An exploration of the effect of convenience for customers when returning products on customer retention for web shops

(1)

1

An Exploration of the Effect of Convenience for

Customers when Returning Products on

Customer Retention for Web Shops

Céline van Baaren Student number: 10250328

Master Thesis

MSc in Business Studies – Strategy Track Date: November 2014

Supervisor: Jaap de Wit Second Supervisor: Erik Dirksen

(2)

2

Abstract

This thesis looks at how convenience experienced by customers when they returned a product influences customer loyalty for web shops. It also examines whether factors such as product type and customer characteristics affect the convenience experienced by customers and how this impacts customer loyalty. A distinction is made between attitudinal and behavioral customer loyalty. A quantitative approach was employed by using a web-based questionnaire to collect data and various statistical calculations were made to analyze the data using SPSS and Andy Hayes’s PROCESS model. Findings show that convenience experienced by customers when they return a product does not directly influence customer loyalty. Only through customer satisfaction with the product return does convenience have a small effect on attitudinal customer loyalty. Another revelation of the findings is that customer characteristics may be more important to consider in forming product return policy strategies than product type or return method. These findings are relevant because they inform whether a convenient return policy strategy actually has the intended effect of customer retention. Knowing this can help managers make a trade-off between the benefits of using convenient product return strategies and the cost dimension that comes with the returning of products, and base their return policy strategy on this trade-off.

(3)

3

1. Introduction

The type of return policy a web shop applies can play a role in customer retention for the web shop. Web shops can attract and retain customers by having strategic return policies in place

(Foscht, Ernstreiter, III, Sinha, & Swoboda, 2013; Jiang & Rosenbloom, 2005; Krikke, Hofenk, & Wang, 2013; Ling (Alice) Jiang, 2013; Mollenkopf, Rabinovich, Laseter, & Boyer, 2007; Ramanathan, 2011). Certain types of return policies can increase product return transactions and as a result increase the costs associated with the product returns for the web shop. It is therefore relevant to examine if the benefits of customer retention, resulting from specific product return policies, outweigh the costs associated with product returns. The aim of the thesis is to inspect the customer retention benefit side of this weigh-off, and research to what extent a specific return policy actually has the intended effect of attracting and retaining customers. This aim is built on the suggestions of Jiang & Rosenbloom (2005) and Ramanathan (2011) to further research the effect of customer retention in the post purchase process.

Creating and retaining a large and loyal customer base is important for web shops because this can be a significant factor for a web shop’s success. Many researchers link customer retention with profitability such as Anderson, Fornell, & Lehmann (1994), Bowen & Chen (2001), Gummesson (1994). Profitability can be viewed as a measure of success; the more profitable a firm the more successful. The more customers that are retained, increases the chances that a firm is profitable and successful. In other words, customer retention is beneficial for web

Case in Point:

Various media sources have spread the rumor that Zalando, a large online web shop that mainly sells clothing, has a return rate that is higher than 50%. Yet, Zalando continues to offer very generous return policies such as free-shipping, 100 days right-of-return policy and a complimentary service hotline which is in place to help customers return their products. This raises many questions. Why does Zalando allow such high return percentages and why do their policies continue to make returning products so easy? An interesting fact is that Zalando is a very profitable web shop, with a reported turnover of 1 billion euros. Do generous return policies have benefits for the web shop that outweigh the associated costs with product returns? What is the trade-off that Zalando makes? Does having generous return policies help retain customers? Do the benefits of customer retention outweigh the costs that are associated with product returns? And do generous return policies have the intended effect of customer retention and the financial benefits that come with that?

(6)

6

shops. Customer retention is often achieved through a positive experience for the customer on its

purchase and post-purchase (Ling (Alice) Jiang, 2013; Ramanathan, 2011). Web shops can

employ different strategies to give their customers this positive product return experience.

One return policy strategy is to make it convenient for the customer to return a product to the web shop. Convenience for the purpose of this thesis is defined in terms of time and effort that is acquired to carry out an action (Berry, Seiders, & Grewal, 2002; Brown, 1990); the more it acquires time and effort for the customer, the less convenience is experienced by the customer and vice versa. Web shops can design their return policy in different ways so that they can influence the convenience experience for customers to different extents.

The research question will therefore be: How does convenience in the customer product

return experience affect customer loyalty for web shops? If convenient return policies do not

have the intended effect of customer retention, there might be reason for web shops to rethink the convenience of their return policy and the associated costs. The exploration of the relationship of product return policies and customer retention can contribute to the knowledge gap about convenience and product returns and provide a direction for further research in the product return and customer retention domain. Having more understanding of this relationship can on a managerial level help make more informed (strategic) decisions about the extent that web shops should make their return policies convenient for the customer.

The rest of the thesis is structured in the following way. In chapter 2 relevant literature and concepts are discussed, and the hypotheses are presented. Chapter 3 explains the research and analytical approach. In chapter 4 the results are presented, followed by chapter 5 in which the results are discussed. Lastly, in chapter 6 the research question is answered and conclusions are drawn about the findings.

(7)

7

2. Literature Review

In this chapter relevant literature is discussed and concepts are explained. First, in paragraph 2.1, return policies as a strategy to retain customers are discussed in more detail. In paragraphs 2.2 and 2.3 the concepts of customer experience and convenience are explained. Paragraph 2.4, elaborates on the question how customer characteristics and product types can influence the convenience that is experienced by customers when products are returned. The concepts of customer retention and customer loyalty are considered and defined in paragraph 2.5 and finally in paragraph 2.6 the hypotheses that will be examined in this thesis are presented.

2.1 Return policy as a Strategy to Retain Customers

The return process of a web shop is a direct way in which a web shop can contact its customers. Customers need to interact with the web shop in order to be able to return an earlier ordered product. This specific interaction opens up a customer experience window that web shops can exploit strategically to retain customers. For example, Mollenkopf et al. (2007) found that interacting with customers over product returns can be an opportunity for web shops to build customer loyalty and can be a service recovery opportunity. A service recovery opportunity is the chance for web shops to appease the customer and restore or improve a situation in which the customer was not satisfied. If the customer is not satisfied with the product he receives, a web shop can appease the customer by allowing the customer to return the product for example. Web shops can steer the product return interaction through the return policies that they maintain.

Previous research indicates that web shops apply a variety of return policies which are mainly based on product type. Davis, Hagerty, & Gerstner (1998) find that the characteristics of the product determine what strategy is used by firms in combination with cost considerations. For example, they suggest firms can use a more convenient return policy for products that have the opportunity for cross-selling, can obtain a high salvage value for the returned product and for products which cannot be consumed within a shorter amount of time than other products who do not share the same characteristics (Davis et al., 1998). Similarly, Yu & Wang (2008) find that linking different return policies to the type of product and customer characteristics can help firms increase sales and reduce returns. These findings explain why firms use different policies based

(8)

8 on one-time interactions with customers, but they do not explore how these policies can be used in that window of interaction to further create customer retention.

Mollenkopf et al. (2007) do look at how firms can impact loyalty intentions of customers through their product return (policy) management. The results of their empirical research show a positive relationship between customer satisfaction with the return experience and their loyalty intentions. They also find that customer effort when returning a product is negatively associated with customer satisfaction. These results emphasize that the customer experience when returning a product has an effect on customer loyalty. To investigate the effect of customer experience when returning products on customer loyalty further, this thesis will look at how a factor such as convenience in the product return experience impacts customer retention.

2.2 Customer Experience as a Strategy to Retain Customers

The importance of providing a good customer experience has been stressed by research. It can affect customer satisfaction (Liljander & Strandvik, 1997), deliver customer loyalty (Mascarenhas, Kesavan, & Bernacchi, 2006) as well as influence expectations (Johnson & Mathews, 1997). Although these findings are not specifically about web shops, customer experiences may also play a role in determining whether customers come back to purchase at a web shop or not. For this reason it is important for web shops who want to attract and retain customers to provide good experiences for customers who return products.

Customer experience in this thesis is defined as “the internal and subjective response that customers have to any direct or indirect contact with the company” (Meyer & Schwager, 2007, p. 118). Customer experience in those terms can be considered as a psychological construct (Rose, Clark, Samouel, & Hair, 2012, p. 309). It is the personal interpretation of the customer of his interaction with the web shop that determines how the customer feels about the experience. The judgments of the experiences will result into customer intentions (Johnston & Kong, 2011). The experience may affect the intention of the customer to repurchase or not and these intentions may then turn into actions (Johnston & Kong, 2011, p. 8). In other words, a good return experience may lead customers to repurchase something again from the web shop.

(9)

9 2.3 Convenience as a Strategy to Retain Customers

One strategy for web shops to retain customers through product returns is to design their return policies in such a way that returning a product becomes convenient for the customer. Berry et al. (2002) argue that consumer’s perceptions of convenience of a service also directly affects the experience and satisfaction of that specific encounter or experience. As discussed in the previous section, customer experience and the satisfaction with that experience can affect customer loyalty. It can, therefore, be argued that the convenience experienced when returning a product, affects the customer experience and satisfaction which can increase customer loyalty. Ling (Alice) Jiang, (2013) finds that convenient online experiences can increase customer loyalty. Although the main focus of this study was not specifically on the return channel, they did find evidence that a convenient return experience may have an effect on customer loyalty.

Convenience, however, is a contextual concept. What is the essence of convenience and how can this be defined? Literature on retailing has suggested two primary factors of importance in delivering convenient service to customers: time-saving and effort minimization (Ling (Alice) Jiang, 2013). In the literature the term convenience mainly signifies the time and effort consumers need to acquire a product rather than the actual characteristics or attributes of a product (Brown, 1990). Researchers viewed convenience as an attribute that reduced the nonmonetary price of a product, such as time and energy that consumers give up to buy goods and services (Berry et al., 2002). Time can be seen as nonrenewable and effort as depletable. There is thus an opportunity cost to convenience. For this reason convenience can be viewed as an important consideration for consumers in their purchasing decisions. It follows that providing convenience to customers can be strategically important for customer retention.

A distinction can be made between the convenience of a product and the convenience of a service. The product characteristics and assets can be convenient, but also the service surrounding a product can be convenient. Since product return policies and processes can be seen as a service provided by a firm for its customers, service convenience mainly refers to the speed and ease of shopping (Ling (Alice) Jiang, 2013). In this thesis service convenience in the product return process for web shops will be examined in terms of time and effort. The speed (time) and ease (effort) of returning products can be viewed in terms of how much effort is spent on the the product handling and the time spent in order to return a product.

(10)

10 The effort it takes to return a product and the impact it can have on the return satisfaction has already been researched by Mollenkopf et al. (2007). They found evidence that customer efforts can be negatively associated with return satisfaction. They have not incorporated the time element with effort, so it can be argued that the concept of convenience, as defined in this thesis, has not been examined by Mollenkopf et al. (2007). This thesis will examine the relationship further by adding the time element.

2.4 The Effect of Different Products and Different Customer Characteristics on Convenience Experienced when Returning a Product

There are factors that can influence the way convenience is being experienced, such as different types of products or customer characteristics (Mollenkopf et al., 2007; Ramanathan, 2010, 2011; Yu & Wang, 2008). For example, it can be much less convenient for a customer to take a couch to a drop off point because it takes more effort than a pair of shoes. Similarly, it might be more convenient for a customer that has more ‘free’ time to return a product, than a customer that is employed full-time and does not have a lot of free time to spend on returning products. The type of return method that is used to return a product can vary in convenience depending on the type of product. For example, it is much easier to get the couch picked up by a return service than having to bring it to a drop-off point, or return a book by postal mail than having to wait at home all day for the mail man. Customer characteristics can also influence the return experience. For a car owner it might be more convenient to return a couch than for someone that does not own a car. There are many more ways in which certain factors can influence the way convenience is experienced when a product is returned.

Previous research has looked into the way that products can influence the return experience of web shops. Ramanathan (2011) examined how both different risk characteristics of products and the way product returns are handled, play an important role in shaping customer loyalty. Ramanathan (2011) categorized the risk characteristics of a product based on price and ambiguity of the product. The higher the price of the product and the less certain the customer is about what it can expect from the product, the higher the risk perception of the product. Ramanathan (2011) finds that the ease of returning a product and the product type (in this case, the type was determined by the risk characteristics) have an influence on the way the product return is experienced, which helps shaping customer loyalty.

(11)

11 One relevant, even though not directly related, study that has been carried out by Weltevreden (2008) can inform us about the convenience of a return method. The research examines the consequences of pick-up and delivery points and what this means for retailers, shopping centers and mobility. He finds that pickup and delivery centers are mainly used for returning online orders in the Netherlands. He also found that 5 minute drive distance was critical to the success of the concept of a pickup and delivery point. Although this research is not about customer loyalty or convenience directly, it can inform us about the following things. It examines one return method used by web shops. Secondly, it finds that time is important for customers to return their product; it was important that the pick-up and delivery point was no further than a 5 minute drive. Thirdly, it shows that people prefer to drive to the drop off point. This could hypothetically be because it takes less effort than other transportation efforts or is just the quickest method. This return method is considered convenient if it does not have a driving distance for more than 5 minutes. It shows that the dimensions of return methods can have an influence on the experience of the customer.

Other research suggests researching product and customer characteristics further in relation to product returns and customer loyalty. For example, Mollenkopf et al. (2007) suggest that future research could examine different return services based on customer segmentation and customer prioritization. They reason that returns may be more frequent for some customers than others, is because customers can have different expectations of return services and customers may generally have different loyalty levels. Similarly, Yu & Wang (2008) suggest that firms should classify their customers in order to fit them with the right return policies.

2.5 Customer Retention and Customer Loyalty

Customer retention and customer loyalty are concepts that are usually used interchangeably in the literature. However, the use of these terms need to be clarified because they do not mean the same. Customer retention, in this thesis, is defined as the activity that a web shop undertakes in order to attract and keep customers coming to the web shop and prevent customers from going to other web shops. Customer loyalty is a more complicated concept. Customer loyalty, according to Bowen & Chen (2001), has a behavioral and attitudinal dimension to it. Behavioral customer loyalty is that customer repeatedly shop at the same shop. Attitudinal customer loyalty is measured emotionally and psychologically. The intention of buying again at the same web shop

(12)

12 and telling other people positive things about the web shop, are examples of attitudinal customer loyalty. Attitudinal and behavioral customer loyalty should be treated as different constructs because customers can make repeat purchases but are not psychologically or emotionally loyal, and customers can be emotionally and psychologically loyal but not make repeat purchases. For this reason a distinction is made between the two concepts and also treated as such in the hypotheses.

Both attitudinal and behavioral customer loyalty can have financial benefits (Bowen & Chen, 2001) for web shops. Reduced marketing costs, increased sales and reduced operational costs are, according to Bowen & Chen (2001), amongst the benefits of customer loyalty that can increase profit. More benefits from loyal customers according to Reichheld & Sasser Jr. (1990) are that loyal customers are less likely to switch for reasons such as price and make more purchases. It can therefore be financially beneficial for web shops to create a loyal customer base, that outweigh the costs of allowing high percentages of product returns.

2.6 Hypotheses

In order to investigate the research question, ‘How does convenience in the customer product return experience affect customer loyalty for web shops?’ different hypotheses are tested. The

first four hypotheses address whether convenience experienced by the customer during the return process affect customer retention. In other words, will customers come back to the web shop if they experienced convenience while returning a product? A distinction in the hypotheses is made between the direct effect of convenience experienced on customer retention and an indirect effect of convenience experienced on customer retention through customer satisfaction with the return process. Mollenkopf et al. (2007) have found that customer satisfaction with returning a product has a positive effect on customer retention. It could be possible that convenience experienced will not directly relate to customer retention, but that it will affect customer satisfaction with the return process which in turn affects customer retention. Separating these effects will give more insight into the role convenience exactly plays in the return process and the influence it has on customer retention.

Another distinction that is made in the first four hypotheses is whether customers have the intention to stay loyal to the web shop or whether customers actually repurchased something at the web shop after having returned a product. The intention of staying loyal to the web shop does

(13)

13

not necessarily mean that customers will actually purchase again (Johnston & Kong, 2011, p. 8)

at a web shop. The indirect aim of this thesis is to inspect whether the benefits of customer retention can outweigh the costs dimension of product returns. It seems logical that the desired effect of convenient return policies is to get customers to repurchase, so it will balance out the cost aspect of product returns. Loyalty intentions without an actual repurchase may not balance out the costs because there is no monetary transaction. Making a distinction between the two variables will give more insight in the actual effect of convenient return policies and customer retention.

The last four hypotheses address whether factors, such as product type, return method and customer characteristics, influence return convenience for the customer. The latter four hypotheses are supposed to give more insight into how convenience is impacted by product and customer characteristics. This can give more details about the relationship between convenience experienced in the return process and customer retention. As mentioned in paragraph 2.4, there are many factors that can influence the convenience experienced. Due to the scope of the thesis, the factors have been limited to four: product type, return method, car owners versus non-car owners and full time employed versus not full-time employed. The last two factors were chosen because they touch on the time and effort dimension of convenience.

The eight hypotheses are stated below:

H1: The more convenient the return experience is for the customer, the more likely the customer will have more intention to stay loyal to the web shop.

H2: The more convenient the return experience is for the customer, the more likely the customer will purchase again at the web shop.

H3: The more convenient the return experience is for the customer, the more likely the customer is satisfied and the customer will have more intention to stay loyal to the web shop.

H4: The more convenient the return experience is for the customer, the more likely the customer is satisfied and the more likely the customer will purchase again at the web shop.

(14)

14 H5: The type of product that is returned to the web shop moderates how convenient a product return is experienced by the customer.

H6: The return method that the customer uses will moderate how convenient a product return is experienced by the customer.

H7: Customers that own a car will experience a product return as more convenient, than customers that do not own a car.

H8: Customers that have a full-time job will experience a product return as less convenient, than customers that do not have a full-time job.

(15)

15

3. Methodological Approach

In this chapter the research approach to answer the research question and test the hypotheses, is explained. First, in paragraph 3.1, a description of the research instruments is given. Second, in paragraph 3.2, the target population and the sample are described. Thirdly, in paragraph 3.3, a description of the sample is given. Finally, in paragraph 3.4, the analytical approach is explained.

3.1 Description of Research Instruments

In order to answer the research question and to test the eight hypotheses, a cross-sectional study was carried out, using a self-administered, internet mediated questionnaire. The questionnaire was implemented using Qualtrics1 and was distributed online through various internet channels, such as Facebook and mail, using the snowball effect and self-selection. Via Facebook and e-mail it was asked if people could like and share the questionnaire amongst their acquaintances, in order to spread the questionnaire out amongst respondents that selected themselves.

The questionnaire was constructed mostly out of self-administered questions, but also partly based on questions carried out in previous research. For the variable constructs of ‘customer loyalty intentions’ and ‘customer return satisfaction’, the questions and scales of Mollenkopf et al. (2007) were used. In their research they developed new scales and modified existing scales contributing to the expansion of knowledge in the area of service offerings in return management. The questions and scales for these constructs have been proven to be valid and for that reason chosen to be used. Additionally the research is also more comparable to that of Mollenkopf et al. (2007) and other related research because for a few constructs the same questions and measures have been used. The questionnaire included 32 questions and 290 items and can be consulted in the appendix.

The questionnaire consisted of two parts. The first part concerned questions regarding demographic characteristics, such as gender, age and occupation, in order to create a typology for the customer segment. The second part of the questionnaire consisted of questions about product returns from web shops and convenience. All items, except for the demographic ones, are measured on a five-point Likert scale anchored at strongly disagree (1) and strongly agree (5) or other similar statements. For multiple choice questions, open answers were included so that

(16)

16 answers not included in the options, would not be excluded, in order to reduce bias. The items of the questionnaire can be viewed in the appendix.

3.2 Description of Sample

Target population

The target research population consists of everyone that is able to buy products from a web shop, including people that do not buy from web shops. Problems with returning products may be a reason for not buying from an online web shop. The target population is therefore adults that have access to a computer with internet and are able to receive packages at their home.

Generalizability of the Sample

The survey received a total of 127 responses with a dropout rate of 10%. In total the research sample used consisted of 120 respondents, corresponding to an approximately 95% response rate. The sample is adequate to do statistical analysis with, but not large enough to effectively generalize for a whole population. In order to be fully generalizable for the population of the Netherlands at a 95% confidence level, .5 standard deviation and a margin of error (confidence interval) of +/- 5% a sample size is required of N=3852. For now, this exploratory research can give an indication and a direction for future research, but should take account that there can be bias in the results because the sample was not large enough.

Sample Characteristics

Table 1 shows the demographic characteristics of the sample. The target population were people that owned a computer irrespective whether they did or did not order anything from a web shop or returned anything to a web shop. There are no data bases that directly give this information to which the sample characteristics can be compared to. For this reason the sample population is compared to the numbers of the general Dutch population because it is the most standardized.

2

Based on calculation from Scott smith, retrieved from the website http://www.qualtrics.com/blog/determining-sample-size/ .

(17)

17 Compared to the numbers of the CBS3 population numbers, the male/female ratio in the sample is highly skewed (see Table 1). According to the CBS the male/female ratio should be 1:1 (50%:50%). The age characteristics of the sample also do not correspond with the numbers of the CBS. According to the figures of the CBS in the Netherlands currently about 25% of the population is between 20 and 40 years old, about 35% of the population is between 40 and 60 years old, and about 17% of the population is between 60 and 80 years old. If you compare these figures to Table 1, the numbers are not similar. In the sample there is a larger group of people in their twenties and thirties than people in their forties’ and fifties, while CBS numbers state that the other way around should be the case.

The difference in population characteristics compared to the figures of the CBS can be explained by the small response rate that results in to having a smaller sample size than is needed to have a representative sample. The cause of this unrepresentative sample is likely due to the method of how the respondents were recruited. What this means for the collected sample is that it not representative for the general Dutch population. Even though the sample is not representative, the results of the sample can give a direction for further research.

Table 1. Sample Characteristics

Gender Location Car Owner

Male 30% Urban 80% Yes 56%

Female 70% Rural 20% No 44%

Age Occupation

Twenties 55% Student 38%

Thirties 5% Full-time job 24% Forties 9.3% Part-time job 18%

Fifties 24.1% No Job 6%

Sixties 3.3% Pension 5%

Seventies 2.5% Unemployed 8%

Eighties 0.8%

3_{CBS is a Dutch central bureau of statistics that keeps track of Dutch demographic statistics. Retrieved data from}

http://statline.cbs.nl/StatWeb/publication/?VW=T&DM=SLNL&PA=37296ned&D1=a&D2=0,10,20,30,40,50,60,(l-1),l&HD=130605-0924&HDR=G1&STB=T on 1st of August 2014.

(18)

18 3.3 Description of Data

The data consisted of information provided by 120 respondents and out of 131 variables. There were 3 cases that had some missing values in the data set. The missing values for these cases were filled in using mean substitution. The missing values for variables where more than two categorical options could be selected were not replaced. It could not be determined if there was a pattern for these specific categorical variables and therefore filling in missing values would increase the chances of bias data. When using those specific variables in the analyses, the three cases were left out.

After the missing values were replaced, some variables and items needed to be transformed to get ready for analysis. First dichotomous variables were recoded with 0 for one category and 1 for the other, so that correlations could be calculated using these variables. Other variables were recoded to reduce the number of categories and make some of the variables more clear. For example, age was recoded into decades.

3.4 Description of Analytical Approach

Creating constructs

In order to have clear variables that could be used to test the hypotheses, three constructs were made to create three new variables ‘customer loyalty intentions’, ‘customer satisfaction experienced with returning the product’, and ‘customer convenience experienced’. The variable ‘customer convenience experienced’ was created based on three items. The items asked to what extent they experienced mental effort, physical effort and time, of the last product they returned. These three questions sum up the definition of convenience applied in this thesis; time and effort. The variable ‘customer loyalty intentions’ is the same construct as the construct used by Mollenkopf et al. (2007). This construct consisted of four items that asked whether the respondents intend to shop again at the web shop where they returned a product; would recommend the web shop to other people after returning the package; whether the web shop remained their first choice to buy a similar product after returning the package; and if they would rather go to another web shop for a similar product after returning the package to the web shop. The variable ‘customer satisfaction experienced with returning the product’ was also constructed with the same items as Mollenkopf et al.(2007) used. The items included four questions that

(19)

19 asked how satisfied the respondents were with the return process, the return options, the amount of time in which they could return the product and the reimbursement.

Before creating the constructs two steps were carried out. First, counter indicative items were transformed so that the variables would point in the same direction and not contradict each other when creating the construct. This was done by recoding the items with counter indicative values. Secondly, before creating the new variables the reliability of the items were checked by using reliability analysis and Pearson correlation. The reliability tests for ‘customer loyalty intentions’ items showed a positive correlation at a significance level of 0.01 (2-tailed). All inter-item correlations were above 0.3 which showed that the inter-items have reliable scales (Field, 2009). The overall Cronbach’s α = .77. According to Kline (2000) values of Cronbach’s α in the range of .7 to .8 probably indicated good reliability. The result shows that the items are reliable to use in the construct. The reliability test for the items of ‘customer convenience experienced when returning a product’ and the ‘customer satisfaction experienced after returning a product’ showed that the item correlations were above .3 and Cronbach’s α above .8. These results also indicate that the items are reliable to use for the construct. After these tests, the constructs were computed into three new variables using the mean of the sum of the items - the items were added up together, and then the mean was calculated.

Analyzing Hypothesis 1 and 2

Hypotheses 1 and 2 are:

H1: The more convenient the return experience is for the customer, the more likely the customer will have intentions to stay loyal to the web shop.

H2: The more convenient the return experience is for the customer, the more likely the customer will purchase again at the web shop.

In order to test whether the variables in the hypotheses affect each other, a Pearson correlation is calculated for the variables in each hypothesis. The variables that are tested for correlation, for H1, are ‘the customer convenience experienced when returning a product’ and ‘customer loyalty intentions’. The variables that are tested for correlation for H2 are ‘the customer convenience experienced when returning a product’ and ‘whether the customer will actually purchase again at

(20)

20 the web shop’. If there is a correlation between the variables, it means that there is an indication that the variables affect each other, however nothing specific can be concluded about causality. The outcome of the correlation test show the strength and the direction of the effects of the variables on each other.

Analyzing Hypothesis 3 and 4

Hypotheses 3 and 4 are the following:

H3: The more convenient the return experience is for the customer, the more likely the customer is satisfied and the customer will have more intention to stay loyal to the web shop.

H4: The more convenient the return experience is for the customer, the more likely the customer is satisfied and the more likely the customer will purchase again at the web shop.

The relationship that is being measured for both hypotheses is depicted in conceptual Model 1. The model tests whether there is an indirect effect of independent variable X on dependent variable Y through mediating variable Mi = aibi. Second, the model tests if there is a direct effect

of independent variable X on dependent variable Y = ci. Finally, the total effect of the model is

calculated. The sum of the indirect effect and the direct effect is the total effect of the model; Total effect = c1 + aibi.. The total effect of the model can help explain how the independent

variable X influences dependent variable Y.

In order to calculate this model the PROCESS command in SPPS created by Andrew Hayes was used. PROCESS is a command in SPSS that is specifically designed to calculate and test many different types of moderation, mediation, mediating moderation or moderating mediation relationships. The computation implements moderation or mediation analysis or a combination of both into one integrated conditional process mode (Hayes, 2013). In other words, this procedure calculates formula that would normally take several steps in SPSS in just one step. To use the PROCESS command various arguments must be provided, which include a list of variables in the model, the role the variables play in the model (independent variable, dependent variable, moderators etc.), and the number of the conceptual model4 that is tested. This procedure

4_{Andrew Hayes’s (2013) PROCESS procedure provides up to at least 78 different conceptual model templates that}

can be used to compute different kinds of moderating, mediating, moderating mediating, and mediating moderating relationships. When trying to compute a relationship, the conceptual model that is tested should be inserted as one of

(21)

21 has an advantage over using regular regression analysis because it does not require extra calculations that are not automatically carried out by regression analysis (Hayes, 2013), hence the choice of this procedure.

The model that is tested for the two hypotheses that is tested is Andrew Hayes model 4. This model was chosen because it depicts the relationships that are being tested for H3 and H4. The models are depicted in Figure 1. Does, for H3, the independent variable X ‘customer convenience experienced’ have a direct and indirect effect on dependent variable Y ‘the intention to stay loyal to the web shop’, through mediating variable M ‘the customer satisfaction experienced with the product return’? Similarly for H4, does the independent variable X ‘customer convenience experienced’, have a direct and/or indirect effect on dependent variable Y

the arguments to be able to make the computation. The type of model that is used is indicated by a number, previously defined by Hayes. The model templates and corresponding numbers can be found online on the following website: http://www.afhayes.com/public/templates.pdf

Model 1. Andrew Hayes Model 4

(22)

22 ‘the likelihood that the customer will actually purchase again at the web shop’, through mediating variable M ‘the customer satisfaction experienced with the product return’. Separately for each hypothesis the variables and the model number were inserted in the PROCESS computation program in order to compute the results in one computation. For the analysis of the computations a 95% confidence level, .5 standard deviation and a margin of error (confidence interval) of +/- 5% is maintained.

Analyzing hypotheses 5 and 6

H5: The type of product that is returned to the web shop moderates how convenient a product return is experienced by the customer.

H6: The return method that the customer uses will moderate how convenient a product return is experienced by the customer.

The relationship that is tested for both hypotheses is depicted in conceptual Model 2. Both hypotheses test whether there is a stronger or weaker effect of independent variable X on dependent variable Y under a certain condition M. This model is calculated with the following equation: Y = bi+b3M.

In order to compute this model the PROCESS computation command by Andrew Hayes is used. The model number that was inserted in the computation is Andrew Hayes Model 1. The models for H5 and H6 are depicted in Figure 2. For H5 the independent variable X is ‘the return method used by the customer’, the dependent variable Y is customer convenience experienced when returning a product and the moderating variable M is the type of product that is returned. For H6 the independent variable X is the type of product that is returned, the dependent variable Y is ‘the customer convenience experienced when returning a product’ and the moderating variable M is ‘the return method used by the customer’. For each hypothesis separately the variables and the model number were inserted in the PROCESS computation program in order to compute the results in one computation. Further analysis for the hypothesis was done by calculating Pearson Correlations for the variables, return method, product type and convenience experienced, to see if and how the variables affected each other. For the analysis of the

(23)

23 computations a 95% confidence level, .5 standard deviation and a margin of error (confidence interval) of +/- 5% is maintained.

Model 2. Andrew Hayes Model 2 Figure 2. Conceptual Models H5 and H6

Hypotheses 7 and 8

H7: Customers that own a car will experience a product return as more convenient, than customers that do not own a car.

H8: Customers that have a full-time job will experience a product return as less convenient, than customers that do not have a full-time job.

Hypotheses 7 and 8 are about comparing groups. The hypotheses do not look for moderating or mediating effects, therefore it is not suitable to use the PROCESS calculation by Andrew Hayes, which is designed only to calculate moderating or mediating effects. Instead, to test hypothesis 7 an independent t-test and a Mann-Whitney U test are calculated and to test hypothesis 8, a one-way ANOVA is calculated and a post hoc test (Tukey HSD) is carried out.

(24)

24 Hypothesis 7 means to test whether car owners experience more convenience when they return a product than non-car owners. In this hypothesis there is one outcome variable with a continuous outcome, customer convenience experienced, one predictor variable with 2 predictor categories, car owner versus non-car owner and different participants are used for each group, making the independent t-test the most suitable test for this hypothesis. A Mann- Whitney U test is carried out to double check the results of the independent t-test and in case the parametric assumptions are not met.

Hypothesis 8 means to test if people with different occupations will experience differences in convenience, more specifically if there is a difference between respondents with a full time job and respondents that do not have a full time job. In this situation more than two groups are compared, making the t-test not suitable to test this hypothesis. Instead one way independent ANOVA was the best way to test the hypothesis because there are more than two predictor categories and each group has different participants. Also post hoc test, Tukey HSD was carried out to give more information about the differences between the groups. The effect size was calculated in order to see how much of the variance between the groups is explained.

(25)

25

4. Results

This chapter shows the results from the data analysis. The results for each hypothesis are presented in separated paragraphs, except for hypothesis 5 and hypothesis 6. The results of hypothesis 5 and 6 are so that they are best presented together. The output of the statistical data analysis for each hypothesis can be referred to in the appendix if they are not presented in more detail in this chapter. The interpretation of the results are discussed in chapter 5.

4. 1 Hypothesis 1

For hypothesis 1, a Pearson Correlation was calculated to see if the variables ‘customer convenience experienced’ when returning a product and ‘customer loyalty intentions’ to the web shop affected each other. The results of the correlation computation can be viewed in Table 2. The outcome of the computation shows that there is no significant affect between the variables ‘consumer convenience experienced’ and the ‘customer loyalty intentions’ to the web shop. H1 is rejected.

4.2 Hypothesis 2

For hypothesis 2, a Pearson Correlation was calculated to see if the variables customer convenience experienced when returning a product and if customers purchased something again from a web shop affected each other. The results of the computation, which can be consulted in Table 2, show that the variables (customer convenience experienced and repurchased from web shop) do not have a significant effect on each other. H2 is rejected.

Table 2. Correlation Matrix H1 and H2

Correlation Matrix H1 & H2

Variable Mean SD 1 2 3

1. Costumer Convenience Experienced 3.92 0.85 _

2. Customer Loyalty Intentions 4.16 0.68 .22 _

3. Repurchased from Web Shop 0.37 0.49 -.14 -.31** _

*Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).

(26)

26 4.3 Hypothesis 3

For hypothesis 3, different relationships were computed. The first relationship is whether customer convenience experienced when returning a product has influence on customer satisfaction experienced when returning a product. The results show that this is a significant relationship, = .28, p< 0.01. The results also show that customer convenience experienced when returning a product has no direct effect on customer loyalty intentions, β= .41, p> 0.05, but that customer satisfaction experienced when returning a product does have a significant effect on customer loyalty intentions, β= .06, p< 0.01. The variable M, customer satisfaction experienced when returning a product, does have a significant mediating effect between customer convenience experienced and customer loyalty intentions, β=.12, LLCI = .03, ULCI = .28. It is a significant indirect effect. However, the total effect of the model is not significant β= .17, P>0.05. H3 can therefore only partially be accepted. Convenience when returning a product does influence the satisfaction with the return experience, which in turn increases customer loyalty intentions. However, the effect of convenience when returning a product is not strong enough to influence customer loyalty intentions directly or indirectly via customer satisfaction with the return process. A graphic presentation of the relationships is presented in figure 3.

For hypothesis 4, three relationships were tested. The first relationship was if customer convenience experienced when returning a product has a direct effect that a customer will purchase again at a web shop. The second relationship was, if there is an indirect effect of customer convenience experienced when returning a product and the likelihood that customers will purchase again at a web shop, through the mediating variable customer satisfaction experienced when returning the product. The third relationship that was tested was the direct effect of customer convenience experienced when returning a product and customer satisfaction experienced when returning the product.

The results show that there is evidence that there is relationship between dependent variable X ‘customer convenience experienced’ and mediating variable M ‘customer satisfaction experienced’; β= .28, p< 0.01. However there is no significant effect of mediating variable M ‘customer satisfaction experienced when returning a product’ on the likelihood that a customer

(27)

27 will purchase again from the web shop where they returned the product β= .12, p> 0.05 (β=.03, LLCI = -.27, ULCI = .39). There is no significant effect of independent variable X ‘customer convenience experienced’ on dependent variable Y ‘the likelihood that a customer will purchase again from a web shop where they returned a product’; β= .38, p>0.05. The total effect of the model is not significant β= -.35, p>0.05. A graphic presentation of the results is depicted in figure 3.

The results show that there is no direct effect of customer convenience experienced when returning a product and the likelihood that the customer will purchase something again from the web shop where they returned the product. There is neither an indirect effect of convenience experienced when returning a product on the likelihood that the customer will purchase something again from the web shop where they returned the product, through the satisfaction of the return experience. The total effect of the model was not significant. Given these results, H4 is rejected.

(28)

28 4.5 Hypothesis 5 & Hypothesis 6

Hypothesis 5concerns the question whether the type of product had a moderating effect on the return method used and the convenience that was experienced when a product was returned. The results show that both return method and product type individually do not have a significant effect on the convenience experienced, product p>.05 and return method p>0.05. The interaction of the two variables with each other also did not have a significant effect on customer convenience experienced, p>0.05. The conditional effect of the type of product on the relationship between return method and customer convenience experienced neither showed a significant effect, p>0.05.

Hypothesis 6 looked at whether the return method had a moderating effect on the relationship between the product type and the convenience experienced when returning a product to a web shop. The results are similar to those of hypothesis 5. Both variables return method and product type individually do not have a significant effect on the convenience experienced, product p>.05 and return method p>0.05 and the interaction of the two variables with each other also did not have a significant effect on the convenience experienced, p>0.05. Likewise, the conditional effects of return method on the relationship between product type and convenience experienced are not significant, p>0.05.

Further analysis of the two hypotheses was done by calculating Pearson correlation for the variables, product type, return method, and customer convenience experienced when returning a product, in order to see how the variables may affect each other. The calculations are shown in Table 3. The outcome of the correlation calculations show that the variables product type and return method do affect each other significantly but that both variables do not affect the variable of customer convenience experienced. This means that models H5 and H6 are not relevant because neither the independent nor moderating variables affect the outcome variable. Hypothesis 5 and hypothesis 6 are rejected.

(29)

29 Table 3. Correlation Matrix H5 and H6

Correlation Matrix H5 & H6

Variable Mean SD 1 2 3

1. Costumer Convenience Experienced 3.92 0.85 _

2. Product Type 3.60 1.10 -.19 _

3. Return Method 1.79 0.47 -.19 .32** _

*Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed).

Hypothesis 7 compares the convenience experienced by car owners and respondents that do not own a car when they returned a product. To test hypothesis 7, an independent t-test and an independent sample Mann-Whitney U test were calculated. The results of the independent t-test show that there are no significant statistical differences between respondents that own a car versus respondents that do not own a car and the convenience experienced when returning a product, t=-1.41, p>0.05. The results of the Mann-Whitney U test was also not significant, p>0.05, and advices to retain the null-hypothesis. Accordingly, hypothesis 7 is rejected. There are no significant statistical differences between car owners and non-car owners and the convenience the two groups experience when returning a product.

Hypothesis 8 compares whether respondents with a full-time job experienced returning product to a web shop than respondents that did not have a full time job. In order to test this hypothesis a 1 way ANOVA calculation and post hoc test Tukey HSD were carried out. The one way ANOVA calculation shows that there is a significant difference in means of respondents with different occupations, F=2.39, p<0.05 and had an effect size of 0.15. This means that about 15% of the difference in convenience experienced is determined by occupation. When testing further which groups significantly differ from each other, using post hoc test Tukey HSD, results show that there is a significant difference in mean between people that have a full-time job and people that are not employed, p<0.05 with a mean difference of 1.4. There are no significant differences in

(30)

30 mean between other groups. When comparing the mean of full time employed respondents, x̅=3.57, and respondents that are not employed, x̅=4.71, the mean is higher for respondents that are not employed. This means that respondents without a job generally have a more convenient experience returning a product than respondents that have a full time job. Hypothesis H8 can be accepted. Respondents that have a full time job have an experience of less convenience when returning a product than respondents that are unemployed.

(31)

31

5. Discussion

In this chapter the results are discussed and put into context of theoretical and managerial implications. Paragraph 5.1 discusses the results and theoretical implications followed by further research suggestions. The managerial implications are discussed in paragraph 5.2. Lastly, paragraph 5.3 pays attention to the research design of the thesis and discuss the impact it may have had on the results.

5.1 Theoretical Implications and Further Research

Convenience and its Effect on Customer Retention

The first four hypotheses in this thesis deal with the question whether convenience, experienced by the customer when returning a product, directly and indirectly effects customer retention. The results show that a convenient return experience does not have an effect on customer retention for web shops. This is in contrast with the suggestion of Ling (Alice) Jiang (2013) that convenient experiences online can increase customer loyalty. However, indirectly through customer satisfaction with the return process, a convenient return experience has a small effect on customer loyalty intentions, but does not have an effect on whether the customer will actually buy again at the web shop. These findings are not surprising. Similar to the results in this thesis, Berry et al. (2002) found that customer’s perception of convenience directly influences customer satisfaction. Likewise, Mollenkopf et al. (2007) found that customer loyalty is indirectly effected by customer effort through customer satisfaction with the return process.

There is a discrepancy between the effect of customer convenience experienced and customer satisfaction with the product return process on customer retention. There is a positive influence on loyalty intentions if the customer has a convenient and satisfying return experience, but these experiences have no effect on whether they actually purchase a product again at the same web shop. Johnston & Kong (2011) can explain this discrepancy as they find in their work that positive experiences with a firm may lead to loyalty intentions, but that it does not necessarily lead to a repurchase, even though the chances are increased. These findings can also be explained through the difference between attitudinal customer loyalty and behavioral loyalty as proposed by Bowen & Chen (2001). There may also be other reasons, outside the scope of this

(32)

32 thesis, why customers do not purchase again at the web shop that have nothing to do with the return experience of the product.

The results may indicate that return policies are a small part in a bigger strategy to retain customers. The return process is only one of many opportunities in which web shops can contact the customer with which they can manage their customer retention strategies. Web shops can have return policies that provide customers with convenient return experiences that may influence the customer’s loyalty intentions minimally, but there are many other factors that have nothing to do with the return experience that can co-determine whether customers will purchase something again at a web shop or not. In this sense a return policy that creates convenience for the customer can be viewed as a tool that can help increase customer satisfaction with the return process and in that way increase the chances of loyalty intentions. It can be argued that this strategic policy is not enough on its own for web shops to retain customers, as it only has a very small impact. In the light of further research, it can be examined to what extent return policies play a role in retaining customers for web shops compared to other customer retention strategies.

Factors Influencing Convenience

The last four hypotheses were tested in order to inform if and what factors can affect convenience experienced which, in turn, can indirectly influence customer retention. The specific factors that were tested are product type and customer characteristics. Even though convenience experienced by the customer does not directly affect customer retention, these factors can inform us about what factors are important in forming return policy strategies because they influence the return experience.

Ramanathan (2011) finds that product characteristics and the returning of those products can influence customer loyalty. The results of the last four hypotheses show that product type and return method do not influence convenience experienced in the product return process. This is unlike the findings of Ramanathan (2011) who finds that different product types do matter. Mollenkopf et al. (2007) theorize that customer characteristics can have an influence on the return experience and through that influence customer loyalty. The results do indicate that customer characteristics can affect how convenient the return process is experienced. It did not matter if customers owned a car or not, but there was a significant difference between people who had a full time job and people who did not have a job and how convenient they experienced

(33)

33 returning a product. It is therefore possible that other customer characteristics that are not addressed in this thesis can influence the return experience. This is something that can be considered for further research.

Research has suggested that return policies are the most effective if they are based on product and customer characteristics. Davis et al. (1998) suggest that the characteristics of a product should be used to determine what strategy is used for product returns in combination with cost considerations. Yu & Wang (2008) suggest that return policies should be linked with customer characteristics and product type. The results show that product type and return method are not of influence on how the returning of a product is experienced. This indicates that the return experience is more dependent on customer characteristics. Of course the type of product and the way the product is returned can matter, but if web shops want to focus on making the return experience more positive in order to increase customer satisfaction, which in turn increases their chances of creating loyal customers, web shops should focus their policies more on their customer’s characteristics than on product characteristics. That being said, from a realistic cost perspective it might not be feasible to maintain return policies that suit many different types of customers. It is something that can be looked into with further research.

5.3 Managerial Implications and Further Research

Convenient Return Policies and the Intended Effect of Customer Loyalty

The practical aim of the thesis was to see if a convenient return policy has the intended effect of customer retention for web shops. This knowledge can help managers make the trade-off between what return policies they should implement and the costs that are associated with those policies. The results show two things. One, is that a convenient return experience indirectly positively influences loyalty intentions. Two, there is no indication that a customer will actually purchase again at a web shop after they had a convenient return experience.

From a cost aspect, it seems more favorable for the web shop if the customers purchased again at the web shop, than if customers just intend to stay loyal but do not actually purchase again from the web shop. Given this logic, it would be advisable for managers not to focus on implementing return policies that create convenience for the customer, because it does not have the desired effect. However, it could be possible that loyalty intentions have other benefits for web shops. For example attitudinal loyalty can reduce marketing costs. In that case, it will be

(34)

34 advisable for managers to develop convenient return policies that contribute to customer satisfaction with the return process in order to increase loyalty intentions. For example, loyalty intentions, as defined in this thesis, partly includes telling and recommending friends about the web shop. Positive word of mouth advertisement could mean free attraction and advertisement for the web shop and spare the web shop costs in other areas of the company. In that case, a positive and convenient return experience by the customer can contribute to a reduction in costs because it indirectly contributes to customer loyalty.

This reasoning is hypothetical and therefore further research should be done in this area. A case study of a specific web shop can be done to see what trade off web shops make internally between return policies, costs and the benefits of customers that only intend to stay loyal. If the results are favorable for customer loyalty intentions in the further research, it is worthwhile for managers to create strategic return policies that make it convenient for customers to return the products, in order to increase customer satisfaction with the return process.

The Convenience of Return Policies

The results of the last four hypothesis show that product type does not matter for the question how convenient customers experience the returning of a product, but that customer characteristics do. This can be an indication that managers should focus their product return strategies on their customer characteristics instead of the type of product if managers want to increase the positive experience of returning products. Managers should ask themselves how feasible it is to adjust return policies to customer characteristics. If the web shop has many different types of products and target audiences, it is difficult to create one return policy that will suit all of them. It is also questionable how well it will come across to customers if web shops maintain different policies for different customers. Some customers may feel discriminated and there might be a possibility to create the opposite effect of customer loyalty.

Hypothesis 7 (H7) and hypothesis 8 (H8) reveal an interesting finding. The purpose of H7 was to measure whether effort was an important factor in convenience experienced by customers and H8 was meant to measure if the element of time is important to customers when they return products. The results showed that customers that have less time experienced the returning of products as less convenient than customers that seemed to have more free time available. The ability to use a car to return a product, does not impact how convenient the return experience is.

(35)

35 For these reasons, physical effort to return a product seems to matter less than the time element for customers. Here lies an area for further research. To what extent does time influence the product return experience? How can managers create policies that suit the time needs of customers – and what are those needs based on time? If time indeed is an important factor in the product return experience, managers should shift the focus to serve the time needs of their customers when they create their return policies.

On another note, managers should also ask themselves, how positive they would like to make the experience of returning products. If the returning of products becomes easy, product returns may increase for web shops and may create the undesired effect of high operational costs. In this case, it would be more desirable to have one return policy that is not too convenient, in order to raise the barrier for customer’s to return a product and reduce operational costs that stem from product returns. Managers should contemplate to what extent customer loyalty benefits their web shop if they compare it with the costs that come with product returns. If managers know the latter, they can decide to what extent they want to cater to the return needs of different customers. The cost-benefit trade off can vary between web shops; there is no single solution that suits all web shops.

5.4 Study Limitations and Further Research

Like every study, this study has limitations. The main limitations have to do with the operationalization of concepts and constructs. Something can also be said about the design of the questionnaire. These limitations are explained further below.

Convenience is a concept that is difficult to define and operationalize. In this thesis convenience was measured in terms of time and effort as defined by the work of Berry et al. (2002). Judging how much time and effort spent on returning a product is reasonable, is a matter of perception, and therefore difficult to determine how much effort or time spent on returning a product is reasonable. There are also other elements that could have been included in the construct of convenience, such as how money is refunded for example. If it takes a lot of time and effort to get reimbursed, it could have influenced the customer’s return experience and satisfaction with the return process.

The way product type is determined in this thesis could have had an influence on the results. In this thesis, product type was based on size. There were three product types based on

An exploration of the effect of convenience for customers when returning products on customer retention for web shops