Acceptance of Suboptimal Food Products: A Choice-Based Conjoint Analysis on the Importance of Appearance, Expiration Date, Price and Organic Labeling

(1)

Acceptance of Suboptimal Food Products:

A Choice-Based Conjoint Analysis on the Importance of Appearance,

Expiration Date, Price and Organic Labeling

(2)

1

Acceptance of Suboptimal Food Products:

A Choice-Based Conjoint Analysis on the Importance of Appearance,

Expiration Date, Price and Organic Labeling

Daphne Timmerhuis June 12, 2020

Master Thesis

MSc Marketing Management & Marketing Intelligence Faculty of Economics and Business

University of Groningen

First Supervisor: Martine van der Heide Second Supervisor: Dr. Felix Eggers

Name: Daphne Timmerhuis Student Number: S2764466

Address: Friesestraatweg 155-18, 9743AA Groningen Phone Number: +31630006143

(3)

2

Abstract

(4)

3

Preface

I would like to express my gratitude toward my thesis supervisors Martine van der Heide and Dr. Felix Eggers. I would like to thank them for their collaboration and their support during my master thesis. The provision of constructive feedback throughout the process and the guidance when I encountered problems has helped me a lot and enabled me to complete the last challenge of my studies.

(5)

4

1. Introduction

1.1. Food Waste Issue

Over the last years there has been drawn considerable attention to the environment, and food waste is causing a big part of the environmental problems. Roughly 1.3 billion tons of food is wasted per year, which equals one-third of all food produced for human consumption (FAO, 2011). Wasted food contributes to excess consumption of freshwater and fossil fuels. This, together with the methane and CO2 emissions from decomposing food, has an impact on the global climate change (Hall et al., 2009). The wastage of food means that the energy and water used to grow, harvest, transport and process the food have also been wasted. By preventing this waste as opposed to diverting it to anaerobic digestion after disposition, about eight times more greenhouse gas emissions can be prevented (Quested et al., 2011). But what causes this food waste?

According to the Food and Agricultural Organization of the United Nations, food waste can be caused by different factors. Quality standards rejecting food items which are not perfect in shape or appearance, make that products do not make it to the store. In Europe, product specifications for fruits and vegetables have been formalized and require levels of quality, ripeness, safety, smell, taste, origin of product, packaging, uniformity across products within one package, shape, skin, size and weight. Products that do not fulfil the specifications are not allowed onto the European market (De Hooge et al., 2018). Next to that, insufficient purchase planning and expiring ‘best-before-dates’ cause large amounts of waste at the consumer level, in combination with the careless attitude of those consumers who can afford to waste food (FAO, 2011).

(8)

7

product packages. These kind of products are often rejected by the customer (Aschemann-Witzel et al., 2015) and cause the customer to contribute to a big share of the global food waste, endangering the world’s future food security and sustainability needs (Jonathan et al., 2011). Williams & Wikström (2011) stated that food waste at the consumer level has a particularly negative environmental impact as all the resources for production, transport, and preparation are wasted.

Figure 1. Suboptimal and Optimal Foods (De Hooge et al., 2017).

To combat this food waste, supermarkets offer suboptimal products against a lower price (Aschemann-Witzel et al., 2017). In the Netherlands, the supermarket chain Albert Heijn sells odd-looking fruits and vegetables, like crooked cucumbers and differently colored apples (Albert Heijn, n.d.-a). The chain also makes use of discount stickers when products get close to their expiration date. Currently Albert Heijn is working on new dynamic pricing technologies, where products get cheaper the closer they get to their expiration date (Albert Heijn, n.d.-b). Also, the supermarket chain Jumbo shows interest in reducing food waste and recently introduced a special shelf for discounted products on their expiration date (Jumbo, n.d.).

1.2. Research Objectives and Problem Statement

(9)

8

difficult. A couple of studies have investigated consumers’ perception toward suboptimal food and the effect that suboptimality has on the purchase intention of the consumer (Stangherlin et al., 2019; Loebnitz et al., 2015). However, no distinction has been made between the effects of the different kinds of suboptimality: whether the product has an unusual look or whether it is close to its expiration date. This study will be among the first to investigate both types of suboptimality and will distinguish between the effects caused by the appearance of the product as well as the expiration date.

Nevertheless, the product choice of the consumer is not only determined by (sub)optimality such as appearance or expiration date, also other factors influence consumers’ product choices. It is widely known that price is an important determinator of consumers’ product choices, and previous research has shed light on the effect of price discounts to encourage the customer to buy specific products (Aschemann-Witzel et al., 2018). Besides, the effect of organic labeling on the attractiveness of a product has been studied (Aschemann-Witzel, 2018; Loebnitz et al., 2015), though contradicting findings make that uncertainty about its influence still exists.

Although a lot of attention has been drawn to suboptimal products and the effect of price and organic labeling on food products, little is known about their relative importance in the decision-making process of the consumer when buying a product. This study sheds light on the following research question:

(1) What is the relative importance of the product features of appearance, expiration date, price and organic labeling with regard to the product choice of the customer?

(10)

9

sustainability issues due to the claim, which causes a positive effect on choosing suboptimal products (Abrahamse et al., 2005). However, if consumers act out of habit or social behavior of others, this effect does not hold (Salmon et al., 2015; Van ’t Riet et al., 2011). The fact that contradicting results mark the field, indicates the matter is still understudied and addresses the need for further research. This study investigates whether a sustainability claim focusing on the environmental impact could influence the choice decision for a suboptimal product. The second research question therefore is:

(2) Can the effect of suboptimal features on the purchase decision be changed by providing a sustainability claim?

To test generalizability of the effects across product types and consumers, several additional factors are investigated. Previous research has focused on different types of suboptimal food, like fruits and vegetables or packaged goods (De Hooge et al., 2017; Symmank et al., 2018). De Hooge et al. (2017) suggest that the product type may have an effect on the intention of the customer to buy suboptimal food products. Therefore, the type of product is taken into account in this study. Furthermore, customer characteristics may be of importance, since several studies have shown the effect of demographics, environmental concern and environmental behavior on the customer’s product choice (Aschemann-Witzel, 2018; Loebnitz et al., 2015). Customer characteristics, as well as different product types, are incorporated in the study to investigate whether the findings can be generalized across different products and consumers. The third and last research question is:

(3) Can the effects be generalized across product types and consumers?

1.3. Contribution and Managerial Implications

(11)

10

whether a sustainability claim influenced their behavior. After the choice tests, participants answered questions which measured consumer characteristics of interest.

New insights are offered regarding the effect sizes and importance of the appearance, expiration date, price and organic labeling in the choice decision of the consumer. This study shows that all features influence the consumer’s product choice, of which price is the most important, followed by the expiration date. Besides, the effect of a sustainability claim on the importance of suboptimal features is investigated, and this claim has proven to be able to reduce rejection toward products with a close expiration date. Furthermore, multiple product types have been tested and differences among consumers are investigated, marking the generalizability of the results.

Managers could use information from this study to better understand the dynamics in the decision-making process of the consumer. From the insights regarding the importance of specific product features, marketing strategies can be adapted so the customer is more prone to buy suboptimal products. New insights about the effect of sustainability claims can enable managers to deliberately implement such claims. Above all, by distinguishing between product types and different consumer groups, managers can better adapt their marketing strategy to the product of interest and can address their target customers in a better way. By implementing strategies which fit the company’s target customers, consumers can be made more prone to buy suboptimal products.

(12)

11

2. Theoretical Background

2.1. Suboptimal Food Products

Suboptimal food products are products that do not meet specific cosmetic specifications, or products which are close to or past its expiration date (De Hooge et al., 2017; Aschemann-Witzel, 2018). Suboptimal products are typically divided into three categories (Van Doorn & Verhoef, 2011). The first category represents fruits and vegetables with an unusual visual appearance and includes products with a deviation from appearance standards like its weight, shape, size, or cosmetic appeal. The second category includes products which are close to or beyond their best-before date, and the third category represents food packages with visual damages, like dented cans or torn wrappers (De Hooge et al., 2017). This research divides suboptimal food into two categories, combining two of the previously named categories into a single one. The first category consists of products with an unusual visual appearance, containing food products with either a deviation from its standard appearance, or with a visually damaged packaging. The second category contains products close to their expiration date.

(13)

12

Suboptimal food products have been proven to cause a considerable part of all food waste in the world. In all of the stages of the supply chain, from harvesting to the post-consumer and end of life stage, food gets wasted (Parfitt et al., 2010). In the product evaluation stage, quality standards cause that suboptimal looking food gets filtered out and does not continue the supply chain. During the distribution stage packages get damaged and in the consumer stage the consumer chooses the product with the ‘cosmetically perfect’ look (Parfitt et al., 2010). To solve the food waste problem, the mechanism behind the purchase decision of the customer is investigated, resulting in less rejecting behavior by the consumer toward suboptimal food and causing the consumer to be more likely to buy these products (Göbel et al., 2015). However, to do so, the factors which cause consumers to choose the products that they do need to be understood, and the relative importance of these factors in the decision-making process needs to be clarified.

2.2. Consumer Choices using Product Utility

To know how important different features are in determining the choice of the consumer, the product utility is used. The product utility is the translation of consumers’ preferences regarding specific characteristics of a product into a number. This number takes the value and the relative weight of each feature into account, and adds them up so that products with different feature compositions can be compared to each other. The product utility can predict purchase decisions in varying conditions, and characteristics which are important can be identified (Eggers & Sattler, 2011). The product with the highest utility is the most valuable to the consumer and is most likely to be chosen. In this study, the term of product utility will be used to indicate the value of a product consisting of different features. By analyzing the effect sizes of each feature, it can be seen how important each feature is in the consumer’s choice decision. Next, product features influencing the product utility, and therefore the purchase decision of the consumer, will be discussed.

2.3. Appearance

(14)

13

produce (Feenstra, 1988; Marx-Pienaar & Erasmus, 2014). According to Parfitt et al. (2010), they shop in a way to obtain the ‘cosmetically perfect’ food, that is fruits and vegetables that are symmetrical in color and shape (Finlayson, 2018). Products which deviate from the usual shape, size or weight are rejected and are classified as suboptimal (Bunn et al., 1990; Göbel et al., 2015). These deviating products can be characterized by for example cosmetic damage scars from traveling, cosmetic damage on the peel, an odd shape or deviating color (Bunn et al., 1990; De Hooge et al., 2017). Regarding the shape of the product, Loebnitz et al. (2015) found that food shape abnormalities influence consumers’ purchase intentions only if the product deviates extremely from the norm. When the product differs just moderately from the usual look, purchase intentions are not affected.

Hussin et al. (2010) found that consumers rate ripeness, freshness, cleanliness and absence of defect and blemish as important attributes for fruits and vegetables. Features like purity, size, appearance or texture are often-used criteria for determining the degree of excellence of a product, although they do not relate to the food safety (Abbott, 1999; Caswell, 1998; Hussin et al., 2010). Even though the suboptimal product remains the same quality and safety assurance as other products, the customer associates these deviations from the usual looking product with quality reduction. This results in rejection of the product due to deviations from the usual size, shape or weight (Göbel et al., 2015).

(15)

14

though the damage is only superficial in nature and does not affect the product’s safety (White et al., 2016). A product with a deviating look or damaged package is therefore expected to have a negative effect on the product utility. According to these theories, the following hypothesis is formed:

H1: The product having an unusual appearance (as opposed to having a usual appearance) has a negative effect on the product utility

2.4. Expiration Date

Next to the appearance of the product being a cause for suboptimal food, also the expiration date is an indicator for the purchase decision of the customer. A lot of food waste is related to these date labels due to consumers rejecting the product because of its close expiration date (Milne, 2013). When the expiration date is soon approaching or has passed, consumers are concerned about the food safety and risk concern. Consumers look for information on the product that indicates freshness, safety and/or viability and so the expiration date influences their decision to purchase (Hall-Phillips & Shah, 2017). Newsome et al. (2014) demonstrated that the majority of consumers (62%) indicates to buy foods with the longest remaining shelf lives. They rated the product with the longest period before reaching the expiration date as the best and the most healthful, suggesting that consumers avoid the purchase of foods that are suboptimal in terms of being close to the best-before date. Extending the best-before date was found to be the top ranked convenience feature by consumers, proving the importance of the expiration date (Koutsimanis et al., 2012). Qi & Roe (2016) found that consumers think that food waste is necessary to reduce the odds of foodborne illness and to ensure that meals taste fresh. This causes that a lot of food is thrown away, even though it is still edible and of good quality (Stangherlin et al., 2019). The food which is unnecessarily thrown away exists mainly of bread, fresh potatoes, milk, home-made and prepared meals, carbonated soft drinks, fruit juice and smoothies, poultry meat, pork meat, cakes, and processed potatoes like chips (WRAP, 2013).

(16)

15

problem, however due the fact that this system was considered to be confusing by customers, this did not work out. Although the intention of the best-before label being to indicate freshness, it fails to do so because some consumers think it indicates safety (Leib et al., 2013). Newsome et al. (2014) found that 71% of the consumers misunderstood the labeling system, while Van Boxstael et al. (2014) showed that only half of the consumers is aware of and appreciates the difference between the two labels. Because of the fact that this labeling system is evidently misunderstood and misinterpreted, consumers on the one hand are increasing the risk for a food disease burden by consuming certain expired ‘use by’ products, which refers to food safety. On the other hand, consumers throw away edible food past or close to its ‘best before’ date (referring to food quality) and consequently waste food, although the product does not pose any negative effects and is suitable to eat (Van Boxstael et al., 2014). This shows, similarly to the wrong interpretation of the visually deviating products, that consumers misinterpret the expiration date as a cue for contamination and as a result reject the product and contribute to the problem of food waste. Milne (2013) calls for better consumer understanding of this labelling system. According to these previous findings, the following hypothesis is stated:

H2: The product being close to its expiration date (as opposed to far from its expiration date) has a negative effect on the product utility

2.5. Price

It is widely known that price and purchase intention are in general negatively related: when charging a higher price, less is sold. Empirical studies show that price is an important factor in choosing a retail food store and that price is just as important as the value perception of the products sold in a supermarket to consumers (Arnold et al., 1983; Zielke, 2010). Also, price is one of the most important types of information when one shops for grocery products (Lappalainen & Gibney, 1998; Lichtenstein et al., 1993). Koutsimanis et al. (2012) even found that price was the most important feature for consumers when making a purchase decision regarding fresh produce.

(17)

16

reductions are an effective strategy to increase the purchase of healthful foods. Hansen & Sørensen (1993) studied the effects of lower prices on the demand for organic food. Their results showed a raise of demand for organic food products in Denmark to one of the highest in the world. Not only for organic products, but also for suboptimal products, a price decrease has a positive effect. Aschemann-Witzel et al. (2018) show that suboptimal foods with a reduced price are well accepted and that the price reduction can encourage the consumer to buy a suboptimal product instead of an optimal product. Consumers paying more attention to the price of the product are more likely to buy the product with a reduced price (Aschemann-Witzel, 2018).

Although the substantial amount of research stating the positive effect of a price reduction, there is also proof of price being positively related to purchase intention. Price has been found to be a major indicator of quality when intrinsic quality cues are missing or when price is the only cue available (Brunsø et al., 2002; Zeithaml, 1988). This means that a higher price indicates a higher expected quality, which would mean that the customer would prefer a higher priced product. However, oftentimes many quality cues are available, and studies taking these into account mostly find a diminishing relationship between price and expected quality (Brunsø et al., 2002). As there are more quality cues investigated in this study than price, i.e. the appearance, expiration date and organic labeling, it is expected that the price in this case will not have a positive effect on the product choice and that there exists a negative relationship between the price and the product utility. Therefore, it is expected that the product utility increases when the price is reduced.

Besides the effect of decreasing the price, Inman et al. (1991) showed that adding a promotion signal to the price cut makes the consumer more likely to choose the product. When a promotional signal is added, some consumers do not even look at the price anymore, but are triggered only by the promotional signal. Hence, it is expected that a lower price accompanied by a signal has a stronger positive effect on the product utility than the lower price solely. The following hypotheses have been developed accordingly:

(18)

17

H3b: The product having a low price (as opposed to having a regular price) complemented by a promotion signal has a positive effect on the product utility

H3c: The product having a low price complemented by a promotion signal has a stronger positive effect on the product utility than the low price solely

2.6. Organic Label

Several studies have looked into the effect of an organic label on food products. Organic food includes food items free from artificial chemicals such as fertilizers, herbicides, pesticides, antibiotics, genetically modified organisms and irradiation (Gad Mohsen & Dacko, 2013). This has an effect on the consumer perception of the organic product. Because organic products use fewer pesticides and artificial fertilizers, they are believed to be less harmful to the environment (Cornelissen et al., 2008). Moreover, these products are seen as natural, local, fresh and pure. Giesen & De Hooge (2019) found a positive effect of authenticity claims on consumers’ purchase intentions and quality perceptions of suboptimal products. This authenticity positioning highlights the product’s genuineness, origin, or naturalness, making it comparable to the organic labeling discussed before.

Furthermore, organic products are considered to be healthy, because no synthetic chemicals are used in the production (Chan, 2001; Suprapto & Wijaya, 2012). Prada et al. (2017) proved this by showing that a whole food product (e.g. lettuce) with an organic label is perceived as more healthful, tastier and less caloric. Also Sirieix et al. (2013) showed that for many consumers, organic food means good food. The fact that organic products are perceived as more healthful is an important factor for choosing organic food products, and it has been proven that consumers are willing to pay more for a product when it is organic (Rana & Paul, 2017). Because of the fact that organic products do not use any pesticides, the consumer believes that the product is risk-free, which results in an increase of the willingness to pay for organic products (Rana & Paul, 2017; Tsakiridou et al., 2008).

(19)

18

Doorn & Verhoef (2011) claim that organic food products are offered for an average price premium of 30 to 60 percent, causing the consumer to show reluctance toward buying these products and explaining the low market shares of organic products. This could mean that an organic label has a negative effect on the purchase intention of the consumer. However, in this study the prices are given to the consumer and they are not raised when the product is organic. Therefore, there is no negative effect of the organic claim as a result of the premium price the product usually comes with. Also, with regard to studies investigating the effect of an organic label on suboptimal products, there has not always been found a significant effect (Aschemann-Witzel, 2018; Loebnitz et al., 2015). Hence, the organic label is not expected to cause any negative effect on the choice of the consumer, but is believed to positively influence their product choice. The following hypothesis is formed:

H4: An organic label presented on the product has a positive effect on the product utility

2.7. Sustainability Claim

Van Giesen & De Hooge (2019) define sustainability positioning as ‘the positioning of suboptimal products in a way that highlights their environmental sustainability’. Examples of environmentally focused sustainability claims are addressing the carbon footprint or food miles, or adding a sticker which reads ‘Fight foodwaste’ on the product (Aschemann-Witzel, 2018; Van Loo et al., 2014).

(20)

19

suboptimal product. As a result, the negative effect of an unusual appearance and a close expiration date will be diminished, causing customers to choose the suboptimal product more often. In case consumers are confronted with a sustainability claim, they will act less rejecting toward suboptimal products and the appearance and expiration date will have a smaller effect on the product utility.

However, Aschemann-Witzel (2018) and Sirieix et al. (2013) did not find an effect of communicating the food waste avoidance aspect on suboptimal products. This could be due to several reasons.

Firstly, behavior is hard to change. The fact that past behavior is a strong indicator for future behavior is a matter which is supported by much empirical evidence (Ajzen, 2011). Given the fact that consumers tend to choose the optimal product in terms of looks and expiration date, changing this status quo is a challenge. Habit is one of the most powerful predictors of eating behavior and when buying products in a habitual way, like food products, people use little information in their decision making and behave based on repeated actions (Stranieri et al., 2017; Van ’t Riet et al., 2011).

Secondly, Sirieix et al. (2013) stated that consumers tend to be skeptical about claims stating products being ‘climate friendly’. They found that consumers did not always trust the claim, though it is also possible that they do not have trust in their ability to solve environmental problems. Despite the environmental claim emphasizing the negative consequences of not buying the suboptimal food, consumers may find themselves feeling unable to impact the issue of sustainability, resulting in less environmentally friendly behavior (Berger & Corbin, 1992).

(21)

20

performed behavior increases the probability that others perform that behavior (Cialdini, 2003). In case of a sustainability claim shown at the suboptimal product, this means that emphasizing the negative effects of not choosing the suboptimal product confirms that other consumers reject the product. As a result, the claim does not increase the likelihood of purchasing the suboptimal product.

There are theories supporting and contradicting the effectiveness of a sustainability claim on suboptimal foods. However, since the majority of articles investigating the effect of a sustainability claim on suboptimal food products found a significant effect, the following hypotheses are proposed:

H5a: The negative effect of an unusual appearance on the product utility is less pronounced in case a sustainability claim is shown

H5b: The negative effect of a close expiration date on the product utility is less pronounced in case a sustainability claim is shown

2.8. Control Variables

Furthermore, several control variables will be taken into account to see to what extent the results are generalizable. There will be tested whether there exist differences between the product types or consumers with regard to the effect sizes or importance of product features.

2.8.1. Product Type

(22)

21

compared to a bent cucumber, dented carton of juice, or broken biscuits. However, they call for better understanding of the way consumers respond to different types of products and the types of suboptimality. To see whether there exists any difference among these different product types, and because these food categories are often investigated, fresh foods and packaged goods will both be taken into account in this study. Besides, since there is a trend of meal kits coming up lately (Distrifood, 2018), this product category will also be taken into the analysis. Meal kits contain ingredients to cook a meal, they are pre-proportioned and packaged and cause for less food waste since they contain the right proportion of food products (Heard et al., 2019). There has not been any research about suboptimal products in this product category, so studying this category can lead to new insights.

2.8.2. Consumer Characteristics

Several studies have shown that consumer characteristics influence the product choices of consumers and the acceptance of suboptimal products. Demographic characteristics such as age and education are found to affect consumers’ product choices (Aschemann-Witzel et al., 2018; Loebnitz et al., 2015). Also, awareness of food waste and pro-environmental identity are often called indicators for choosing a product (Loebnitz et al., 2015). Stangherlin et al. (2019) stated that when grocery shopping, consumers strive either for choosing the optimal, perfect product, or they are more accepting toward suboptimal products because of their environmental concern. The influence of environmental factors is also investigated by De Hooge et al. (2017), who showed that consumers with a higher commitment to environmental sustainability, or environmentally friendly behavior, showed a higher preference for suboptimal products. Because of the possible relevance of these factors, several consumer characteristics will be taken into account to see whether results are generalizable, or whether there exist differences between consumer groups. Information about consumers’ demographics, environmental concern and sustainable behavior will therefore be taken into account.

(23)

(24)

23

3. Method

In this chapter the methodology used in the study will be explained. The survey consisted of two parts: in the first part participants indicated their preferred product for three types of products in a choice-based study, and in the second part information about the participant was collected. Before the study started, participants were assigned to one of the conditions regarding the sustainability claim. In this chapter, first the data collection will be discussed. Secondly, the attributes and their levels used in the choice study will be specified, after which the experimental design will be explained. Lastly, the scales used to measure consumer characteristics will be discussed.

3.1. Data Collection

In order to gain insight into the influences of different product features on the purchase decision of the consumer, data was collected through an online consumer survey. The survey was developed using the software tool preferencelab.com and was spread mainly through social media. The survey was shared in Facebook groups, on LinkedIn and in WhatsApp groups. Data was collected over a period of 10 days and participants took on average 10.7 minutes to complete the survey.

3.2. Attributes and Levels

In the choice study, consumers indicated their preferred products among several alternatives shown. The products differed in their appearance, expiration date, price and organic labeling. The attributes and corresponding levels will be discussed accordingly. Before the study took place, a pretest was conducted to check whether the levels of the appearance and expiration date of each product were perceived the way they were intended.

3.2.1. Appearance

(25)

24

The pictures of the apple in the study were adopted from the study of Loebnitz et al. (2015). Loebnitz et al. (2015) demonstrated that for appearance to make a difference in consumers’ purchase intentions, the product needs to deviate extremely from the norm; moderate abnormalities do not have an effect. However, from the pretest, in which the perception of the appearance of the products was tested, the extremely abnormal apple was perceived significantly different from the unusual version of the crackers and the meal kits. This is interesting since in the pictures of the meal kits, also the extremely abnormal product variants of Loebnitz et al. (2015) were used. To prevent differences in perception, the moderately abnormal version of the apple was used instead of the extremely abnormal version in order to equalize the perception of the unusual products.

For the crackers, a picture was made from a package of ‘Bolletje Oerknäck’ in a usual condition and in a damaged condition, where it was visually evident that the package was deformed. From the pretest was concluded that the usual and unusual version were perceived significantly different from each other. For the meal kit, a private label of Jumbo’s Mexican Burritos was used and the private label logo was removed. The meal kit was then edited so it looked like it was filled with either usual looking vegetables or unusual looking vegetables. Pictures of vegetables were used from the study of Loebnitz et al. (2015). The normal and extremely abnormal variants of the eggplant and the carrot from the study were used, and a usual and deformed image of a paprika were used. For this product, the usual and unusual variants were also perceived significantly different in the pretest. The images of the products used are shown in Figure 3.

(26)

25

3.2.2. Expiration Date

In the study, the expiration date was given as one of the attributes of the product. It was shown below the pictures of the products and could take on the levels of close and far. For the apple and the meal kit, the expiration date was either ‘in 1 day’ (close) or ‘in 7 days’ (far). From the pretest was concluded that these were perceived significantly different from each other. However, the pretest also revealed that for the crackers, the expiration date being in 7 days was not significantly different from the expiration date in 1 day. Besides, the expiration date being in 1 day was perceived as shorter for the crackers than for the apple and meal kit. This led to adjustment of the expiration date levels of the crackers: the close expiration date was set to ‘in 3 days’ and the far expiration date to ‘in 2 months’. These values were also closer to the real-life expiration date of the product in stores.

3.2.3. Price

Price took on the levels of regular, lower or reduced in this study. Prices were kept as realistic as possible, adopting the prices from the website of the Jumbo, the second largest supermarket chain in The Netherlands, for the regular price level. Apples were normally sold for 2.19 euros per kilo on average, where a kilo consisted of 5 to 7 apples. Assuming there are 6 apples in a kilo, the regular price for one apple would be about € 0.35. The value of 35 cents was therefore set as the regular price for an apple. Crackers had a regular price of € 1.77 and the Mexican Burrito meal kit had a regular price of € 4.69.

For the lower priced level, the regular prices were lowered by 35%. The reduction of 35% was chosen because Albert Heijn, the largest supermarket chain in The Netherlands, uses this percentage to increase the sales of products with a close expiration date (Albert Heijn, n.d.-b). The lower price levels for the apple, crackers and Mexican burrito meal kit were therefore set to € 0.23, € 1.15 and € 3.05 respectively.

(27)

26

reduction and sometimes only showing a lower price for their suboptimal products, this seems relevant. The price level was, like the expiration date, shown below the images of the products of the choice set.

3.2.4. Organic Label

The organic label was retrieved from the supermarket chain Jumbo since this logo appeared to be clearer than other organic logos, since it showed not only a green logo, but also the word ‘biologisch’ (meaning ‘organic’) appeared on it. The name of the supermarket chain was removed and replaced by the word ‘organic’, so the meaning of the organic label would be clear in both Dutch and English. The organic label attribute consisted of two levels: the logo was present or the logo was not present. The organic label was, like the expiration date and price, placed below the picture of the product as an attribute. An image of the organic label used in the study is shown in Figure 4.

Figure 4. Organic label used in the study

3.3. Experimental Design Choice Study

In this part, the design of the choice study will be discussed. The use of conjoint analysis will be explained and the assignment of consumers to one of the conditions regarding the sustainability claim will be illustrated.

3.3.1. Conjoint Analysis

(28)

27

that consumers make natural choices representing their preferences, and that they make trade-offs between different products and their characteristics (Eggers et al., 2016).

Attributes in the study consisted of the appearance of the product, the expiration date, the price, and organic labeling. This total of four attributes lies within the boundaries of the advised maximum number of attributes for a conjoint study (Eggers & Sattler, 2011). Each attribute existed of two or three levels and random levels of the stimuli were taken to determine the choice sets. The attribute levels were mutually exclusive. There was chosen for minimal overlap between the product alternatives in order to create an efficient choice design (Huber & Zwerina, 1996). Each choice set consisted of three product alternatives, plus a ‘none-option’ with the statement: ‘With these options, I would go to a different supermarket to find a better offer’. The none-option was added to strengthen the realistic nature of the experiment, since the participant in a real-life scenario is not forced to buy the product at the supermarket. Figure 5 shows an example of a choice set which could be presented to the participant. Since the attribute levels are randomized for each choice set, the choice sets differed between participants.

Figure 5. Example of a choice set

(29)

28

number of choice sets to use is equal to the total number of levels across all attributes minus the number of attributes plus one. In this study the minimum number of choice sets would therefore be 6. In this study’s choice task, participants indicated their preferred product among the product alternatives for a total of 9 choice sets. This, together with the different products being presented, prevented the participant from becoming tired or bored from the repeated task and enables them to also complete the remaining part of the survey. Besides, a lot of information could be retrieved from the amount of choice sets. Three different products were used in the study to control for the effect of the product type on consumer choices. Participants gave their preferences for an apple, crackers and a Mexican burrito meal kit, representing the different product types of fresh produce, packaged products and meal kits respectively. The first three choice sets showed apples, the second three choice sets showed crackers, and the last three choice sets showed Mexican burrito meal kits.

3.3.2. Assignment to Sustainability Condition

(30)

29

3.4. Consumer Characteristics

After the first part of the study, which consisted of the choice task, participants started the second part of the study. This part focused on the consumer characteristics of the participant. Information about participants’ demographics, environmental concern and sustainable behavior were measured. The scales and measurement of these consumer characteristics will be explained accordingly. Part two of the survey as presented to the participants is shown in Appendix B.

3.4.1. Demographics

The demographics section collected information about the descriptive statistics of the participant, including the participant’s gender, age, household size, nationality, education and occupation.

3.4.2. Environmental Concern

Environmental concern was measured using a scale adopted from Urban & Ščasný (2012). Participants were asked how concerned they were with various environmental problems such as waste generation, air pollution and climate change. The items were rated on a five-point scale ranging from 1 = ‘Not concerned’ to 5 = ‘Very concerned’.

3.4.3. Sustainable Behavior

For the sustainable behavior, a scale by Diekmann & Preisendörfer (2003) was used. Participants were asked how often they performed certain forms of sustainable behavior, rating items like ‘recycling paper’, ‘buying refill bottles’ and ‘using energy saving light bulbs in

the household’. Participants filled in the answers using a four-point scale: never – sometimes

(31)

30

3.5. Manipulation Check

(32)

31

4. Results

In this chapter, the results of the study will be discussed. Firstly, the results of the manipulation check will be analyzed and the sample characteristics will be described. Secondly, the effects and importance of the food product attributes will be examined, followed by the effect of the sustainability claim. Lastly, the generalizability between product types and the results of the latent class analysis will be explained.

4.1. Manipulation Check

To check whether the manipulations in the study were interpreted the way they were supposed to by the participants, a manipulation check was done. All the usual versions of the products were evaluated on average between ‘moderately usual’ to ‘very usual’ on a seven-point scale from 1 = ‘Very unusual’ to 7 = ‘Very usual’ (apple M = 6.67, SD = 0.91; crackers M = 6.52, SD = 1.03; meal kit M = 6.35, SD = 1.19). The unusual variant of the apple was on average found to be ‘Slightly unusual’ (M = 3.04, SD = 1.50) and the unusual variant of the crackers (M = 3.99, SD = 1.74) and meal kit (M = 4.09, SD = 1.83) was on average found to be ‘Neutral’. A t-test was performed to compare the usual version to the unusual version for each of the products. These tests showed that the appearances significantly (p < 0.001) differed from each other for each product type, so that the usual versions were perceived more usual than the unusual variants of the products.

(33)

32

the levels taken for the appearances and expiration dates for each product were perceived the way they were intended.

4.2. Sample Characteristics

The survey was completed by 193 respondents, who filled in a total of 1737 choice sets plus the questions about consumer characteristics. From these respondents, 13 were eliminated due to the fact that they did not notice the different appearances in some or all of the product choice sets. Two of those respondents indicated that this was caused by filling in the survey on their mobile phones, on which the pictures appeared very small. The final sample consisted of 180 respondents, filling in a total of 1620 choice sets.

In the study, 59 males (32.8%) and 120 females (66.6%) participated and one person identified as other (0.5%). Since the survey was mostly distributed among fellow students, friends and relatives, the sample is relatively young. The average age of participants was 25.9 years (min. = 15, max. = 69), 106 of the participants (58.9%) was a student and 67 (37.2%) were employed and most of the respondents were Dutch (93.3%). Table 1 gives a summary of the demographics of the sample. Participants were evenly distributed among the sustainability claim conditions. The sustainability claim condition was randomized among respondents and made sure that there was an even distribution where 91 respondents (50.6%) got the instruction text without the sustainability claim and 89 respondents (49.4%) got the instructions with the sustainability claim.

Table 1. Sample demographics

Variable Level Frequency Percentage

Gender Male Female Other 59 120 1 32.8% 66.6% 0.5% Age Min. = 15 Mean = 25.85 Median = 24 Max. = 69 Household size Single person

(34)

33 Other - 5 persons - More than 5 12 5 6.8% 2.8% Nationality Dutch Other 168 12 93.3% 6.7% Education No formal qualification

Secondary education (Middelbare school) Secondary vocational education (MBO/MAVO) College (HBO)

Undergraduate degree (e.g. BSc) Graduate degree (e.g. MSc) Doctoral degree (e.g. PhD) Other 2 43 19 30 48 37 1 - 1.1% 23.9% 10.5% 16.7% 26.7% 20.5% 0.6% Occupation Student Employed Unemployed Retired 106 67 6 1 58.9% 37.2% 3.3% 0.6%

4.2 Choice-Based Conjoint Analysis

4.2.1. Conjoint Analysis with Main Effects

To test for the effect of the product attributes on the product utility, a part-worth model was made, since the attributes used in the study are nominal. A part-worth model assumes that goods and services are combinations of attributes, to which the consumer attaches part-worth utilities. The utility V of consumer n for product i is the sum of the part-worth utilities and can be expressed by Formula 1, where k is the number of attributes, x is a dummy indicating the specific attribute level of product i, and β is the part-worth utility of consumer n for attribute

k.

Formula 1. Part-worth utility of a product

(35)

34

data well. The LL of the NULL model is -2245.8, while the LL of Model 1 is -1293.4. This means the estimated model is a better fit than the NULL model (p < 0.001). A summary of Model 1 can be found in Table 2.

From the results, the estimates and the importance of each attribute can be derived and the reference levels are recovered. The outcomes show that all attributes have a significant effect on the product choice. Consumers are more likely to choose a product with a usual appearance (β = 0.2286, p < 0.001) and less likely to choose a product with an unusual appearance (β = -0.2286, p < 0.001), which confirms H1. The expiration date being far away has a positive effect on the choice of the consumer (β = 0.7642, p < 0.001), while the expiration date being close has a negative effect on the product choice (β = -0.7642, p < 0.001). This confirms H2. The coefficients of the price attributes reveal that consumers prefer a lower price (β = 0.447, p < 0.001) or a reduced price (β = 0.8123, p < 0.001) over a regular priced product (β = -1.2592, p < 0.001). This confirms H3a and H3b. To compare the reduced price level to the lower price level, the difference between these betas was divided by the pooled standard error, which results in a z- and p-value. The p-value is significant which means that consumers significantly prefer the reduced price over the lower price (β = 0.365, p < 0.001), confirming H3c. Further, consumers are more likely to choose a product with an organic label (β = 0.365, p < 0.001) than a product without an organic label (β = -0.365, p < 0.001), confirming H4. A visual representation of the attribute betas can be found in Figure 6. The utilities of each attribute level are shown, where positive utilities mean that the level is preferred by consumers.

(36)

35

The relative importance RI of product attribute i can be derived by taking the maximum difference between part-worth utilities called the utility range UR within an attribute, and dividing it by the sum of the maximum difference of all attributes j. The relative importance was expressed in percentage weight and was calculated with use of the equation shown in Formula 2.

Formula 2. Relative importance of attributes

This means that for this model, consumers implicitly assign an importance of 8.73% to the appearance, 29.2% to the expiration date, 48.12% to price and 13.95% to organic labeling. Price is by far the most important attribute, followed by the expiration date. An organic label is found of less importance, and the appearance of the product is the least important of the four attributes. A visual representation of the attribute importance is given in Figure 7.

Figure 7. Attribute importance in percentage weight

Table 2. Results of Model 1

Coefficients Estimate St. Error Z-value P-value Sig. Appearance

(37)

36 Unusual -0.2286 0.0338 -6.7573 0.0000 *** Expiration date Far 0.7642 0.0372 20.5614 0.0000 *** Close -0.7642 0.0372 -20.5614 0.0000 *** Price Regular -1.2592 0.0638 -19.7386 0.0000 *** Lower 0.447 0.045 9.9487 0.0000 *** Reduced 0.8123 0.0447 18.1723 0.0000 *** Organic label Yes 0.365 0.034 10.7364 0.0000 *** No -0.365 0.034 -10.7364 0.0000 *** None Option -2.651 0.2076 -12.7707 0.0000 *** p < 0.001 ‘***’, p < 0.01 ‘**’, p < 0.05 ‘*’, p < 0.1 ‘.’ LL = -1293.4, R2_{= 0.424, AIC = 2598.82}

4.2.2. Conjoint Analysis with Interaction Effect of Sustainability Claim

Model 1 can be extended with interaction effects of the sustainability claim to find out whether the claim has an effect on the attributes indicating suboptimality. Model 2 includes interaction effects of the sustainability claim with the features which indicate suboptimality (appearance and expiration date). The model shows a better fit than Model 1 (LL = -1289.3, p = 0.016). The McFadden R2_{is 0.426 and the adjusted R}2_{is 0.423, which are slightly higher than} for Model 1, meaning that this model better fits the data. Besides, the Akaike information criterion shows a lower value (AIC Model 1 = 2598.82, AIC Model 2 = 2594.58), also confirming a better fit.

(38)

37

slightly more (β = 0.0685, p = 0.049) when a sustainability claim is included. These findings confirm the expectation that the rejection of close expiration dates can be diminished using a sustainability claim, which confirms H5b. The effect of the claim on the utility of the appearance and expiration has been visualized in Figure 8 and results of the model can be seen in Table 3. Besides the interaction of a sustainability claim with the appearance and expiration date, the interaction with the remaining attributes has also been checked. These interactions did not show any significant results.

Figure 8. Utilities of the attributes Appearance and Expiration date with interaction of Sustainability Claim

Table 3. Results of Model 2 (SC = Sustainability Claim)

Coefficients Estimate St. Error z-value p-value Sig. Appearance Usual 0.2256 0.0339 6.6491 0.0000 *** Unusual -0.2256 0.0339 -6.6491 0.0000 *** Expiration date Far 0.7673 0.0373 20.5602 0.0000 *** Close -0.7673 0.0373 -20.5602 0.0000 *** Price Regular -1.2641 0.064 -19.7523 0.0000 *** Lower 0.4511 0.0451 10.0124 0.0000 *** Reduced 0.813 0.0448 18.1473 0.0000 *** Organic label Yes 0.367 0.0341 10.7636 0.0000 *** 0,23 -0,23 0,28 -0,28 -0,40 -0,30 -0,20 -0,10 0,00 0,10 0,20 0,30 0,40 U S U A L U N U S U A L

APPEARANCE

Without sustainability claim With sustainability claim

0,77 -0,77 0,70 -0,70 -1,00 -0,80 -0,60 -0,40 -0,20 0,00 0,20 0,40 0,60 0,80 1,00 F A R C L O S E

EXPIRATION DATE

(39)

38

No -0.367 0.0341 10.7636 0.0000 ***

Sustainability effect int.

SC * App Usual 0.0592 0.0336 1.7627 0.0780 . SC * App Unusual -0.0592 0.0336 -1.7627 0.0780 . SC * Exp Long -0.0685 0.0348 -1.919 0.0486 * SC * Exp Short 0.0685 0.0348 1.919 0.0486 * None Option -2.651 0.2076 -12.7707 0.0000 *** p < 0.001 ‘***’, p < 0.01 ‘**’, p < 0.05 ‘*’, p < 0.1 ‘.’ LL = -1289.3, R2_{= 0.426, AIC = 2594.58}

4.3. Generalizability among Product Types

Next, a model is generated to measure whether the different product types influence the attribute effects. Model 3 includes the main effects of all attributes, complemented by interaction effects of the product types with all attributes. The McFadden R2_{of Model 3 is} 0.451 and the adjusted R2_{is 0.443. These values show that a bigger part of the data is} explained by this model than by the previous models. The AIC value of Model 3 is 2303.33, which is a lower value than the AIC of Model 2 and also indicates that this model better explains the data. A summary of the outcomes of Model 3 can be found in Appendix C.

(40)

39

to other products. They acted less rejecting toward a close expiration date (β = 0.2614, p < 0.001) and disliked the regular price more (β = -0.3486, p < 0.001) in case of choosing between meal kits. Figure 9 shows a visualization of the attribute importance of each product type.

p < 0.001 ‘***’, p < 0.01 ‘**’, p < 0.05 ‘*’, p < 0.1 ‘.’ Figure 9. Attribute importance per product type

4.4. Consumer Segmentation

In the following part, it is investigated whether consumer characteristics play a role in the choice behavior of consumers, and whether consumers can be divided into segments to help explain their behavior. First, the models used are clarified and subsequently the results are discussed.

4.4.1. Model Description

(41)

40

a sustainability claim with (sub)optimal attributes and the interaction effects of the product type with all attributes. A summary of the outcomes of Model 4 is represented in Appendix E.

Next to the latent class analysis, also several simpler models have been made to check for the individual impact of each consumer characteristic. Model 1 was used as a basis and was in turn extended with interaction effects of one of the consumer characteristics with all attributes. For instance, for the model where the effect of gender was investigated, the main effects of the appearance, expiration date, price and organic labeling were inserted and complemented by the interaction effects of gender with every attribute. A comparison of each extended model with Model 1 can be found in Appendix D.

4.4.2. Results Latent Class Analysis

The results of the latent class analysis reveal the existence of two classes, with relative class sizes of 7.8% for Class 1 and 92.2% for Class 2. Consumers in Class 1 tend to choose a product with a usual appearance (β = 1.6772, p < 0.001) and with a lower (β = 0.6077, p = 0.003) or reduced (β = 0.7654, p < 0.001) price instead of a regular price (β = -1.3731, p < 0.001). They tend to prefer a product without an organic label (β = 0.9909, p < 0.001), this effect is even marginally stronger when the product concerns crackers (β = 0.4567, p < 0.091). The time to expiration has no significant effect for consumers in Class 1 (β = -0.1543, p = 0.711). When the consumer is exposed to a sustainability claim, the product having a usual appearance is found less important (β = -0.8271, p = 0.005). The appearance (43.19%) of the product is most important to consumers in Class 1, followed by the price (27.51%). They prefer the absence of an organic label (25.45%). This relatively small segment can therefore be identified as non-organic and cosmetically perfect product seeking consumers.

(42)

41

sustainability claim, they seem to assign less importance to the expiration date (β = -0.2169, p < 0.001), which means that the claim has the desired effect. However, interestingly the claim also causes consumers in this class to assign importance to the appearance of the product where they did not care about before, resulting in a preference for a usual appearance (β = 0.2004, p < 0.001). The expiration date (41.68%) of the product is most important to consumers in Class 2, followed by the price (39.35%) and the presence of an organic label (17.11%). These consumers can be recognized by their preference for products with a far expiration date and an organic label. The main attribute effects for Class 1 and Class 2 are visualized in Figure 10 and the importance of the features per class is shown in Figure 11.

Consumers in Class 2 tend to act differently toward product features depending on the product type. For the apple, a usual appearance was found to be more important (β = 0.1991, p < 0.001) and the expiration date was found of less importance (β = -0.2662, p < 0.001). Consumers were also less price sensitive when the product concerned an apple, which could be caused by the lower price levels of the apple. For the crackers, consumers found the expiration date being far away even more important (β = 0.5833, p < 0.001).

(43)

42

ᴼ = ‘not significant’ Figure 11. Attribute importance per class in percentage weight

When looking at the consumer characteristics, which are summarized in Appendix F, it can be seen that consumers in Class 2 are – in comparison to Class 1 – more likely to be female (β = 0.6249, p = 0.006) and are slightly younger (β = -0.0228, p = 0.039). They have obtained a higher education level (β = 0.2952, p < 0.001) and a smaller household size (β = -0.3879, p < 0.001). Consumers in Class 2 seem to be more environmentally concerned (β = 1.3581, p < 0.001). The fact that this class prefers products with an organic label and is more willing to buy a product with a close expiration date when exposed to a sustainability claim is therefore not surprising. However, despite their environmental concern they perform less sustainable behavior than Class 1 (β = -0.7239, p = 0.008).

4.4.3. Results Interaction Models

(44)

43

(45)

44

5. Discussion

It has been proven that suboptimal features of a product have a negative influence on the product choice of the consumer (Milne, 2013; Parfitt et al., 2010). However, what their relative importance is with regard to other product characteristics, and whether a sustainability claim influences this relationship did not receive much interest or consensus. This chapter will return to the research questions of this paper and will address the findings and the forthcoming managerial implications. Table 4 shows the overview of the outcomes of the hypotheses proposed in the study.

Table 4. Results of hypotheses analysis

Hypothesis Results

H1: The product having an unusual appearance (as opposed to having a usual

appearance) has a negative effect on the product utility

Supported

H2: The product being close to its expiration date (as opposed to far from its

expiration date) has a negative effect on the product utility

Supported

H3a: The product having a low price (as opposed to having a regular price) has a

positive effect on the product utility

H3b: The product having a low price (as opposed to having a regular price)

complemented by a promotion signal has a positive effect on the product utility

H3c: The product having a low price complemented by a promotion signal has a

stronger positive effect on the product utility than the low price solely

Supported

H4: An organic label presented on the product has a positive effect on the product

utility

Supported

H5a: The negative effect of an unusual appearance on the product utility is less

pronounced in case a sustainability claim is shown

H5b: The negative effect of a close expiration date on the product utility is less

pronounced in case a sustainability claim is shown

Rejected

Supported

5.1. Main Effect of Product Features

(46)

45

important feature for consumers when making a purchase decision. The expiration date (29.20%) was the second most important feature in their choice-decision, for which consumers prefer a farther expiration date. The presence of an organic label (13.95%) was preferred over the absence of such a label, although it was not found of great importance. The appearance (8.73%) of the product was the least important for consumers.

With regard to the features indicating suboptimality, this study finds that consumers indeed prefer a usual appearance over an unusual appearance, which is in line with previous research. However, in contrast to the statements by Feenstra (1988) and Marx-Pienaar & Erasmus (2014), this does not seem to be the main feature consumers pay attention to when making a choice decision. This could be caused by the fact that aforementioned studies did not quantify the price of the products and drew superior attention toward the appearance of the products. Milne (2013), Hall-Phillips & Shah (2017) and Newsome et al. (2014) found that a product with a farther expiration date is preferred over a product with a close expiration date. This study’s results are in line with these findings and proves that the expiration date is an important feature when deciding which product to buy. This seems logical as date labels are a mean to protect the health of consumers from unseen microbiological threats (Milne, 2013). However, the problem of food waste has gained awareness over the last years and new labeling systems were introduced to raise the understanding of perishability if products (Newsome et al., 2014). This could result in more environmentally friendly behavior by consumers and make the expiration date becomes less important in their product choice.

5.2. Effect of a Sustainability Claim

(47)

46

the outcomes of this study with previous studies investigating the effect of a sustainability claim on suboptimal products. Van Giesen & De Hooge (2019), Aschemann-Witzel et al. (2018) and Loebnitz et al. (2015) used unusual appearance to identify products as suboptimal, and found a significant effect of a sustainability claim. However, in this study the claim does not diminish the rejection toward unusual looking food. The study of Aschemann-Witzel (2018) used the expiration date as a cue for suboptimality and they did not find any effect of a sustainability claim. However, this study finds a significant effect of a sustainability claim on a product with a close expiration date.

This study shows that a sustainability claim can make products with a close expiration date slightly more attractive to consumers. Loebnitz et al. (2015) stated that this could be caused by increasing environmental awareness, which results in an increased purchase intention toward suboptimal food products. In case there was presented a product with an unusual appearance, a sustainability claim did not have a positive effect. In contrast to the expectations, the sustainability claim slightly increased the negative effect of the appearance on the product utility. One possible explanation could be that due to the claim, consumers paid more attention to the appearance of the product where they would otherwise not pay attention to. As a result, they could have assigned more importance to that feature. Another reason could be that consumers were confirmed about the suboptimality of the product and acted accordingly. Cialdini (2003) proposed that the sustainability claim confirms consumers that the product is rejected by others, and due to the fact that consumers like to conform to others, they reject the product as well.

5.1. Generalizability among Product Types and Consumers

(48)

47

prefers the absence of an organic label. As Aschemann-Witzel et al. (2018) and Loebnitz et al. (2015) proposed, demographic characteristics such as gender and education level influence these product choices. This study has shown that consumers who are environmentally concerned are more likely to choose an organic product, and are more prone to choose a product with a close expiration date when seeing a sustainability claim. This confirms the theory of Loebnitz et al. (2015) that environmental factors indicate product choice.

5.4. Managerial Implications

Managers could use the findings of this study to improve their policy and adapt their marketing strategy in a way to make suboptimal food products more attractive to the consumer, in order to contribute to the environment. This study quantified the relative importance of different product features in the choice decision of the consumer, and showed that price is the most important feature. This means that a product with a close expiration date or unusual appearance can be most effectively made more attractive by offering it for a lower or reduced price. The emphasis on the price reduction has a stronger effect than stating the lower price solely, which implies the preference for a price reduction strategy using a promotional signal. The study found that overall, the expiration date is of bigger importance to the consumer than the appearance, which means that the price reduction would have an even stronger effect on the suboptimal product in terms of its appearance. When the company wants to sell more suboptimal products, showing an organic label can also advance the sales.

(49)

48

When adapting the marketing strategy to increase the sales of suboptimal food product, managers could take into account the product category of the product they try to sell, and the target customer of the product. Actions to improve sales by for instance introducing a price reduction, could have a different impact depending on the product category and their different importance of features. Not only do attribute effects differ among product types, they also vary among consumers. Consumers can be segmented into two groups based on their product preferences. While the majority of consumers mainly assigns importance to the expiration date and values an organic label, the other smaller group takes into account the appearance and dislikes an organic label. Managers should know the preferences of the consumers they want to target, before developing their strategy.

5.5. Limitations and Further Research

This research has several limitations. Firstly, with regard to the sample of the study, the survey was mainly filled in by young and Dutch people. Previous research has revealed that younger people and consumers from The Netherlands or Norway are more prone to buy suboptimal food products, this could have influenced the results of this study (De Hooge et al., 2017). Future research could investigate whether the results stay the same in case of a more diverse sample in terms of age and nationality.

Secondly, the levels of expiration date and price could have an influence on the outcomes of this study. The levels were adapted to the product presented to the participants; however, the deviating levels between products could have an influence on the behavior of the consumer. This effect has not been taken into account in this study and it is therefore hard to determine the actual effect of these features. Future research could take price levels and expiration dates which are equal across product types and adapt the products to prevent unrealistic prices or expiration dates for the products.

Acceptance of Suboptimal Food Products: A Choice-Based Conjoint Analysis on the Importance of Appearance, Expiration Date, Price and Organic Labeling

Acceptance of Suboptimal Food Products:

A Choice-Based Conjoint Analysis on the Importance of Appearance,

Expiration Date, Price and Organic Labeling

Acceptance of Suboptimal Food Products:

A Choice-Based Conjoint Analysis on the Importance of Appearance,

Expiration Date, Price and Organic Labeling

Abstract

Preface

Contents

1. Introduction

1.1. Food Waste Issue

1.2. Research Objectives and Problem Statement

1.3. Contribution and Managerial Implications

2. Theoretical Background

2.1. Suboptimal Food Products

2.2. Consumer Choices using Product Utility

2.3. Appearance

2.4. Expiration Date

2.5. Price

2.6. Organic Label

2.7. Sustainability Claim

2.8. Control Variables

3. Method

3.1. Data Collection

3.2. Attributes and Levels

3.3. Experimental Design Choice Study

3.4. Consumer Characteristics

3.5. Manipulation Check

4. Results

4.1. Manipulation Check

4.2. Sample Characteristics

4.2 Choice-Based Conjoint Analysis

APPEARANCE

EXPIRATION DATE

4.3. Generalizability among Product Types

4.4. Consumer Segmentation

5. Discussion

5.1. Main Effect of Product Features

5.2. Effect of a Sustainability Claim

5.1. Generalizability among Product Types and Consumers

5.4. Managerial Implications

5.5. Limitations and Further Research