The parameters !! are the part-worth utilities to be estimated for the brand, quality, price and transparency levels. One level within each attribute constitutes as the reference category for the other attribute levels. The following attribute levels are chosen: price-level € 1.49, three stars for quality, an opaque company and the brand Albert Heijn.
The coefficients, !!, include an index I to account the fact that in this study part-worth utilities are estimated at the individual respondent level using a
Hierarchical Bayesian Estimation (HB). The basic assumption in most HB applications is that the individuals (in this case, consumers) are members of one population. More specifically, the individual-level parameters, !!, are proposed to originate from a distribution with unknown population mean B and unknown covariance matric Ω. This can be written as follows:
(3) !! = ! + !ζ!!!!!!~!!(0, !)
In other words, HB is able to provide estimates of individual part-worths given only a few choices by each individual (Lenk et al., 1996). This is possible through deriving information from population information, describing the preferences of other
respondents in the same dataset (Omre, 2016). Nevertheless, HB generates utility values that can contribute to an accurate analysis of attribute importances (Sawtooth Software, 2013). Therefore, this model is positively reviewed in many articles (Lenk et al., 1996).
The HB analysis that is performed to calculate the individual attribute utilities is conducted using Sawtooth Discover Software. Subsequently, the generated utilities are analyzed in SPSS. Descriptive statistics, skewness, kurtosis and normality tests are computed. Generally, normality tests showed normal distributions for all variables. To find support for the hypotheses, various analyses have been done. To test whether differences in utilities for levels of attributes were significant, Paired Samples T-Test and Repeated Measures Anova are used. A Paired Samples T-Test is conducted for attributes with two levels and a Repeated Measures Anova is conducted for attributes with three or more levels. Before conducting these analyses,
assumptions of normality and sphericity are checked. Special attention is paid to the results extracted from the Repeated Measures Anova, because these results are required to test the first and second hypotheses.
Furthermore, the relative importances of the product attributes are calculated using Omre’s method (2010). Descriptive statistics, skewness, kurtosis and normality tests are computed for these new variables. Next, a Repeated Measures Anova is conducted to test the third hypothesis. Furthermore, a Bivariate Pearson Correlation is used to test the fourth hypothesis. Eventually, a Cluster Analysis is conducted to test the fifth hypothesis. Grouping similar consumers is a fundamental marketing activity, and a Cluster Analysis is a convenient method for identifying homogeneous groups (Mooi & Sarstedt, 2011).
Results
In table 3, the mean utilities are displayed for the different levels of the four attributes. These utilities are estimated at the individual respondent level using HB analysis. To test whether the different levels used for price, quality, brand and transparency are perceived as expected, statistical analyses are done in SPSS. Attribute quality
consisted of two levels, so a Paired Samples T-test is conducted. Prior to conducting the analysis, the assumption of normality is examined. The assumption is considered satisfied, as the skewness and kurtosis levels for three stars quality are estimated at -.83 and .80, respectively, and skewness and kurtosis levels for four stars quality are estimated at .83 and .80, respectively. It can be assumed that this variable is normally distributed. Furthermore, there are no significant outliers. Subsequently, Paired Samples T-test is conducted, and showed a significant difference between three stars quality and four stars quality perceived by previous consumers, t(194) = -9.80, p < .001. Hence, the mean utility for four stars quality (M= .41, SD= .59) was
significantly higher than the mean utility for three stars quality (M= -.41, SD= .59). For the attributes with three levels, a Repeated Measures Anova is conducted. Before conducting this analysis, several assumptions are checked. Skewness and kurtosis levels for price level 1 (i.e., € 1.49) are estimated at .54 and -.21,
respectively. Skewness and kurtosis levels for price level 2 (i.e., € 1.99) are estimated at .16 and .39, respectively. Finally, skewness and kurtosis levels for price level 3 (i.e., € 2.49) are estimated at -.36 and -.67, respectively. Therefore, the normality assumption is satisfied. In addition, there are no significant outliers. The final assumption that is checked is known as sphericity. Mauchly's Test of Sphericity indicated that the assumption of sphericity has been violated, χ2(2) = 309.40, p < .001. Due to this violation, degrees of freedom are corrected using Greenhouse-Geisser estimates of sphericity (ε = .56). Greenhouse-Geisser correction is used, because ε <.75 (Girden, 1992). There are significant differences between the levels of price,
F(1.11, 215.71) = 311.49, p < .001. The mean utility for price level 1 (M= 1.44, SD=
1.17) was significantly higher than the mean utility for price level 2 (M= .17, SD= .32) and was also significantly higher than the mean utility for price level 3 (M= -1.60, SD= 1.20). Finally, the mean utility of price level 2 (M= .17, SD= .32) was significantly higher than the mean utility of price level 3 (M= -1.60, SD= .32).
Next, a Repeated Measures Anova is used to compare the levels for the brand attribute. Again, several assumptions are checked. The assumption of normally distributed scores is satisfied for Tony Chocolonely and Cote d’Or. Skewness and kurtosis levels for Tony Chocolonely are estimated at .35 and -.47, respectively, and skewness and kurtosis levels for Cote d’Or are estimated at .37 and .95, respectively. Albert Heijn is not normally distributed. Kurtosis level was above 1, and appeared to be caused by outliers. The outliers are examined to assure no data entry or instrument errors have been made. A normality test of Albert Heijn without these outliers showed a normal distribution, with skewness and kurtosis levels of -.35 and -.78, respectively. Mauchly's Test of Sphericity indicated that the assumption of sphericity has been violated, χ2(2) = 23.34, p < .001. Therefore, degrees of freedom are corrected using Huynh-Feldt estimates of sphericity (ε = .90). Huynh-Feldt correction is used, because ε >.75 (Girden, 1992). There are significant differences between the brands,
F(1.81, 351.37) = 95.27, p < .001. The mean utility for Tony Chocolonely (M= 1.53, SD= 1.91) was significantly higher than the mean utility for Cote d’Or (M= -.27, SD=
1.49) and was also significantly higher than the mean utility for Albert Heijn’s chocolate bar (M= -1.26, SD= 1.53). Finally, the mean utility of Cote d’Or (M= -.27,
SD= 1.49) was significantly higher than the mean utility of Albert Heijn (M= -1.26, SD= 1.53).
The levels of transparency are examined using a Repeated Measures Anova. Website transparency, Packaging transparency and opaqueness have normal
distributions. Skewness and kurtosis levels for Website Transparent are estimated at .16 and .66, respectively. Skewness and kurtosis levels for Packaging Transparent are estimated at .61 and -.30, respectively, and skewness and kurtosis levels for Opaque are estimated at -.28 and -.72, respectively. However, Translucent is not normally distributed because the kurtosis level was above 1. It appeared to be caused by an outlier. This outlier was examined and due to the relatively large sample size in this study (N = 195), the absence of a normal distribution is unlikely to influence further statistical analyses (Elliot, 2007). Mauchly's Test of Sphericity indicated that the assumption of sphericity has been violated, χ2(5) = 490.02, p < .001. Therefore, degrees of freedom are corrected using Greenhouse Geisser estimates of sphericity (ε = .43). There are significant differences between the levels of transparency, F(1.30, 215.71) = 255.48, p < .001. The mean utility for Packaging Transparent (M= 2.00,
SD= .61) and was also significantly higher than the mean utility for Opaque (M=
-1.47, SD= 1.33). Finally, the mean utility of Translucent (M= -.70, SD= .61) was significantly higher than the mean utility of Opaque (M= -1.47, SD= 1.33). In addition, the mean utility of Packaging Transparent (M= 2.00, SD= 1.64) was
significantly higher than the mean utility of Website Transparent (M= 1.68, SD= .50). Thus, there is empirical evidence found for the first and second hypothesis.
Prior to testing the third hypothesis, importances of the attributes are
calculated. The relative importances per attribute are shown in table 4. Furthermore, the descriptive statistics of the relative importance variables are displayed in table 5. To test if the means significantly differ from each other, a Repeated Measures Anova is conducted. The variables are normally distributed. However, the Mauchly’s Test of Sphericity indicated that the assumption of sphericity has been violated, χ2(5) = 157.81, p < .001. Degrees of freedom are corrected using Huynh-Feldt estimates of sphericity (ε = .76). There were significant differences between the means, F(2.29, 444.76) = 55.05, p < .001. The relative importance for transparency (M= .32, SD= .20) is significantly higher than the relative importance for quality (M=.10, SD= .09). However, there is no significant difference between the relative importance for transparency and price. Thus, the third hypothesis is not supported. However, this analysis shows another noteworthy, not hypothesized result. It appears that the relative importance of brands (M= .33, SD= .21) is significantly higher than the relative importance for quality (M= .10, SD= .09) and price (M= .24, SD= .17). Hence, both transparency and brand are relatively important.
Furthermore, it is examined if there is a negative correlation between the relative importance of the transparency on one hand and brands on the other. Scatter plot showed no linear relationship between those variables, so it was not possible to use a Bivariate Pearson Correlation to test this fourth hypothesis. Therefore, a Spearman’s correlation coefficient is applied. There is a monotonic relationship between the two variables, so the assumption of monotonicity is considered satisfied. Results showed that there was a negative relationship between relative importance of transparency and brands, which was statistically significant, rs = -.48, p < .001. Thus, when the relative importance of transparency is increasing, the importance of brands is decreasing. Moreover, when the relative importance of brands is increasing, the relative importance of transparency is decreasing. As a result, the fourth hypothesis is supported.
Table 3
Descriptive statistics utility price, quality, brand, and transparency
Attributes N Min Max M SD
Price Low Medium High 195 -.81 -.69 -4.20 4.34 1.05 -1.49 1.43 .17 -1.60 1.17 .32 1.20 Quality Low High 195 -2.23 -.84 .84 2.23 -.41 .41 .59 .59 Brand Tony Chocolonely Cote d’Or Albert Heijn 195 -2.55 -3.94 -4.44 5.82 4.10 4.41 1.53 -.27 -1.26 1.91 1.49 1.53 Transparency 195 Opaque Translucent Transparent (website) Transparent (packaging) -4.20 -2.06 -1.10 -.93 1.92 2.06 2.01 6.08 -1.47 -.70 .17 2.00 1.33 .61 .50 1.64 !
Table 4
Relative importance of product attributes
Attribute Relative importance
Transparency 32 %
Price 24%
Quality 10%
Brand 33%
Table 5
Descriptive statistics relative importance of transparency, price, quality, and brand
Relative importance N Min Max M SD Transparency 195 .05 .89 .32 .20 Price 195 .01 .74 .24 .17 Quality 195 .00 .44 .10 .09 Brand 195 .03 .92 .33 .21
Next, a Cluster Analysis is used to determine homogenous groups. The clustering variables are the relative importance of transparency and the relative importance of brands. The number of clusters is known, specifically the fifth hypothesis proposes that there are two distinct groups (i.e., one that considers transparency as relatively more important than brands and one that considers brands as relatively more important than transparency.). A Hierarchical Cluster analysis and K-Means Cluster analysis are conducted to check if these groups exist and if these groups significantly differ from each other.
With regard to the Hierarchical Cluster analysis, a between-groups linkage is chosen to find differences between the clusters. Furthermore, Euclidean distance is selected, since it is the most commonly used type of measure when it comes to analyzing ratio or interval-scaled data (Mooi & Sarstedt, 2011). Box’s Test of Equality of Covariance Matrices was significant, so the assumption of equality of covariance matrices was violated, F(3, 523486.89) = 19.31, p < .001. For this reason, Mixed Design Anova cannot be used. As to the relative importance for brands, there
was Homogeneity of Variance as assessed by Levene's Test for Equality of Variances,
F(1, 193) = .51, p = .48. For the relative importance of transparency, Homogeneity of
Variances has been violated, F(1, 193) = 43.74, p <.001. However, cluster sizes are relatively equal (i.e., not four times greater) and the difference in variance of cluster 1 and 2 is no more than ten (Tabachnick & Fidell, 2013). One-way Anovas are used, although this increases the chance of Type 1 error. The One-way Anovas showed that the clusters significantly differed for the relative importance of transparency, F(1, 193) = 77.21, p < .001, and for the relative importance of brands, F(1, 193) = 453.42,
p < .001. The relative importance of transparency is significantly higher in cluster 1
(M= .40, SD= .21) than in cluster 2 (M= .17, SD= .08). Moreover, the relative importance of brands is significantly higher in cluster 2 (M= 57, SD= .13) than in cluster 1 (M= .20, SD= .11). The descriptive statistics of the clusters drawn from the Hierarchical Cluster analysis are displayed in table 6.
K-Means Cluster analysis also showed that the two clusters significantly differ for the relative importance of transparency, F(1, 193) = 179,28, p < .001, and for the relative importance of brands, F(1, 193) = 213.21, p < .001. The descriptive statistics of the clusters are displayed in table 7. Finally, to ensure that clusters from the
analyses overlap, a crosstab is computed. This crosstab showed that cluster 2 (N = 68) of the Hierarchical Cluster analysis in its entirety is represented in cluster 1 (N = 100) of K-Means Cluster analysis. Additionally, cluster 2 (N = 95) from K-Means Cluster analysis is represented in cluster 1 (N = 127) from Hierarchical Cluster analysis. There is an 84% overlap between the clusters of these two types of cluster analysis. Altogether, the fifth hypothesis is supported.
Table 6
Descriptive statistics clusters from Hierarchical Cluster analysis Cluster N Min Max M SD Relative importance of 1 127 .05 .89 .40 .21 transparency 2 68 .05 .38 .17 .08 Total 195 .05 .89 .32 .20 Relative importance of 1 127 .03 .48 .20 .11 brands 2 68 .34 .92 .57 .13 Total 195 .03 .92 .33 .21
Table 7
Descriptive statistics clusters from K-Means Cluster analysis Cluster N Min Max M SD Relative importance of 1 100 .05 .45 .18 .09 transparency 2 95 .07 .89 .46 .19 Total 195 .05 .89 .32 .20 Relative importance of 1 100 .09 .92 .48 .18 brands 2 95 .03 .41 .17 .10 Total 195 .03 .92 .33 .21
Discussion
This final part elaborates on the results of this research. A general conclusion is provided, and the theoretical and practical implications of these results are presented. Finally, limitations of this research and suggestions for future research are discussed.
General Conclusion
Nowadays companies have to deal with an increased demand for transparency among consumers (Cohn & Wolfe, 2013). Due to the emergence of the Internet, companies are under more scrutiny than before, and as a consequence, companies may feel more pressure to operate transparently (Fournier & Avery, 2011). However, most research indicates that transparency can have positive effects for companies. For example, transparency may result in an increase in purchase intention (Kang & Hustvedt, 2014; Bhaduri & Ha-Brookshire, 2011), customer satisfaction (Simintiras et al., 2015; Matzler et al., 2006), and consumer loyalty (Singh, 2015). Now that more companies are becoming transparent, a question that arises is: to what extent transparency is relevant in day-to-day choices among consumers. Cohn and Wolfe (2013) show that transparency and honesty are important in consumer decision-making, even more important than brands. However, these results are obtained through a survey. The data collected through a survey is unlikely to be as detailed as those collected by other research methods (Lewis & Saunders, 2012), such as a CBC. This trade-off analysis was used to examine the relative weight of transparency against other product attributes such as brand. Although Cohn and Wolfe (2013) argue that the importance of brands is decreasing due to an increase in transparency, previous research has not provided sufficient empirical proof to demonstrate that transparency is becoming more important in consumer decision-making, and therefore, the relative importance of brands is decreasing. To close this research gap, the following research question was proposed: How does transparency influence consumers’ choice for brands?
The purpose of this study was to expand the transparency literature and add to the theory of transparency on consumer behavior. This research contributes to this literature through three main findings.
Firstly, results showed that consumers prefer transparent companies to translucent and opaque companies. The utilities for opaque and translucent way of doing business were negative, and the utilities for the two transparent ways of doing business were positive. These findings correspond with Cohn and Wolfe’s (2013) findings that consumers take transparency into consideration when purchasing a product. It is also consistent with Carter and Curry’s (2010) findings, which states that consumers generally value transparent prices more than opaque prices. However, this only was the case when the allocation was perceived as fair. Moreover, this research indicates that transparency through information disclosed on the packaging is preferred to transparency through information disclosed on the homepage of the website.
Secondly, it was hypothesized that the product’s price and quality are more important than the company’s transparency. This hypothesis is not supported, because transparency is found to be significantly more important than quality. Consumers thus value companies communicating openly more than the product’s quality perceived by previous consumers. No significant difference was found between relative importance of transparency and the relative importance of price. These findings do not correspond with Cohn and Wolfe’s (2013) findings, as they found that price and quality are more important than transparency. However, Bhaduri and Ha-Brookshire (2011) found that consumers who perceived a higher value in purchasing a transparent product were willing to compromise on quality. Moreover, when confronted with risk, consumers often rely on a quality signal and are more motivated to select a brand as a form of insurance (Richardson et al., 1996). The provided information reduces the importance of mentioning that previous consumers perceived the quality. As consumers possess more information, consumers are less in need for a quality signals (Delveccio, 2001). Furthermore, results showed that the three most important product attributes were brand, transparency, and price. This is also not consistent with the findings of Cohn and Wolfe (2013), which state that the three most important factors of purchase intentions are price, quality and transparency. Based upon the results of this research, it seems that the relative importance for transparency and brands in consumer
decision-making are approximately equal. This additional result is discussed in the next paragraph.
Thirdly, the results showed that the relative importance for transparency and brands significantly correlate. There was a negative relationship, which means that an
increase in relative importance of transparency results in a decrease in relative importance of brands and vice versa. As a consequence, it was examined if two different groups exist, that is, consumers who consider transparency as relatively more important than brands and consumers who consider brands as relatively more important than transparency. Results showed that these distinct consumer segments exist in the age of transparency. These findings provide further support for the findings of Cohn and Wolfe (2013), which state that if the relative importance of transparency is increasing, the relative importance of brands is decreasing. Brands contribute to reducing the consumer’s risk of making a purchase mistake (Fischer, Völckner & Sattler, 2010). However, consumers are less in the need for a quality guarantee, such as brand names, in the age of transparency. Information reduces the perceived risk of purchase. Therefore, brands are becoming less important while transparency is becoming more important. Nevertheless, brands still function as social demonstrance (Fischer, Völckner & Sattler, 2010). For instance, brands are used to project self-image (Levy, 1959), to keep and enhance self-image (Shrauger, 1975), to communicate to other the type of person they are or would like to be (Escalas & Bettman, 2005. When brands are used as social demonstrance, transparency becomes less important. It might also be the case that some consumers are extremely loyal to the brand. These consumers think a certain brand continues to offer the best choice alternative, and they naively avoid communications from competitors that argue their product is most efficient, lowest price, and so forth (Oliver, 1999).
Theoretical and Practical Implications
The findings of this research have numerous theoretical implications. So far most studies have compared an opaque and transparent company. However, this study contributes to the existing literature to be one of the first in investigating the relative importance of transparency against price, quality, and brands. Close attention has been devoted to the brands in the age of transparency. To examine the relative
importance of transparency against the other attributes, a CBC analysis has been used. CBC is a popular marketing tool that is able to measure the utilities of different product attributes. After the CBC, data was tested with statistical tests and relative importances of the product attributes were calculated. The findings, an increase in the relative importance of transparency results in a decrease in the relative importance of
brands and vice versa and the existence of two consumer segments, builds upon the findings of Cohn and Wolfe (2013), which state that brands are becoming less important in the age of transparency. Moreover, these findings build on the works of Fischer, Völckner and Sattler (2010), which propose that brands function as risk reduction and as social demonstrance. As consumers use brands for social demonstrance purposes, transparency is becoming less important.
Furthermore, after conducting a CBC, a cluster analysis was performed. These results showed that there are two different groups in the age of transparency. On the one hand, a consumer segment that considers transparency as relatively more
important than brands. On the other hand, a consumer segment that considers brands as relatively more important than transparency. This study is the first to demonstrate these consumer segments based upon CBC data.
This study also shows that transparency is more important in consumer decision-making than quality. These findings build on the findings of Bhaduri and Ha-Brookshire (2011), which state that consumers who perceived a higher value in purchasing a transparent product were willing to compromise on quality. These findings also support Delvecchio’s (2001) findings, which propose that as consumers possess more information, consumers are less in need of quality signals. Reviews by previous consumers can be seen as quality signals. Consumers often rely on these quality signals, when confronted with risk of purchasing, but when more information becomes available, the provided information is used to mitigate this risk.
In addition, literature is expanded because this study is one of the first in testing and comparing different levels of high transparency. Based upon the mean utilities, transparency through information disclosed on the packaging was preferred to transparency through information disclosed on the homepage of the company’s website. These findings build upon the finding of McNeill and Wyeth (2011), which state that packaging is a driving force in product choices. It also provides support for Behaeghel (1991) and Peters’ (1994) consideration that packaging is the most
important communication tool.
The findings have practical implications for marketers. Existing literature has mainly focused on comparing transparent products with opaque products. Considering it is expected that transparency is only increasing, the results argue that brands are becoming less important for consumers decisions. Marketers should notice that it is easier for consumers to switch to alternatives. However, brands are still important in
the daily lives of consumers, because brands are used for social demonstrance. Hence, marketers should position a brand more as a symbolic device and less as a quality guarantee.
Moreover, with the expectation that transparency is only increasing, the results also provide some insights in which communication tool is preferred to use for
communicating information about business practices. Based on this research, marketers are advised to choose disclosing information on the product’s packaging instead of disclosing information on the website’s homepage.
Limitations and Future Research
This research has some limitations. Fortunately, limitations develop new opportunities for future research, especially in the still growing field of transparency.
First of all, Omre (2010) states that when calculating importances from CBC data, it is recommended to use part-worth utilities resulting from HB estimation. However, one of the concerns with standard importance analysis is that is considers the extremes within an attribute, not taking into account whether the part-worth utilities follow rational preference order. The importance calculations capitalize on random error, and attributes with little to no importance can be biased upward in importance. Therefore, it is preferred to use sensitivity analysis.
On the one hand, a biased increase in importance could be the case for the brand attribute. For the brand attribute a private label was selected. It could be that respondents were less familiar with Albert Heijn’s products because they shop at another retail store or supermarket. Therefore, the range of the brand attribute could be wider. On the other hand, the quality attribute only contained two levels. This could also result in a biased importance calculation, because the range could be narrower. Future research could allow for this limitation of the standard importance analysis and use a sensitivity analysis to calculate importances of CBC data.
Furthermore, utilities are calculated for one type of product (i.e., chocolate bars). However, consumer choices are likely to deviate from this research, since consumers can choose from a variety of products in the real world. Additionally, a chocolate bar is a Fast Moving Consumer Good (FMCG). Consumers may be less motivated to read information disclosed on the website or packaging for a FMCG
decision. Therefore, future research could include a durable good to compare the relative importances of attributes for these two types of products.
This research included two different levels of high transparency. However, a company can be transparent in multiple ways, for instance through the amount of information a company provides. Therefore, future research could include other high transparency levels, but perhaps more interestingly, future research could further investigate why utility for transparency through information disclosed on the
packaging was higher than the utility for transparency through information disclosed on the website’s homepage. Finally, future research could focus on the consumer segments that were uncovered in this research. So far no previous research has showed that the existence of these two distinct groups in the age of transparency. Questions remain what other characteristics of these consumer segments are.
References
Aaker, D. A. (1991). Managing Brand Equity. New York: The Free Press.
Aaker, D. A. (1996). Measuring Brand Equity across Products and Markets.
California Management Review, 38(3), 102-120.
Allenby, G. M., Rossi, P. E., & McCulloch, R. E. (2005). Hierarchical Bayes Models: A Practitioners Guide. Journal of Bayesian Applications in Marketing. Retrieved from:
https://www.researchgate.net/profile/Greg_Allenby/publication/228150862_Hierarchi
cal_Bayes_Models_A_Practitioners_Guide/links/5436b48d0cf2bf1f1f2d0820.pdf
Alves, S., Aspinall, P. A., Thompson, C. W., Sugiyama, T., Brice, R., & Vickers, A. (2008). Preferences of Older People for Environmental Attributes of Local Parks: The Use of Choice-Based Conjoint Analysis. Facilities, 26(11/12), 433-453.
Aspinall, P. A., Hill, A.R., Nelson, P., O’Brien, C., O’Connell, E. O., McCloughan, L., Azuara-Blanco, A., Brice, R., Green, S., & Steeds, C. (2005). Quality of Life in Patients with Glaucoma: A Conjoint Analysis Approach. Visual Impairment
Research, 7, 13-26.
Behaeghel, J. (1991). Brand Packaging: The Permanent Medium. Architecture Design and Technology Press: London.
Bertini, M., & Gourville, J. T. (2012). Pricing to Create Shared Value. Harvard
Business Review, 90(6), 96-104.
Bhaduri, G., & Ha-Brookshire, J. E. (2011). Do Transparent Business Practices Pay? Exploration of Transparency and Consumer Purchase Intention. Clothing & Textiles
Bhaduri, G., & Ha-Brookshire, J. E. (2015). The Role of Brand Schemas, Information Transparency, and Source of Message on Apparel Brands’ Social Responsibility Communication. Journal of Marketing Communications, 1-18.
Bishop, W. R. (1984). Competitive Intelligence. Progressive Grocer, 63, 19-20.
Boniface, B., Gyau, A., & Stringer, R. (2012). Linking Price Satisfaction and Business Performance in Malaysia’s Dairy Industry. Asia Pacific Journal of
Marketing and Logistics, 24(2), 288-304.
Carter, R. E., & Curry, D. J. (2010). Transparent Pricing: theory, tests, and
implications for marketing practice. Journal of Academic Marketing Science, 38, 759-774.
Carter, C. R., & Rogers, D. S. (2008). A Framework of Sustainable Supply Chain Management: Moving Toward New Theory. International Journal of Physical
Distribution & Logistics Management, 38(5), 360-387.
Christensen, L. T. (2002). Corporate Communication: The Challenge of
Transparency. Corporate Communications: An International Journal, 7(3), 162-168.
Cobb-Walgren, C. J., Ruble, C. A., & Donthu, N. (1995). Brand Equity, Brand Preference, and Purchase Intent. Journal of Advertising, 24(3), 25-40.
Cohn & Wolfe. (2013). From transparency to full disclosure. London.
Connolly, A., & Davidson, L. (1996). How Does Design Affect Decisions at Point of Sale? Journal of Brand Management, 4(2), 100-7.
Degeratu, A. M., Rangaswamy, A., & Wu, J. (2000). Consumer Choice Behavior in Online and Traditional Supermarkets: The Effects of Brand Name, Price, and Other Search Attributes. International Journal of Research in Marketing, 17(1), 55-78.
