Localizing or globalizing product packaging.
An experiment looking at local vs. global packaging strategies and the role of product type in the Dutch market.
Paul Ivo Schötteldreier 10655530
Submission date: 22-06-2018
Master Thesis MSc Business Administration: Marketing (track) University of Amsterdam
1. Statement of originality
Statement of originality
This document is written by Student Paul Ivo Schotteldreier 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
This thesis is dedicated to my grand-mother who was very dear to me and always believed in me. Unfortunately she was battling illness and passed away during the writing of this thesis.
Inge Heider geb. Röder 15-08-1946 - 05-05-2018
2. Table of contents
1. Statement of originality 1
2. Table of contents 3
3. List of tables and figures 4
4. Abstract 5
5. Introduction 6
6. Theoretical Framework 9
6.1 Social Identity Theory 9
6.2 Brand positioning strategies 9
6.3 Packaging 10
6.4 Localization/globalization of packaging 11
6.5 Packaging language 11
6.6 Hedonic vs. utilitarian products 12
6.7 Brand likeability 12
6.8 Packaging likeability 13
6.9 Willingness to pay 14
6.10 Conceptual model and hypotheses 16
7. Methodology 18 7.1 Research method 18 7.2 Sampling method 18 7.3 Respondents 18 7.4 Pre-test 19 7.5 Product selection 20
7.6 Hedonic and utilitarian measurements 20
7.7 Trustworthiness of the manipulated product 21
8.4 Control variables 27
8.5 Main and interaction effects 30
9. Conclusion 36 9.1 Brand likeability 36 9.2 Packaging likeability 37 9.3 Willingness to pay 38 10. Discussion 39 10.1 Limitations 39 10.2 Implications 39 10.3 Future research 40 11. References 41 12. Appendix 50
12.1 Appendix A: Stimulus materials 50
12.2 Appendix B: Complete survey 54
12.3 Appendix C: Pre-test stimulus 58
3. List of tables and figures
1. Figure 1: Conceptual model 16
2. Table 1: Overview hypotheses 17
3. Table 2: Distribution of participants over the four conditions 19 4. Table 3: Utilitarian and hedonic levels of selected product groups 21 5. Table 4: Main and interaction effects on brand likeability 31 6. Table 5: The average scores on brand likeability 31 7. Table 6: Main and interaction effects on packaging likeability 32 8. Table 7: The average scores on packaging likeability 33 9. Table 8: Main and interaction effects on willingness to pay 34 10. Table 9: The average scores on willingness to pay 34
4. Abstract
Should products and their brand positioning strategies be tailored to each market the product is sold in, or could a global standardised approach be more effective? This question has kept both marketers and scientists debating. However, the question of variable packaging seems to be rather untouched in this field of research. Since packaging is one of the main outlets of brand positioning strategies in the FMCG business, conducting a study that looks at the effectiveness of either a globalised or localised packaging strategy could be extremely valuable. The studies that have been conducted so far are executed in non-western countries with different alphabets. Thus, a global (English) packaging strategy implies a significant change in the packaging itself. To broaden the scientific knowledge, conducting a study in a western non-English speaking country is necessary.
This study included another independent variable, which was product type. This variable has been proven to be influential in studying the differences in local and global packaging strategies and was therefore part of this study. The study comprised of a 2 (local/global positioning) x 2 (hedonic/utilitarian product) between subjects 2- factorial design. After analysing the results (N=213), no significant differences between the hedonic and utilitarian products within the global and local product packaging strategies were found. The dependent variables for which these effects have been studied were brand likeability, packaging likeability and willingness to pay. Unfortunately this study could not contribute significant results to the paradigm but it did provide deeper understanding of the matter and paved the way for future studies.
Keywords – Local, Global, Brand Positioning, LCCP, GCCP, Packaging. Product type, Hedonic, Utilitarian, Language
5. Introduction
“Language discrimination by Albert Heijn.” This is the headline of several articles from 2011 mentioning the uproar started by the Dutch Language Defence Fund complaining about the standardised packaging many of private brands of the biggest supermarket chain in the Netherlands, Albert Heijn (Volkskrant, 2011). The fund cites two arguments as to why this is a poor development: that people cannot read what is inside the packaging anymore and that the Dutch language is under attack. In particular translating local products into English seems like a wrong move for some customers. As a response, Albert Heijn decided that the local products would undergo a change of packaging and would be tailored to the Dutch market again. This is a challenge for many companies that expand outside their country of origin. Should they enter the market with a standardised product, or enter the country with a localised product?
The issue of globalizing or localizing branding approaches is a topic that leads to discord outside the business environment. There are many studies that look into the
differences in both the local and the global branding approaches (Fan, 2002; Okazaki, 2004; Eckhardt, 2005; Fastoso and Whitelock, 2012; Jiang and Wei, 2012). However, these studies portray a lot of dissension about this issue. On many variables, researchers cannot seem to agree with one another on what variables are actually most effective when choosing either a globalized or localized branding approach. A variable that is rarely touched upon is product packaging. Especially in the FMCG industry, packaging is an extensive part of branding (Speece & Silayoi, 2007). The studies that are done regarding localizing vs. globalizing product packaging are mostly conducted in non-western countries (Khan, Lockshin & Corsi, 2017; Kahn & Lockshin, 2017; Lockshin, Khan & Lee, 2015; Hu, Chen & Yang, 2016). These studies show concurring results on the use of either localized or globalized packaging strategies. However, most of these studies are conducted in countries that have a different alphabet than that of most western countries. Therefore creating a more significant change in both packaging approaches. Together with the significant cultural differences between the non-western countries and the western countries of origin of the products, it can be concluded that these factors are paving the way for greater chances of finding significant results.
Nonetheless, the study of Katz and Frost (1992) shows that orthographical difference is not subordinate to alphabet difference. Doing a study that looks into the effects of both packaging approaches in a non-English speaking western country has not been conducted but could enhance the theoretical base for this issue.
Studying the effect of either using a globalized or a localized packaging approach in a western non-English speaking country has some benefits for the business side as well. Figures show that in the FMCG industry an increasing proportion of the budget is spent on packaging due to the recovery of the global economy (EY, 2016). Adapting all packaging to each market a company operates in would mean higher operating costs. Hence, trying to push to a global packaging approach would be more cost efficient. But the downsides seem to be unknown. There is little empirical data that looks deeper into this possible trade off between cost and attitude in non-English western countries.
Another variable that is proven to be influential when studying the difference between localised and standardised packaging in both western and non-western countries is the
difference between the product types. First, it is necessary to understand what is meant by this variable. Products can be categorized on many variables based on the function they fulfil or the form in which them come. Search vs. experience goods for example or food vs. non-food. In this study the product types that are taken into account can be defined as the product either being hedonic or utilitarian. These two terms distinguish products not based on the form of the product itself but by the perception of the consumer of that product related to the function that this product is fulfilling (Batra & Athola, 1991). As an explanation for why this form of product type distinguishing is taken into account in a study concerning global vs. local packaging, the study of Leclerc, Schmitt and Dubé (1994) states that language has a significant role in the evaluation of both product types. Since the hedonism factor of a particular product is closely associated with the foreignness of that product and since utilitarian products thrive on understandable communication, having either a hedonic or utilitarian product may play an important role in evaluating the language on the packaging of a product. These findings have been confirmed by the study Lockshin et al. (2015), which shows that for hedonic products, foreign brand names are evaluated more favourably opposing to utilitarian products.
dependent variable. The study of Li, Li and Kambele (2012) shows that having a higher brand evaluation leads to customers who are more willing to pay a higher price. Since studies have proven that choosing a certain packaging strategy will lead to different brand and product evaluations, it would be reasonable to assume that these findings could potentially be translated into a higher willingness to pay. This has not been studied yet and therefore the following RQ will be proposed:
RQ: What is the effect of either choosing a globalised (English) or localised (Dutch)
packaging strategy for both hedonic and utilitarian products on brand likeability, packaging likeability and willingness to pay?
Although there are some studies already done in this field, a major contribution of this study would be the fact that it is done in a non-English speaking western country. Previous studies have mostly been conducted in non-western countries (Khan et al., 2017; Kahn & Lockshin, 2017; Lockshin et al., 2015; Hu et al., 2016). However, the Dutch language has the same alphabet as the English language, meaning that the execution of localizing packaging has different consequences for a western country with an ISO basic Latin alphabet than for a non-western country with a different alphabet. This study could fill the gaps and broaden our understanding of the localization of product packaging.
This study could also be very valuable to marketers seeking the most efficient way to enter a new market with their product. A quick walk through a local supermarket shows that there is a lot of inconsistency between the languages used on the packaging. This study could determine what strategy would be most effective and could lead to a consensus of what strategy to choose.
6. Theoretical framework
The main theoretical concepts about the Social Identity Theory, brand positioning strategies, product types, brand and packaging likeability and willingness to pay will be discussed and explained in the following section.
6.1 Social identity theory
The phenomenon of localising products has been around since the beginning of globalisation, and marketers and researchers have been diving into the matter ever since. Why is there a difference and what works better? The foundation and occasion of the studies that followed the rise of globalisation is laid by the study of Turner, Brown and Tajfel (1979), which shows that people tend to favour things that are relatable to their own social group. These findings are backed up by the Social Identity Theory (Turner, 1975), which states that people see themselves as in-group members. Their self-identity becomes one with their social identity as a group member. Hence, customers favour things that are in line with their social identity as a group member. This conclusion was interfering with globalisation because it showed that a globalised strategy is not working for everyone. This has been proven to be the case for many characteristics of this social identity. Poon, Evangelista and Albaum (2010) showed that using advertising that made use of characters with the same ethnic background as the customer evoked more favourable attitudes to the advertised product. However, these relationships are a bit more complex, and differ according to the formats and levels in which these culturally defined relationships are portrayed.
6.2 Brand positioning strategies
When going for a local positioning, the brand is tailored to the market in which the product is sold. This can happen on several levels, language for example. But also cultural traditions, race and religion are cultural factors that influence the brand’s position of a product (Alden et al., 1999). An example would be the orange M&M version that was made for the Dutch market (Distrifoods, 2014). An example executed on a brand name level would be Unilever’s ice cream brand Ola, which has different brand names in each country it is operating
(Advergize, 2014). GCCP
A global positioning on the other hand would entail a standardised positioning that is tied to one global strategy. This strategy is often used by tech companies for example who do not bind themselves to geographic locations but more to the technology they produce. The brand name, product and advertising strategy for Apple products, for example, are the same in each country in which Apple is operating (BrandQuarterly, 2015).
FCCP
A third strategy that could be executed is using a foreign positioning. This positioning strategy embraces cultural aspects of a specific country and incorporates them into the brand itself. A case example would be Singapore Airlines. The company embraces their culture of origin and doesn’t adapt when entering in to other markets (their Asian looking brand ambassador/mascot is used in all operating countries) (Alden et al., 1999).
6.3 Packaging
These three brand-positioning strategies can be expressed in different formats (Simonson & Schmitt, 1997). Advertisements are the most studied (Akaka & Alden, 2010; Cui, Yang, Wang, & Liu, 2012; Nijssen & Douglas, 2011; Okazaki, Mueller, & Taylor, 2010; Zhou, Poon, & Wang, 2015) and notable format but there are more. A less studied and significantly different format is packaging. This format is less dimensional and has a lower emotional connectedness to customers (Griffith, Chandra and Ryans, 2003). This does not make it less of an important format. Packaging plays a key role in influencing consumer choice decisions (Underwood and Klein, 2002; Ampuero and Vila, 2006; Ballantyne et al., 2006; Clement,
2007; Mueller, Lockshin & Louviere, 2010; Steenkamp, Van Heerde & Geyskens, 2010; Venter, Merwe, Van der Beer, Kempen & Bosman, 2011) and is the last point of contact before purchasing, making it a key format for communicating qualities and benefits of products (Ampuero and Vila, 2006; Clement, 2007). Packaging can be defined as a container that holds and protects the product, encompasses attributes such as shape, design and symbols and facilitates handling and commercialisation (Underwood, 2003; Ampuero and Vila, 2006; Aday and Yener, 2014). Since packaging is also part of the formats in which the three
elements (LCCP, GCCP and FCCP) proposed by Alden et al. (1999) can be expressed, analysing which can be perceived as most effective in what context can be valuable. 6.4 Localization/globalization of packaging
Packaging and other branding formats have many variables in which they can differ.
According to Alden et al. (1999), these variables can be categorized in to three components, which are language, aesthetic styles and story themes. Cues like symbols, colours, logos and labels, which are part of the components’ aesthetic styles and story themes, have been subject to an array of studies with regard to packaging (Underwood and Klein, 2002; Clement, 2007; Clement, Kristensen & Grønhaug, 2013; Littel and Orth, 2013). However, one of the main components is language, which has only been looked at by a limited number of studies in combination with localisation and globalisation (Khan et al., 2017; Kahn & Lockshin, 2017; Lockshin et al., 2015; Hu et al., 2016; Leclerc et al., 1994). The study of Leclerc et al. (1994) is the only study conducted in a western country (France). However, the study only focuses on the translation of the brand name. This study aims to localize other text elements on the product packaging so that literature gaps can be tapped.
Hence, the descriptive text stating the type of product and the slogans will be manipulated to create a global and a local version of the product. Studying language on the packaging on the descriptive level in a non-English speaking western country has not been done yet. In this study globalisation of packaging language is defined as a packaging written in English. The language English is chosen due to it scoring the highest on globalness compared to all languages (Crystal, 2012).
6.6 Hedonic vs. utilitarian products
Another variable that is proven to be influential when studying the difference between localised and standardised packaging in non-western countries, is the difference between the product in it either being hedonic or utilitarian. First, it is important to clearly define the theory and definition of these concepts. The foundation for these terms can be traced back to early philosophy and psychology in which hedonism is defined as a motivation that people can carry which allows them to move away from pains and maximize pleasure (Riediger, Schmiedek, Wagner & Lindenberger, 2009). The term utilitarianism often refers to the psychological definition of a motivation, that allows people to make choices based on the outcome which has to be worth more than the suffering invested into making the choice (Mill, 2016). Within the field of marketing, both words have synonyms like practical or enjoyable and are mostly applied to consumer behaviours and products (Hirschman & Holbrook, 1982; Ahtola, 1985). Hence, in this study hedonic products are defined as products that fulfil a need for personal expression, convey status and are related to sensory pleasure (look, taste or smell nice). Therefore, these products are more pleasure and experiential based (Park, Jaworski, & MacInnis, 1986). The term utilitarian product is defined as a product that fulfils a need for problem solving or problem prevention and therefore is of a practical or useful nature (Park et al., 1986).
6.7 Brand likeability
Some studies apply this variable to the localization of brand packaging with result. The studies of Khan et al. (2017) and Khan et al. (2015) show that the brand evaluation differs for both product types when using either a local or global packaging strategy. As an explanation, both studies mention the fact that the value of the packaging of a product differs between either a hedonic or utilitarian product. For hedonic products, attention grabbing and prominent
packaging is important and for utilitarian products informative and relatable packaging is important. These studies link these findings to language by stating that a foreign language is enhancing the perceived excitement and that a local language is enhancing readability. This conclusion is backed up by studies that found out that for hedonic products, writing the packaging text in the foreign language is increasing the brand evaluation more than for utilitarian products (Khan et al., 2017; LeClerc et al., 1994; Khan et al., 2015). These studies elucidate these differences by stating that it is a fact that when consumers seek for fun in their consumption, a foreign language may enhance the perception of hedonism and therefore increase the evaluation of that product/brand. As for the utilitarian products, no study has yet found significant results in either using localized or globalized packaging (Khan et al. 2017; Khan et al., 2017; Khan et al. 2015). An explanation for not finding any differences is given by Muniz and O’guinn (2001) and Khan et al. (2015). These studies state that consumers are more concerned with product functionality rather than implicit attributes such as packaging cues when buying utilitarian products, which eventually is carrying more weight than the understandable language benefit. The consumption motive of utilitarian product is after all to fulfil a need and therefore the brand will not be judged differently since localizing or
globalizing has no effect. Therefore the following hypotheses are proposed.
H1a: For a hedonic product, the brand is rated more positively when using standard international packaging than it would be for local packaging.
H1b: For a utilitarian product, the brand is not rated more or less positively when using standard international packaging than it would be for local packaging.
6.8 Packaging likeability
when the packaging is globalized compared to localized packaging. Parts of these findings are due to the self-congruity theory (Sirgy, 1986). This theory supposes that products are chosen if they fit the self-image (Kressmann, Sirgy, Herrmann, Huber, Huber & Lee, 2006). When the consumer’s motive to purchase a global product is out of hedonic interest linked to the packaging being written in a foreign language, having a hedonic product with global packaging will lead to a higher level of congruity and thus a higher level of packaging
likeability (Khan et al., 2017). Another explanation for these findings is based on the fact that hedonic products are evaluated more on elements such as packaging aesthetics rather than functional attributes (Hagtvedt and Patrick, 2009). Given the fact that foreignness is also associated more with hedonism (Gerritsen, Korzilius, Van Meurs & Gijsbers, 2000), having a hedonic product with a global packaging will lead to more favourable packaging evaluations than the localized packaging. As mentioned above, this is not the case for utilitarian products. These products are valued on their performance and are less amenable to local vs. global packaging strategies. Therefore evaluating the packaging is not susceptible to either one of the packaging strategies. Thus, the following hypotheses are proposed.
H2a: For a hedonic product, the packaging is rated more positively when using standard international packaging than it would be for local packaging.
H2b: For a utilitarian product, the packaging is not rated more or less positively when using standard international packaging than it would be for local packaging.
6.9 Willingness to pay
As a third dependent variable, willingness to pay has been incorporated into this study as well. Willingness to pay is a variable that naturally follows the variable brand likeability. It is no unknown phenomenon that a brand with high likeability has the ability to set higher price points (Dodds, Monroe, & Grewal, 1991). According to the study of Krishna (1991)
willingness to pay can be defined as the financial offer a customer is willing to make in return for the intended product. No attempt is made yet to see if a hedonic product with a global packing strategy will lead to a higher willingness to pay. Only the study of Khan et al. (2017) suggests that the willingness to pay based on both the product type as well as the packaging strategy is something that could turn out to be significant based on the fact that it leads to higher brand and packaging likeability as well. The study also measures the importance of the
different attributes on the consumer choice. Results show that respondents rate price as the second most important choice attribute after packaging. This result tells us that customers may be willing to pay more for hedonic products in global packaging than in local packaging due to the perceived foreignness and quality (Khan et al., 2017). The positive evaluation of a hedonic product with a global packaging strategy could enhance the willingness to pay. There are no evident results that could suggest the same for utilitarian products. Again, these
products are rated on their functionality (Khan et al., 2017). Customers will decide what they pay based on how good they think the product will work. There are numerous variables that play a role in this judgement but packaging language has not been proven to be one of them (Underwood & Klein, 2002). Therefore the following hypotheses are proposed.
H3a: For a hedonic product, the willingness to pay is higher when using standard international packaging than it would be for local packaging.
H3b: For a utilitarian product, the willingness to pay is not higher or lower when using standard international packaging than it would be for local packaging.
6.10 Conceptual model and hypotheses
The conceptual model portrays all variables that are incorporated into this study together with all expected relations and hypotheses. This overview shows which directions and which type of relationships this study aims to find.
Name Hypotheses H1a H1b H2a H2b H3a H3b
For a hedonic product, the brand is rated more positively when using standard international packaging than it would be for local packaging.
For a utilitarian product, the brand is not rated more or less positively when using standard international packaging than it would be for local packaging. For a hedonic product, the packaging is rated more positively when using standard international packaging than it would be for local packaging.
For a utilitarian product, the packaging is not rated more or less positively when using standard international packaging than it would be for local packaging. For a hedonic product, the willingness to pay is higher when using standard international packaging than it would be for local packaging.
For a utilitarian product, the willingness to pay is not higher or lower when using standard international packaging than it would be for local packaging. Table 1: Overview hypotheses
7. Methodology
The methodology section will entail the procedures and antecedent measurements and steps that were taken before conducting the main analyses. This part consists out of the description of the method, respondents, design, procedure, measurements and pre-tests.
7.1 Research method
This study aimed to determine if manipulating an independent variable had a direct effect on a dependent variable. An experiment is the only research method that is able to determine if there are truly causal effects (Yin, 2013) and therefore an experiment has been conducted. This study tried to prove which brand positioning strategy (global vs. local) led to the highest brand likeability, packaging likeability and willingness to pay. To incorporate the proposed moderator, four conditions were set up that were manipulated on both the brand positioning strategy (local/global) and the type of product (hedonic/utilitarian). This made this study a 2 (local/global positioning) x 2 (hedonic/utilitarian product) between subjects factorial design. 7.2 Sampling method
The respondents (N=213) of this study were recruited by sharing a Facebook and Instagram post, which included a link to the survey. Respondents were also asked directly by mail or direct message. Thus, this sampling method could be defined as a convenience sample, meaning that the respondents were not recruited fully randomly. Doing so was not possible due to time and financial resources, which were too limited. Most respondents were students and friends and were part of the personal network. However, due to snowballing, a lot of non-related and more varied respondents took part in this study as well.
7.3 Respondents
The respondents were on average 25.6 (SD=8.5) years old. Of all respondents, 46.5 percent were male and 53.5 percent were female. The respondents were quite highly educated. 55.9 percent were educated on University level (WO in Dutch). 27.2 of all respondents were educated on Higher Vocational Education level (HBO in Dutch). The remaining 16.9 percent were either educated on high school or Intermediate Vocational Education level. The
respondents had salaries that were below average. 47.9 percent stated that they earned between zero and 1000 euros a month on average (after taxes). 32.4 percent stated that they earned between 1001 and 2000. The remaining 16.4 percent stated that they earned more and 3.3 percent stated that they did not have an income at all.
The sample did not include any product related limitations. Since both the products used in the study (Batteries and Chocolate) are not particularly bound to a certain age or educational group, including all ages and educational levels seemed justified. However, respondents needed to speak Dutch and be Dutch. This limitation was set, because people that do not speak Dutch are not able to be aware of the manipulation in the packaging and thus will be interfering with the internal validity of this study. The survey was written in Dutch to make sure only Dutch speaking respondents could fill it in. There was also a question
included in the survey that tried to check this. Results showed that 100 percent of the respondents spoke Dutch and were living in the Netherlands.
The respondents were assigned to the four conditions completely at random. Meaning that a randomizer was used to determine which respondent was assigned to which condition without being biased. After the randomisation, all respondents were assigned to either one of four conditions (see table 3). All condition had 50 respondents or more, making the groups statistically comparable (Yin, 2013).
Hedonic Utilitarian
Local 57 54
Global 52 50
Table 2: Distribution of participants over the four conditions 7.4 Pre-test
7.5 Product selection
Ten products had been pre-tested. All of them were selected from previous research in which these products already had proven that they were perceived as either hedonic or utilitarian. The study of Voss, Spangenberg, and Grohmann (2003) showed that alkaline batteries and paperclips were accessed as being very utilitarian. Beer and wine (both red and white) proved to be a product scoring high on hedonism. Besides these five products, toothpaste and
detergent showed to be perceived as utilitarian by Sloot, Verhoef and Franses (2005). The last products were retrieved from the study of Batra and Ahtola (1991), this study showed that ice cream and chocolate were perceived as hedonic and water as utilitarian. Using different studies lowered the chance of choosing out-dated or insignificant results. To make sure the brand is not interfering with the results, all products were part of the private brand of Albert Heijn (see all product images in appendix C).
7.6 Hedonic and utilitarian measurements
After selecting the products, a scale retrieved from Voss et al. (2003) measured what hedonic/utilitarian scores the products evoked. This scale existed out of a ten-item seven-point bipolar scale. The utilitarianism was measured by the following five items:
effective/ineffective, helpful/unhelpful, functional/not functional, necessary/unnecessary and practical/impractical. The items measuring hedonism were not fun/fun, dull/exciting, not delightful/delightful, not thrilling/thrilling and enjoyable/unenjoyable. The respondents had to access each product and rate the product on the seven-point ten-item scale.
The respondents of the first pre-test (N=27) had an average age of 28,9 (SD=12,3) years old. 68,6 percent of the respondents were male and 31,4 percent were female. 87,3 percent were educated on University level and 12,7 percent were educated on Higher Vocational Education level. All respondents stated that they lived in the Netherlands. As shown in table 2, only three products scored significantly higher on either the hedonic or utilitarian scale. Therefore the product used in the main study had to be chosen from these three. As for the hedonic product, chocolate has been chosen since it is the only product scoring significantly higher on the hedonic scale, t (25) = -1.297, p = 0.001 , CI = [-1.26, 0.96]. As for the utilitarian scale, two products scored significantly higher (batteries and paperclips). However, since batteries scored the highest individual score on the scale that measured perceived utilitarianism (see table 2), and scoring significantly higher, t (25) =
6.289, p = 0.009, CI = [2.51, 5.20] than the hedonic scale, the product was chosen to be part of the main study.
Table 3: Utilitarian and hedonic levels of selected product groups (n = 27) 7.7 Trustworthiness of the manipulated product
Besides the pre-test determining the products that were used in this study, another pre-test was conducted that tried to see if the manipulated products (chocolate and batteries) were perceived as trustworthy. This was done because the brand name and the language on the packaging were manipulated which could have led to a decrease in believability of the
product portrayed and an increase of the awareness of the manipulation (Schmitt, 1994). This pre-test was conducted amongst 12 University students of which 75 percent were male and 25 percent female. The respondents had to rate the images of the products by stating if they believed the product was a real product. All respondents stated ‘yes’ for all conditions.
Average utilitarian level Average hedonic level Classification (1 = low, 7 = high) (1 = low, 7 = high)
Wine (red) 3.70 4.28 Hedonic
Wine (white) 3.53 4.03 Hedonic
Detergent 5.18 2.25 Utilitarian
Toothpaste 5.23 3.23 Utilitarian
Water 3.92 1.95 Utilitarian
Chocolate 1.71 5.15 Hedonic*
Ice cream 3.96 4.32 Hedonic
Beer 3.55 3.65 Hedonic
Batteries 5.55 1.69 Utilitarian*
Paperclips 5.45 1.72 Utilitarian*
positioning (local and global) and product type (hedonic and utilitarian). Global brand positioning was defined as the language on the packaging being in English (Crystal, 2013). The packaging of the products that were in that condition (both chocolate and batteries) was already written in English and required no change except for the manipulation of the brand name and logo. The local condition however had to be translated to Dutch. The items that were translated were the descriptive texts (slogan, product name, warranty claims etc.). To make sure the translation was credible, three Dutch natives checked the translation making sure the words meant exactly the same in Dutch as they did in English. The brand name and logo were manipulated to create a fake brand so that people would not be biased, creating distorted results. The original chocolate brand was named Woolworths Gold (Woolworth, 2018). This Australian private brand was unknown in the Netherlands. However, the brand name could bias the results and therefore the brand name was changed to Degani, which is a non-existent, neutral brand name. The manipulated pack of batteries was originally sold by Rayovac (Rayovac, 2018). Since this brand is sold in the Netherlands, the brand name was changed to Voltage. This neutral brand name had both meaning in Dutch and English and therefore making is suitable and unbiased to respondents. The images that respondents were exposed to also included a light blue background and the logo of Albert Heijn, the biggest supermarket chain in the Netherlands. This was done to incite the perception that the product was being sold in the Netherlands. The image also contained a small text (in Dutch) stating ‘new in our product range’ invigorating the perception of a Dutch supermarket selling the product even more (all stimulus materials shown in appendix A). The reason for inciting this perception was again to prevent biased results. Without this perception, respondents could perceive the product as being sold in a foreign country. The study of Kim, Eves and Scarles (2009) showed that consumers act differently on language cues in foreign countries as they would in their domestic country, thus excluding this perception prevented skew results. 7.9 Measurements
This study contained three dependent variables all measured with different scales. These three variables were brand likeability, packaging likeability and willingness to pay.
The variable brand likeability was measured by using a scale retrieved from Spears and Singh (2004). The five-item seven-point bipolar scale measured the likeability of the brand that people had been exposed to. The scale existed out of the following items:
unappealing/appealing, bad/good, unpleasant/pleasant, unfavourable/favourable and unlikeable/likeable. These items were translated into Dutch under supervision of Dutch natives. To see if the items would make a usable scale, a factor analysis has been conducted. This analysis and scree plot showed that all five items were contributing to one scale with an Eigen value higher than one (Eigen value = 3.98). The scale predicted 79.6 percent of
variance. All items had factor loadings which where higher that 0.45. Thus, all items could be taken into the reliability analysis. This analysis showed that the scale could be perceived as reliable (Cronbach’s alpha = 0.94). The scale was not ameliorable by excluding one of the items and therefore all items were merged to create an average scale measuring brand likeability. This scale had a minimum score of one and a maximum score of seven. Packaging likeability
The second dependent variable that was used in this study endeavoured to measure packaging likeability. The existing scale was retrieved from the article of Yoo and Donthu (2001) and consisted out of a three-item seven-point bipolar scale. This scale included the items
bad/good, awful/nice and unattractive/attractive. These items have been translated into Dutch under supervision of Dutch natives. A factor analysis has been conducted to see if all items formed one scale. This analysis and scree plot showed that all three items were contributing to one scale with an Eigen value higher than one (Eigen value = 2.58). The scale predicted 86 percent of variance and included only items with factor loadings higher than 0.45. Hence, all items could be taken into the reliability analysis, which showed that the scale could be perceived as reliable (Cronbach’s alpha = 0.92). The scale was not ameliorable by excluding one of the items and therefore all items were merged to create an average scale measuring
maximum amount that they would be willing to spend in euros on the product shown. This method was based on the method of Breidert Hahsler and Reutterer (2006).
7.10 Attention, control and manipulation checks
Three attention checks were conducted to see if the respondents payed attention to the stimulus material. First respondents were asked what type of product they had seen. The question was followed by four options (chocolate, batteries, cookies or paperclips). The same was done for the brand name on the packaging (Voltage or Degani). To make sure
respondents were really paying attention, an extra questions asking the name of the shop that sold these products (Albert Heijn) was asked.
Two manipulation checks were conducted that tried to see if the manipulations had worked. Since the perceived hedonism and utilitarianism was already pre-tested, only the perceived localness vs. globalness and brand name were checked. Asking participants in what language the packaging was written checked the perceived localness/globalness. Asking respondents if they were known to the brand portrayed, checked the manipulation of the brand name.
The control variables that were part of this study were purchase frequency of the portrayed product and the level of English and level Dutch that respondents spoke. The level of English and Dutch was asked because people’s level of either one of these languages could interfere with the results since the language on the packaging was manipulated. Both were measured on a single-item seven-point bipolar scale, which asked how the respondents would describe their level of English or Dutch (very bad/very good). Secondly, the consumption rate of the hedonic or utilitarian product shown could have played a role as well. If people are regularly buying a product, they might have a predefined preference for either one of these products making the results on the depended variable obtrusive. This variable was measured by asking respondents how often they were buying products from the same product category a month and had six options (never, 1 or less, 2-4, 4-6, 7-8, more than 8). These options were based on the average consumption rates of both batteries and chocolate. The average for batteries in Europe or the Netherlands was 0.8 times a month and for chocolate 2.4 times a month (Globalqyresearch, 2016; Statista, 2018). Thus, creating a scale covering this scope.
7.11 Procedure
Respondents were recruited online and were asked to attend the survey by clicking on a link. The survey was built with Qualtrics and existed out of 21 questions. First, respondents had to agree with the terms and conditions of the study that secured their privacy and informed them about the research that followed. After that, the demographic variables were measured
including age, gender and level of income. A text encouraging the respondents to take a close look at the upcoming image followed. Then the respondents were assigned to one of the four conditions randomly. After the exposure to the conditions, the dependent variables brand likeability, packaging likeability and willingness to pay were measured. All dependent variables included a page break so respondents would not know what the next questions were about and thus minimalizing learning effects. After the dependent variables, the control variables were measured. These variables included familiarity with the brand, amount of product category purchases a month, country of residence, speaking Dutch, level of speaking Dutch, speaking English, level of speaking English, speaking other languages and number of languages spoken by respondents. These questions were followed by the manipulation checks. These included questions that checked if respondents knew which product was show, what the name of the brand was, what the language on the packaging was and what the name of the store was in which people could buy the product.
After the initial questionnaire, respondents had the chance to enter their e-mail address to win the incentive prize. This incentive was a JBL speaker, which was worth 59 euros. By entering their e-mail address they were part of the raffle that was executed on the 3th of June 2018 and the winner received an e-mail that asked for their address and name so that the prize could be delivered.
8. Results
In this section the analyses that have been conducted to answer the research question and the hypotheses will be described including the results that came out of these analyses.
8.1 Randomisation checks
All respondents of this study (N=213) have been assigned to the four conditions randomly using a randomizer. To see if this was actually the case some variables have been checked to see if the conditions were not significantly different. The variables that functioned as
randomisation checks were age, educational level and level of income. Before analysing the variable age, a Kolomogorov-Smirnov normality check has been conducted. This analysis showed that the variable age did not fulfil the assumption of normality for the condition local/hedonic; D(57) = 0.34, p < 0.001, local utilitarian, D(54) = 0.37, p < 0.001,
global/hedonic, D(50) = 0.39, p < 0.001 and global/utilitarian, D(52) = 0.30, p < 0.001. The Levene’s test of homogeneity of variance showed that the homogeneity of variance
assumption could not be fulfilled either, F (3, 209) = 7.74, p < 0.001. Hence, a non-parametric analysis has been conducted. A Kruskal-Wallis H-test showed that there was no statistically significant difference in age between the four conditions, χ2(3) = 1.474, p = 0.688. The variable educational level was checked by a cross tabulation, Chi-Square (15) = 14.40, p = 0.496. Results showed that there were no differences between the educational levels of respondents between the four conditions. This was also done with the variable level of income, Chi-Square (21) = 24.75, p = 0.258. This variable also showed no significant differences between the four conditions. Hence, all respondents were assigned to the conditions randomly.
8.2 Attention checks
To see if the respondents were aware of the conditions, three attention checks have been conducted. These checks included reproducing the product, brand name and distribution channel. First the awareness of the product that respondents were exposed to was checked. To do so a cross tabulation, Chi-Square (1) = 213.00, p < 0.001, was conducted. This analysis showed that there was a significant effect of the condition respondents were in on the perceived product. All respondents had stated the right answer after asking which product
they had seen. Secondly, the same was done for the brand name of the product. A cross tabulation, Chi-Square (3) = 184.36, p < 0.001, was executed showing a significant effect of the conditions on the perceived brand names. 95.2 percent was aware of the brand name Voltage (utilitarian) and 90.2 percent was aware of the brand name Degani (hedonic). After this analysis, the perceived distribution channel was checked to see if people were paying attention during the manipulation. Off all the respondents, 87,8 percent of the respondents were aware of the distribution channel being Albert Heijn. Meaning that the targeted 80 percent was achieved (Yin, 2013). All three analyses showed that the respondents were aware of the manipulation and paid full attention.
8.3 Manipulation checks
To see if the manipulation in this study had worked, two manipulation checks were conducted. Since the levels of perceived hedonism and utilitarianism of the product were already pre-tested, only the perceived packaging language was checked. To do so, a Fishers-exact test was conducted since both variables formed a 2x2 table, Fisher-Fishers-exact p < 0.001. Hence, a significant effect of the conditions on the perceived language which was strong (tau = 0.571). 69.6 percent of the respondents were aware of the English (global) packaging and 86.5 percent were aware of the Dutch (local) packaging. This implies that the manipulation of brand language was executed successfully. A second manipulation check was done to see if people were unknown to the manipulated brand name. A question asking if people were familiar with the brand name of the product, showed that 78,9 percent of all respondents were not familiar with the brand portrayed. Although this number was not as high as expected, the percentage was still high enough to conclude that the manipulation had worked.
Level of Dutch
The level of Dutch was measured with an ordinal scale and was controlled for both brand and packaging likeability. Before analysing the level of Dutch, a Kolomogorov-Smirnov
normality check was conducted. This analysis showed that the variable level of Dutch did not fulfil the assumption of normality since the scores on brand likeability were not distributed normal on all levels of Dutch; score 6 (good) D(38) = 0.19, p = 0.002 and score 7 (very good), D(170) = 0.14, p < 0.001. The Levene’s test of homogeneity of variance showed that the homogeneity of variance assumption could be fulfilled, F (3, 208) = 2.04, p = 0.109. Since the ANOVA analysis is quite robust (Yin, 2013), a two-way ANOVA including level of Dutch and packaging language seemed justified and was carried out. This analysis showed that for brand likeability, no significant results were found, F (4, 205) = 1.65, p = 0,164.
The same was done for the variable packaging likeability. To see if the scores of packaging likeability were normally distributed for the different levels of the scale measuring level of Dutch, a Kolomogorov-Smirnov normality check was conducted. This analysis showed that the variable level of Dutch did not fulfil the assumption since the scores on packaging likeability were not distributed normally for all levels of Dutch; score 6 (good), D(38) = 0.19, p = 0.074 and score 7 (very good), D(170) = 0.14, p < 0.001. The Levene’s test of homogeneity of variance showed that the homogeneity of variance assumption could not be fulfilled either, F (3, 208) = 3.61, p = 0.014. However, histograms showed that the brand likeability scores on the two values of level of Dutch were somewhat distributed normally. Since the ANOVA analysis is quite robust (Yin, 2013), a two-way ANOVA including level of Dutch and packaging language seemed justified and was carried out. This analysis showed that for brand likeability, no significant results were found, F (4, 205) = 0.98, p = 0,418. Hence, no significant results were found for the level of Dutch respondents spoke for both brand and packaging likeability.
Level of English
After that, the same was done for the level of English. Again, a Kolomogorov-Smirnov normality check was conducted. This analysis showed that the variable level of English did not fulfil the assumption since the scores on brand likeability were not distributed normally for all levels of English; score 4 (not that good or bad), D(16) = 0.17, p = 0.200, score 5 (quite
good), D(44) = 0.12, p = 0.174, score 6 (good), D(82) = 0.15, p < 0.001 and score 7 (very good), D(67) = 0.21, p < 0.001 The Levene’s test of homogeneity of variance could not be conducted because of the constant scores on 2 values of the scale measuring level of English. The plotted histograms showed that the scores were not distributed normally and therefore bootstrapping has been applied to the variable brand likeability. After that a two-way ANOVA analysis including level of English and packaging language has been conducted which did not result in significant different results for the variable brand likeability, F (3, 202) = 0.79, p = 0,502.
The same was done for the dependent variable packaging likeability. Again, a Kolomogorov-Smirnov normality check was conducted. This analysis showed that the variable level of English did not fulfil the assumption since the scores on brand likeability were not distributed normally for all levels of English; score 4 (not that good or bad), D(16) = 0.17, p = 0.200, score 5 (quite good), D(44) = 0.12, p = 0.174, score 6 (good), D(82) = 0.15, p < 0.001 and score 7 (very good), D(67) = 0.21, p < 0.001 The Levene’s test of homogeneity of variance could not be conducted because of the constant scores on 2 values of the scale measuring level of English. The plotted histograms showed that the scores were not
distributed normally and therefore bootstrapping has been applied to the variable packaging likeability. After that a two-way ANOVA analysis including level of English and packaging language has been conducted which did not result in significant different results for the variable packaging likeability, F (4, 205) = 1.60, p = 0,190. Hence, no significant results were found for the level of English respondents spoke.
Purchase frequency of the portrayed product
Lastly, the variable monthly purchase frequency of the portrayed product has been tested to see if this variable would have had any influence on the brand likeability, packaging
ANOVA has been conducted including the packaging language and product type as individual variables. The variable purchase frequency turned out not to be significant and therefore did not play a role in the main analyses, F (1, 205) = 1.83, p = 0.177.
The same was done for the variable packaging likeability. A Kolomogorov-Smirnov normality check was conducted which showed that the purchase frequency did not fulfil the assumption. The scores on packaging likeability were not distributed normally for all levels of purchase frequency; ‘never’, D(42) = 0.13, p = 0.075, ‘less than 1’, D(122) = 0.13, p < 0.001 and ‘2-4 times’, D(42) = 0.16, p = 0.010. However, the Levene’s test of homogeneity showed non-significant results, F (3, 209) = 0.60, p = 0.613. Hence, since an ANOVA is quite robust (Yin, 2013), a three-way ANOVA has been conducted including the packaging language and product type as individual variables. The variable purchase frequency turned out not to be significant and was therefore not taken into account in the main analyses, F (1, 205) = 3.67, p = 0.056.
Lastly, the variable willingness to pay has been checked in this manner. A
Kolomogorov-Smirnov normality check, showed that the variable willingness to pay did not fulfil the assumption. The scores on willingness to pay were not distributed normally for all levels of purchase frequency; ‘never’, D(42) = 0.13, p = 0.082, ‘less than 1’, D(122) = 0.21, p < 0.001 and ‘2-4 times’, D(42) = 0.22, p < 0.001. The Levene’s test of homogeneity showed significant results, F (3, 209) = 4.64, p = 0.001. However, since an ANOVA is quite robust (Yin, 2013), a three-way ANOVA has been conducted including the packaging language and product type as individual variables, F (1, 205) = 1.56, p = 0.331. Hence, no significant results were found so there were no effects of the purchase frequency that had to be accounted for.
8.5 Main effects
Three two-way ANOVA analyses have been conducted to answer the hypotheses proposed. Each analysis has been done for one of the three dependent variables
Brand likeability
First the dependent variable brand likeability was tested. Before doing so, the normality of both the packaging language and product type had to be tested. A Kolomogorov-Smirnov normality check showed that the variable packaging language did not fulfil all of the
assumptions. The scores on brand likeability were not distributed normally for both levels of product language; local, D(111) = 0.16, p < 0.001 and global, D(102) = 0.13, p < 0.001. However, the Levene’s test of homogeneity showed non-significant results, F (1, 211) = 0.07, p = 0.792. The same results were found for product type, Kolomogorov-Smirnov normality check; hedonic, D(109) = 0.16, p < 0.001 and utilitarian, D(104) = 0.13, p < 0.001, Levene’s test of homogeneity; F (1, 211) = 0.01, p = 0.921. Taking into account that both dependent variables did suffice to support the homogeneity assumption and the fact that an ANOVA analysis is quite robust, a two-way ANOVA has been conducted. The only significant result was that of the product type, F (1, 209) = 26.26, p < 0.001 (see table 4). However, the average scores on brand likeability (see table 5) did not differ significantly within the two types of packaging languages (local vs. global). Hence, there was no interaction effect.
Therefore the hypothesis, H1a: For a hedonic product, the brand is rated more positively when using standard international packaging than it would be for local packaging, had to be
rejected. The hypothesis, H1b: for a utilitarian product, the brand is not rated more or less positively when using standard international packaging than it would be for local packaging, was accepted.
Table 4: Main and interaction effects on brand likeability
df Mean Square F Sig. Eta2 Brand likeability Corr. Model 3 10.73 9.01 <0.001* 0.11 Pac. Language 1 0.87 0.72 0.395 0.01 Product type 1 31.28 26.26 <0.001* 0.11 Pac. Language * 1 0.03 0.02 0.885 0.00 Product type
Packaging likeability
Secondly, the same process was applied to the variable packaging likeability. Starting with the normality of both the packaging language and product type which had to be tested. A Kolomogorov-Smirnov normality check showed that the variable packaging language did not fulfil all of the assumptions. The scores on packaging likeability were not distributed
normally for both levels of product language; local, D(111) = 0.11, p < 0.001 and global, D(102) = 0.13, p < 0.001. However, the Levene’s test of homogeneity showed non-significant results, F (1, 211) = 0.07, p = 0.921. The same results were found for product type,
Kolomogorov-Smirnov normality check; hedonic, D(109) = 0.14, p < 0.001 and utilitarian, D(104) = 0.13, p < 0.001, Levene’s test of homogeneity; F (1, 211) = 1.78, p = 0.184. Taking into account that both dependent variables did suffice for the homogeneity assumption and the fact that an ANOVA analysis is quite robust, a two-way ANOVA has been conducted. The only significant result was that of the product type, F (1, 209) = 21.39, p < 0.001 (see table 6). However, the average scores on brand likeability (see table 7) did not differ significantly within the two types of packaging languages (local vs. global). Hence, there was no interaction effect. Therefore the hypothesis, H2a: For a hedonic product, the packaging is rated more positively when using standard international packaging than it would be for local packaging, was rejected. The hypothesis, H2b: For a utilitarian product, the packaging is not rated more or less positively when using standard international packaging than it would be for local packaging, was accepted.
Table 6: Main and interaction effects on packaging likeability
df Mean Square F Sig. Eta2 Packaging likeability Corr. Model 3 12.20 7.16 <0.001* 0.09 Pac. Language 1 0.12 0.07 0.793 0.01 Product type 1 36.44 21.39 <0.001* 0.09 Pac. Language * 1 0.01 0.01 0.964 0.00 Product type
Table 7: The average scores on packaging likeability Willingness to pay
The last variable that was tested was the willingness to pay. First, the normality of both the packaging language and product type which had to be tested. A Kolomogorov-Smirnov normality check showed that the variable packaging language did not fulfil all of the assumptions. The scores on packaging likeability were not distributed normally for both levels of product language; local, D(111) = 0.16, p < 0.001 and global, D(102) = 0.13, p < 0.001. However, the Levene’s test of homogeneity showed non-significant results, F (1, 211) = 0.07, p = 0.792. The same results were found for product type, Kolomogorov-Smirnov normality check; hedonic, D(109) = 0.19, p < 0.001 and utilitarian, D(104) = 0.09, p = 0.039, Levene’s test of homogeneity; F (1, 211) = 2.92, p = 0.089. Taking into account that both dependent variables did suffice to the homogeneity assumption and the fact that an ANOVA analysis is quite robust, a two-way ANOVA has been conducted. The only significant result was that of the product type, F (1, 209) = 109.39, p < 0.001 (see table 8). However, the average scores on brand likeability (see table 9) did not differ significantly within the two types of packaging languages (local vs. global). Hence, there was no interaction effect. Therefore the hypothesis H3a: For a hedonic product, the willingness to pay is higher when using standard international packaging than it would be for local packaging, had to be
rejected. The hypothesis H3b: For a utilitarian product, the willingness to pay is not higher or N M SD SE 95% CI Packaging likeability Hedonic/global 52 4.76 1.23 0.18 [4.40; 5.11] Hedonic/local 57 4.80 1.25 0.17 [4.46; 5.14] Utilitarian/global 50 3.92 1.38 0.19 [3.56; 4.28] Utilitarian/local 54 3.98 1.36 0.18 [3.63; 4.33]
Table 8: Main and interaction effects on willingness to pay
Table 9: The average scores on willingness to pay
df Mean Square F Sig. Eta2 Brand likeability Corr. Model 3 266.04 36.67 <0.001* 0.35 Pac. Language 1 1.61 0.22 0.395 0.01 Product type 1 793.60 109.39 <0.001* 0.34 Pac. Language * 1 0.18 0.03 0.964 0.00 Product type
*significant at the level of 0.05
N M SD SE 95% CI Willingness to pay in € Hedonic/global 52 2.71 0.91 0.37 [1.97; 3.44] Hedonic/local 57 2.48 0.95 0.36 [1.77; 3.20] Utilitarian/global 50 6.51 4.32 0.38 [5.76; 7.27] Utilitarian/local 54 6.40 3.09 0.37 [5.68; 7.12]
Name Hypothesis Outcome H1a H1b H2a H2b H3a H3b
For a hedonic product, the brand is rated more positively when using standard international packaging than it would be for local packaging.
For a utilitarian product, the brand is not rated more or less positively when using standard international packaging than it would befor local packaging.
For a hedonic product, the packaging is rated more positively when using standard international packaging than it would be for local packaging.
For a utilitarian product, the packaging is not rated more or less positively when using standard international packaging than it would for be local packaging.
For a hedonic product, the willingness to pay is higher when using standard international packaging than it would be for local
packaging.
For a utilitarian product, the willingness to pay is not higher or lower when using standard international packaging than it would be for local packaging.
Declined Accepted Declined Accepted Declined Accepted
9. Conclusion
Three two-way ANOVA analyses have been conducted to answer the research question: What is the effect of choosing either a globalised (English) or localised (Dutch) packaging strategy for both hedonic and utilitarian products on brand likeability, packaging likeability and willingness to pay? The outcomes of these analyses were analysed and compared with the pre-defined hypotheses and discussed in the following section.
9.1 Brand likeability
The effect of localising or globalising product packaging has proven to be moderated by the product type (hedonic vs. utilitarian) in previous studies (Khan et al., 2017; Kahn & Lockshin, 2017; Lockshin et al., 2015; Hu et al., 2016; Leclerc et al., 1994) in relation to the brand likeability of the product shown. According to the study of Khan et al. (2017), reasons for these findings can be traced back to the fact that consumers seek for fun in their consumption, which was enhanced by a foreign language. Including this foreign language affected the perception of hedonism and therefore increased the evaluation of that product/brand. In this study this was unfortunately not the case. The results on the variable brand likeability were higher for a local/utilitarian product than they were for a global/utilitarian product
(chocolate). However, these differences were far from significant. The same results were shown within the utilitarian product range (batteries). Therefore the hypothesis H2a: For a hedonic product, the brand is rated more positively when using standard international
packaging than it would be for local packaging, was not accepted. The hypothesis H2b: For a utilitarian product, the brand is not rated more or less positively when using standard
international packaging than it would be for local packaging, was accepted. An explanation can be found in the fact that all previous studies have been conducted in non-English speaking non-western countries. (Khan et al., 2017; Kahn & Lockshin, 2017; Lockshin et al., 2015; Hu et al., 2016). Global products were rated as luxury goods and obtained a certain social status. This was not the case here in the Netherlands and therefore the hedonic value that the
globalness of the product would contain was diminished. Another reason for not finding any significant results could be the fact that in this study, only the descriptive text on the
packaging was translated. The trade off that has been made between the internal validity and the chances of finding significant results has turned out negative. A brand name carries higher emotional value than descriptive text and would therefor be more suitable for evoking effects
concerning brand evaluation (Griffith et al., 2003). However, brand names are more easily translated when a language carries a different alphabet, which is not the case in Dutch. Since no study has been conducted in a non-English western country, the only reference for
prediction was the study of LeClerc et al. (1999). This study showed that the same results were portrayed for the brand name of products in advertisements. However these results have proven to be non-applicable to the field of packaging. This study contributed to that field of research and gave clear indications that could be useful in future studies.
9.2 Packaging likeability
Secondly, the variable packaging likeability has been taken into account. This variable has been proven to be a significant predictor of the likeability of the packaging of products. Some studies (Khan et al., 2017; Kahn & Lockshin, 2017; Lockshin et al., 2015) predicted that the evaluation of a hedonic product with a global packaging strategy would be higher because of the fact that hedonic products are evaluated more on elements such as packaging aesthetics rather than functional attributes (Hagtvedt and Patrick, 2009). Given the fact that hedonism is also associated more with foreignness (Gerritsen et al., 2000), having a hedonic product with a global packaging would lead to more favourable packaging evaluations than the localized packaging. This would not be the case for utilitarian products where packaging has a
subordinate role and where functionality is a key factor. Because of this functionality, having text that is easier to read (Dutch) makes the packaging more attractive since customers are more easily informed about the performance of the product (Khan et al., 2017; Kahn & Lockshin, 2017) Therefor two hypotheses have been proposed, H3a: For a hedonic product, the packaging is rated more positively when using standard international packaging than it would be for local packaging, which was rejected. The hypothesis H3b: For a utilitarian product, the packaging is not rated more or less positively when using standard international packaging than it would be for local packaging, was accepted. A possible explanation for
scored high on the utilitarian score but it could be that consumers already assign chocolate to a certain country, which could have led to distorted results. This was not checked for. 9.3 Willingness to pay
The last variable was tested was the willingness to pay. The existing literature had not tapped into this variable yet. Studies limited themselves to the likeability of both the brand and the packaging. However, according to the study of (Kahn & Lockshin, 2017) brand likeability is a predictor of willingness to pay. If a customer likes a brand of a product he is willing to pay more for that product. Therefore the following hypotheses have been proposed.
The first hypothesis (H4a) stated: For a hedonic product, the willingness to pay is higher when using standard international packaging than it would be for local packaging, was
declined. The second hypothesis (H4b) stated: For a utilitarian product, the willingness to pay is not higher or lower when using standard international packaging than it would be for local packaging, was accepted. The analyses unfortunately showed no significant differences. This was expected because the brand likeability did not differ significantly as well. The study of Dodds et al. (1991) defined this to as a condition for a higher willingness to pay. So the reasons for not finding any results could be traced back to the same reasons for not finding any result in the brand likeability. Besides this reason, another possible reason would be the low score on brand name awareness. Although both brand names were fictive, more than 20 percent stated they knew the brand. This could have led to distorted results because brand names are very influential in the willingness to pay of consumers for that brand (Dodds et al., 1991).
10. Discussion
10.1 Implications
Although none of the analyses conducted in this study led to significant results, we could conclude that this study did lay a foundation for future studies to be conducted in this field. The study was a succession to the study conducted by (Khan & Lockshin, 2017) who found significant results and stressed the importance of expanding this research into other countries and by adding variables like willingness to pay. However, the research topic was too
unexplored to set up hypotheses that turned out to be significant. But this result can be seen as a clarification of the use of languages on packaging in western countries. The results show that the language on the packaging is not that important to customers in western countries. However, many variables have not been studied yet in the field of product packaging. This study could function as a base for possible future research into this matter. Another
contribution that this study made would be the fact that this study had managed to show the hedonic and utilitarian values of certain products. This could be an enhancement for future studies that want to make use of these variables as well. Lastly, the personal contribution to the work of HJ Lockshin in this study could be considered valuable. His attempts to tap into this field of research have been broadened by analysing the same variables in a western country like the Netherlands.
10.2 Limitations
This study did impose some limitations. It did not manage to gain results that were in line with the pre-defined hypotheses. Some of the reasons for these findings already have been mentioned by in the conclusion section. However, it is necessary to understand deeply why these finding did not turn out the way they were expected. Therefore a small overview is
