The relative importance of the service clues and the influence of past experience and frequency of use

(1)

The Clues to the

Customer Experience

-The relative importance of the service clues and the

influence of past experience and frequency of use

Master Thesis

By

Roy G.A. Menger University of Groningen Faculty of Economics and Business

(2)

In environments where customers purchase not just a tangible product, but instead pay for a service that has both tangible and intangible aspects and takes place over prolonged periods of time (for instance, a dinner at a restaurant, or in the case of the current study, a bus drive). In these situations, it is important for managers to not only look at the specific outcomes of the service, but at the customer experience as a whole, and then take steps to improve the individual aspects of the service that make up the customer experience. Based on previous research, this paper defines these aspects as ‘clues’, and aims to explore the effect of these clues on the customer experience, as well as the influence of past customer experience and the frequency with which the customer uses the service. The clues have previously been divided into three categories: functional-, mechanic and humanic clues. This paper has split up the humanic clues into humanic employee clues and humanic customer clues, based on the premise that the customer experience is partly reliant on the behavior and appearance of other customers. While functional clues assess the core aspects of the service, mechanic clues include the sensory presentation of the service and humanic employee clues involve the performance and behavior of the service employees. As the findings suggest that humanic

customer clues do not significantly affect the customer experience, they are not mentioned

here in further detail.

(3)

Measurement scales were developed, and customers were asked to fill in an online questionnaire, while recalling their most recent bus drive. The findings have several important implications:

It was proven that customer experience is indeed partly reliant on past experience. When customers have had negative experiences in the past, this therefore negatively affects their future customer experience, and vice versa. Furthermore, for customers who have had more positive past experiences, the humanic employee clues decrease in importance. Additionally, the findings indicate that while the effect of functional clues is not affected by frequency of use, mechanic clues become less important, and humanic employee clues become more important as frequency of use increases. A second approach was used to examine differences between low-frequent and high-frequent users, and an interpretation of the outcomes is presented in the following figures:

Low-frequent users high-frequent users

The figures show that in order to improve the customer experience for low-frequent customers (in order to convert them into high-frequent customers), managers should primarily pay attention, to the mechanic clues (furniture, color schemes, smells, climate, music). When the aim is to keep high-frequent customers coming back and possibly improve word-of-mouth ratings, the aim should be to invest firstly in mechanic clues, followed by functional- and humanic employee clues. As the findings show that humanic clues become more important when customers make more frequent use of the service, employee-training investments could ensure the satisfaction and future patronization of the most loyal customers. As past experience has proven to contribute to future experiences, large initial investments seem preferable over smaller investments over time.

Past Experience Past Experience Mechanic Clues Mechanic Clues Functional Clues Humanic Employee Clues

(4)

This master thesis is the product of six months of research and one-and-a-half years of studying marketing management, including an internship at Qbuzz Groningen Drenthe.

After having received my bachelor’s degree in Business Administration at the University of Groningen, I decided to move on the MSc Marketing management program at the same university, as marketing has been my passion for some time. After the first year, I had finished all my courses which meant that only the master thesis stood in the way of me graduating. I decided that I wanted to write my master thesis at a company, and while this meant I would have more on my plate than ever before, I was eager to get started and learn as much as I could about working in an actual business environment. Qbuzz Groningen/Drenthe gave me that opportunity. They not only gave me all the freedom I needed in order to write my thesis and graduate, but more importantly, the opportunity to learn and develop my skill set. I sincerely hope that this thesis will not only add to existing academic literature, but will also help Qbuzz in their journey towards optimizing the customer experience.

(5)

Management Summary... 1

Preface ... 3

1. Introduction ... 6

2. Theoretical Framework ... 8

2.1 Customer Experience ... 8 2.1.1 Conceptual model ... 8 2.2 Service Clues ... 9 2.2.1 Functional Clues ... 10 2.2.2 Mechanic Clues ... 10 2.2.3 Humanic Clues ... 11 2.2.4 Relative Importance ... 11 2.3 Past Experience ... 12 2.4 Frequency of Use ... 14

3. Research Design ... 15

3.1 Data Collection ... 15 3.2. Measurement Scales ... 16 3.4 Method of Analysis ... 18

4. Results ... 20

4.1 Main Effects ... 20 4.2 Interaction Effects ... 21

4.3 Frequency of Use Reexamined ... 22

5. Discussion and Conclusion ... 24

5.1 Discussion ... 24

5.2 Theoretical Implications ... 26

5.3 Managerial Implications ... 27

5.4 Limitations and Directions for Future Research... 28

References ... 29

Appendix ... 33

A: Questionnaire... 33

(6)

1. Introduction

While customer satisfaction and service quality have been researched extensively throughout the years, customer experience is a relatively new concept that has received a rather limited amount of attention. The customer experience is, however, becoming increasingly important and is often favored over measures such as customer satisfaction. “When it comes to customer journeys, organizations that are able to skillfully manage the entire experience reap enormous rewards, such as enhanced customer satisfaction, increased revenue and greater employee satisfaction” (Rawson et al., 2013). This is perhaps even more so the case in service industries. The touch points in service encounters are usually far more numerous than in situations where a customer merely buys a product (Johnston & Kong, 2011). These touch points can trigger a reaction in the customer, consciously or subconsciously, and affect their overall experience of the service. Berry, Wall and Carbone (2006) call these triggers ‘clues’. When the company providing the service has insights into which clues carry the most weight in the mind of the consumer and thus have the largest contribution to the overall customer experience, they can effectively make organizational decisions based on these insights. One of the aims of the current research is therefore to investigate the relative importance of the service clues on customer experience.

The influence of these dimensions on the overall customer experience has received some attention (Mathwick et al., 2001; Chua et al., 2014). Additionally, Chauhan & Manhas (2014) investigated the multidimensionality of customer experience, and concluded that the domain of customer experience offers a rich agenda for future research. Bitran et al. (2008) called for future research on the division of a service encounter in the mind of the consumer. Furthermore, Kumar et al. (2014) examined past and current customer experience in relation to the state of the economy, and proposed that future studies could use alternate approaches for detecting the prevalence of satisfaction and service experience information. The current study explores the multidimensionality of customer experience from a consumer perspective, and aims to fill the research gap that exists when it comes to the hierarchies of customer experience antecedents in the mind of the consumer.

(7)

frequency with which the customer uses the service could be of importance. Customers who have been exposed to the service relatively often, might have a higher tolerance for service failure (Anderson & Sullivan, 1993). On the other hand, they may have higher expectations, and thus are more critical of the performance than customers who only use the service occasionally. The relevance of this variable is also supported by van Pham and Simpson (2006) , who note that frequency of use is an interesting variable that requires further research in the field of expected and perceived service quality.

These gaps in research lead to the following research question: In what order do the

service clues affect the customer experience and to what extent is the relationship between the clues and the customer experience influenced by past experience and frequency of use? The

research question will be answered through empirical research which will be conducted in the context of a bus company called Qbuzz. A survey will be developed and distributed digitally among customers (bus travelers) of this company.

(8)

2. Theoretical Framework

In this chapter, an overview of relevant academic papers will be given. The constructs will be defined and, based on arguments from existing literature, hypotheses will be developed in sections 2.2, 2.3 and 2.4.

2.1 Customer Experience

In this thesis, customer experience is defined as “the customer’s cognitive and affective assessment of all direct and indirect encounters with the firm relating to their purchasing behavior” (Klaus & Maklan, 2012). This definition is consistent with definitions used by Verhoef et al. (2009). Customer experience has also been called ‘experiental value’. Experiential value is defined as “customers’ perceptions of value that arise from experience and is the result of direct or indirect interaction during the consumption process” (Mathwick et al., 2001). Furthermore, Gentile, Spiller and Noci (2007) explain that the customer experience originates from a set of interactions between a customer and a product, a company, or part of its organization, which provoke a reaction. This experience is strictly personal and implies the customer’s involvement at different levels (rational, emotional, sensorial, physical and spiritual). This explanation includes several aspects of customer experience. First, a set of interactions refers to the touch points of the service, and the clues that affect the experience. Second, the customer has to react to these interactions, or, clues in a certain way. The model in this thesis proposes that different types of clues have different effects on the overall experience.

2.1.1 Conceptual model

In figure 2.1 the conceptual model is displayed. The dependent variable in the model is customer experience. Based on existing literature (Berry, Wall & Carbone, 2006), the model proposes that customer experience results from different types of clues. Section 2.2.1, 2.2.2 and 2.2.3 elaborate upon these effects. Furthermore, existing literature has indicated that the effect of the service clues is mitigated by the past experience of the customer. This effect is further explained in section 2.3. Lastly, the model proposes that when a customer makes more frequent use of the service, the effect of the functional- and mechanic clues diminishes, while the effect of the humanic employee- and customer clues is strengthened. This effect is elaborated upon in section 2.5.

(9)

2.2 Service Clues

Berry, Wall and Carbone (2006) identified three types of clues from which the customer experience is formed. These types of clues are: functional clues, mechanic clues and humanic clues. Examples of the different types of clues are presented in table 2.1. The impact of these types of clues on customer experience is supported by Wu and Liang (2009). They investigated the antecedents of customer experience in a restaurant setting and found significant influences of restaurant environment, service personnel performance and other customer interaction on customer experience. Andersson and Mossberg (2004) proposed that the service experiences consist of three ‘rings’. At the core, there is the must. Around the core is a ring including satisfiers. The outer ring is called delight. These three rings work together to create a customer experience, but each ring plays a different role in the process.

Customer Experience Clues

Customer Experience

Figure 2.1 Conceptual model

(10)

Table 2.1

Type of clue Example

Functional Reliability of the service, departure and arrival times

Mechanic Comfortable seats, cleanliness of the vehicle

Humanic employee Friendliness and helpfulness of the bus driver

Humanic customer Behavior of other passengers

2.2.1 Functional Clues

Functional clues are said to support the core of any service because they address the problem that brings the customer to the market. They reveal the reliability and competence of the service. Anything that indicates or suggests the technical quality of the service is a functional clue (Berry, Wall, Carbone, 2006). In a restaurant setting, for example, these clues refer to the quality of the food (Chua, Jin, Lee & Gog, 2014). Different types of service clues play different roles in the consumers’ minds (Berry, Wall, Carbone, 2006). Functional clues affect the calculative perceptions of quality. In practice, this means that functional clues are about meeting customers’ expectations. This reasoning is supported by Andersson & Mossberg (2004). As stated, they view the service experience as three rings, with at its core the must. When this attribute is missing, the service cannot perform its basic function. The functional clues are therefore likely to positively contribute to the customer experience.

H1a: Functional clues are positively related to customer experience.

2.2.2 Mechanic Clues

Mechanic clues concern the sensory presentation of the service (Berry, Wall, Carbone, 2006). In the case of the current research, mechanic clues encompass for example the cleanliness and the interior of the vehicle. Bitner (1992) proposed that positive responses to the atmospherics, physical design and décor elements can lead to positive beliefs and feelings associated with the organization, its products and its people. The servicescape literature furthermore suggests that ambient and design factors contribute heavily toward the experience of the service. Extreme ambient factors, such as very high- or low- temperatures may lead to the display of avoidance behavior (Ezeh & Harris, 2007). Like functional clues, mechanic clues are therefore believed to positively affect the customer experience.

(11)

2.2.3 Humanic Clues

Brocato, Voorhees & Baker (2012) suggests that assessments of customer experience that merely focus on a firm’s performance are incomplete in contexts where customers share the service facility. This finding is supported by Nicholls (2010), who found that positive customer to customer interactions have a positive impact on the customers’ evaluations of the service experience. Therefore, in this research, the humanic clues are split up in two categories: humanic employee clues and humanic customer clues.

2.2.3.1 Humanic employee clues

Humanic clues come from the behavior and appearance of service providers (Berry, Wall, Carbone, 2006). Previous research indicated that courteous, knowledgeable and efficient service employees could provide an image of service excellence (Keng et al. 2007). Humanic (employee) clues can be used to exceed customer expectations, and thus come into play when customers are pleasantly surprised by the humanic aspects of the service. The following is proposed:

H1c: Humanic employee clues are positively related to customer experience.

2.2.3.2 Humanic customer clues

As stated, other customers can heavily influence the way in which the service is experienced (Brocato et al., 2012; Nicholls, 2010). Furthermore, Wu & Liang (2009) also found a significant, positive influence of other customer interactions on experiential values. The servicescape literature proposes that these social factors can serve to either enhance or inhibit the service experience (Bitner, 1992). The current research encompasses this social effect as part of the humanic clues through the following hypothesis:

H1d: Humanic customer clues are positively related to customer experience.

2.2.4 Relative Importance

Functional clues alone are not enough to keep the customer satisfied, but are

fundamental to the service experience (Berry, Wall, Carbone. 2006). As proposed by Andersson & Mossberg (2004), these aspects are the must of a service, and respond to a base need of a customer.

Mechanic clues, on the other hand, influence an affective component of the customer

(12)

Lastly, humanic (employee and customer) clues are used to exceed customer expectations. Customers can therefore be pleasantly surprised when they perceive the customer service as being above par.

Interestingly, Andersson & Mossberg (2004) found that customers’ willingness to pay increases as the order of the need increases. Customers were thus more willing to pay for intellectual and social stimulation than for basic physiological comfort. Moreover, Chua et al. (2014) found that the impact of humanic clues on experiential value was greater than the impact of functional clues on the same outcome variable.

This leads to the following hypotheses:

H2a: Mechanic clues will have a larger impact on customer experience than

functional clues.

H2b: Humanic clues will have a larger impact on customer experience than

mechanic clues.

2.3 Past Experience

(13)

customer who has had negative experiences in the past is more likely to have negative experiences in the future, and vice versa. This effect is incorporated in the model and results in the following hypothesis:

H3: Past experience is positively related to customer experience

It has been confirmed in literature that tolerance for inferiority moderates the relationship between service failure and customer dissatisfaction (Zhang, Lam, Chow. 2009). The article stated that customers whose have the personality trait of low tolerance for

inferiority respond less strongly to service failure. Personality traits are beyond the scope of

the current research, but the experience that the customer has had in the past may have a similar influence on the relationship between the service clues and customer experience. Furthermore, previous research has found that when the perceived value of previous experience increases, the negative effect of dissatisfaction with the service recovery on trust and affective commitment weakens (Vázquez, Suarez & Diaz., 2010). One explanation for this could lie in the cognitive interpretation bias, which indicates that customers have a desire to uphold their levels of trust in, and commitment to, the firm (Ahluwalia, 2000). Reasoning from this perspective, the opposite would also hold: Positive service experiences (clues) would have a stronger effect on customers with previous negative experiences, since they have a desire to strengthen their trust in the company.

This effect is included in the model and is hypothesized as follows:

H4a: Past experiences mitigate the relationship between functional clues and

customer experience

H4b: Past experiences mitigate the relationship between mechanic clues and

customer experience

H4c: Past experiences mitigate the relationship between humanic employee clues

and customer experience

H4d: Past experiences mitigate the relationship between humanic customer clues

(14)

2.4 Frequency of Use

Previous research has suggested that usage frequency or frequency of use is related to concepts of customer satisfaction and service failure. Frequency of use (Van Pham & Simson, 2010) refers to the amount of times during a given period the customer makes active use of the service. In a study on customer satisfaction with commodities such as microwaves and VCR’s, it was found that usage frequency significantly contributes to the customer satisfaction (Jung, 1991). Rust & Williams (1994) found that the length of the patronage has an effect on the relationship between customer satisfaction and repurchase intention, implicating that long-time customers are more forgiving and that attention to customer satisfaction for new customers is critical. This finding is also supported by Anderson & Sullivan (1993). They found that customers who are more experienced will be more forgiving when it comes to negative experiences. These findings demonstrate the negative sides of the equation, which is to be expected as they can be called hygiene factors in terms of Herzberg’s

motivator-hygiene theory (Herzberg, 1968). These so-called hygiene factors are the causes for

dissatisfaction and comprise the core aspect of a service. In the case of the service clues, these comprise the functional and to an extent, the mechanic clues. As mentioned earlier, Andersson and Mossberg (2004) found that willingness to pay increased as the order of the need fulfilled by the service increased. Alternatively, when higher order needs (Maslow, 1987) are fulfilled, the service clues fulfilling these needs could be called motivators (Herzberg, 1968). In the current research, these motivators comprise the humanic (employee and customer) clues, as they provide social and intellectual stimulation, and it is likely that more frequent customers, who have already fulfilled their base needs in the past, are looking towards these ‘motivator-oriented’ service clues for their customer experience. This results in the following hypotheses:

H5a: Frequency of use mitigates the effect of functional clues on customer

experience

H5b: Frequency of use mitigates the effect of mechanic clues on customer experience

H5c: Frequency of use strengthens the effect of humanic employee clues on customer

experience

H5d: Frequency of use strengthens the effect of humanic customer clues on customer

(15)

3. Research Design

3.1 Data Collection

The data for this research was collected among customers of a bus company in the Netherlands. The geographic area in which this bus company operates is a combination of the Dutch provinces Groningen and Drenthe.

3.1.1 Qualitative Research

One way to generate items that should make up the scales related to the service clues, is to conduct interviews (Churchill,1979). Therefore, to start of the process of item generation, interviews were conducted with nine regular bus travelers. The questions that were asked were: which aspects are most important to you when travelling with a bus, which aspects

could potentially ruin a bus drive for you, and which aspects could positively surprise you during a bus drive. The answers that were given were very helpful in determining the items

that should make up the service clue constructs. Temperature and the behavior of fellow passengers were mentioned by four of the respondents. Punctuality, comfortable seating and the friendliness of the bus driver were mentioned by all respondents. When asked the

question: “What could positively surprise you in a bus drive?”, one respondent said: “[…] you really expect a bus to bring you from A to B on the predetermined times without delays. Because this is often not the case, I am positively surprised when this (departing and arriving on time) does happen.” This quote symbolizes hypothesis H4a.

3.1.2 Quantitative Research

The data for the current research was gathered by means of survey (appendix 1). The survey consisted of 35 Likert-scale items revolving around the service clues and the customer experience. The survey was distributed digitally, in order to lower the barrier for participating. A combination of convenience sampling and snowball sampling was used in order to obtain the sample. The participants were asked to think about their last bus drive, and in that state of mind, fill in the survey. The items and scales that made up the instrument are presented and explained in section 3.2.

(16)

an average age of 26 (M = 25.6, SD = 10.8). Furthermore, the average travel frequency (in the last month) was 18 (M = 17.83, SD = 19.16).

3.2. Measurement Scales

The scale developed by Chua, Jin, Lee and Hog (2014) has served as a basis for the items related to the service clues. The items used to measure the humanic customer clues come directly from the research by Brocato et al. (2012). They identified three dimensions in which other customers can influence the service experience. Furthermore, the ‘satisfaction with travel scale’ will be used to measure the overall customer experience with the trip. The ‘satisfaction with travel scale’ was first introduced by Jakobsson Bergstad et al. (2011), and was later improved by Ettema et al. (2011), by adding an affective dimension. Olsson, Friman, Pareigis & Edvardsson (2012) subsequently validated the ‘satisfaction with travel scale’. Since this scale is directly applicable in the context that is used in this paper, it will be applied as a means to measure the overall customer experience, as it has been in previous research (Bergstad et al., 2010; Ettema et al., 2011; Olsson et al., 2012). The scale related to past experience is based on the past experience scale used by Huang & Hsu (2009). The measurement scales can be found in table 3.1.

Table 3.1 Measurement scales

Construct Item

Mechanic Clues

The interior of the vehicle was attractive The noise in the vehicle was bothersome*

The information supply in the bus was appropriate The cleanliness of the vehicle was appropriate The seating in the vehicle was comfortable The accessibility of the vehicle was appropriate The temperature in the vehicle was to my liking I could find a place to sit with ease

The timeliness of the drive was acceptable

Functional Clues

The bus drive brought me to my destination on time I wish the departure and arrival times of the trip had been different

The bus drive was consistent and reliable

Humanic Employee Clues

The bus driver created a positive atmosphere* The bus driver was friendly to me

I took offense to the behavior of the bus driver The bus driver made the drive less enjoyable

The driving style of the bus driver was appropriate* I felt safe during the drive*

Humanic Customer Clues

The behavior of the other passengers was appropriate The other passengers were friendly

(17)

The other passengers’ behavior was pleasant

Customer Experience (STS)

The bus drive was the best I can think of The bus drive had a high standard The bus drive worked well

The bus drive made me relaxed The bus drive made stressed I had confidence in the bus drive I was enthusiastic about the bus drive

Past Experience

Your overall evaluation on past bus drives is positive Your overall evaluation on past bus drives is favorable You are satisfied with your past bus drives

You are pleased with your past bus drives

Descriptive items

What is your age? What is your gender?

With what bus company was your last bus drive? Was this bus drive in the city or in the country? What was the purpose of your last bus drive? Could you have made this journey by other means (car/bike)?

* deleted after Factor Analysis

In order to find the underlying constructs, a Principal Component Analysis was conducted on the items used to measure the independent variables. The items believed to measure the dependent variable were not included in the factor analysis but were handled separately. The factor analysis initially found nine factors, based on Eigenvalues higher than one, but because the measurement scales were so clearly defined, a second factor analysis was conducted with a predetermined number of factors (five).

The KMO statistic for the factor analysis was ,791 and the Bartlett’s test of sphericity was significant, which indicated that a factor analysis was indeed appropriate..

Three items (The bus driver created a positive atmosphere and The driving style of the

bus driver was appropriate and I felt safe during the drive) showed cross loadings (appendix

B). These items were deleted from the dataset. Furthermore, one item (The noise in the

vehicle was bothersome) was omitted due to the fact that it did not load significantly on any

(18)

Table 3.3 Internal consistency

Factor Construct α

1 Functional Clues ,792

2 Mechanic Clues ,795

3 Humanic Employee Clues ,652

4 Humanic Customer Clues ,899

5 Past Experience ,937

The items that were based on the ‘satisfaction with travel scale’ were not included in the factor analysis, since this is a scale that has been used and validated before by Olsson, Friman, Pareigis & Edvardsson (2012). A Pearson correlation matrix was consulted to make sure the items did in fact correlate with each other. All of the items correlated significantly with each other at the .05 level. To check for the internal reliability of the dependent construct, a Cronbach’s alpha was generated. The α of this construct was .872, which means that the internal reliability is sufficient.

In order to get to variables as input for the regression model, sum variables are made in which the average of the items is taken. The means and standard deviations of the resulting regression variables are presented in table 3.4.

Table 3.4 Constructs

Construct Mean SD

Functional Clues 5,18 1,29

Mechanic clues 5,19 ,94

Humanic Employee clues 5,53 1,17

Humanic Customer Clues 5,03 1,24

Past Experience 5,17 1,09

Customer Experience 4,61 ,86

3.4 Method of Analysis

Multiple regression will be used to test the hypotheses. The first model includes only the main effects.

CEi = α0 + β1FCi + β2MCi + β3HECi + β4HCC i+ β5PEi + εi (1)

CE= Customer experience

FC= Functional clues

MC= Mechanic clues

HEC= Humanic employee clues

(19)

PE= Past experience FOU = Frequency of use αo = Intercept

εi = Error term i = Respondent

Second, the proposed interaction effects of past experience and frequency of use are examined. Two models are constructed, one including past experience as a moderator (2) and one including frequency of use (3):

CEi = α0 + β1FCi + β2MCi + β3HECi + β4HCCi + β5PEi + β6(FCi*PEi) + β7(MCi*PEi) + β8(HECi*PEi) + β9(HCCi*PEi) + εi (2)

CEi = α0 + β1FCi + β2MCi + β3HECi + β4HCCi + β5PEi + β6(FCi*FOUi) + β7(MCi*FOUi) + β8(HECi*FOUi) + β9(HCCi*FOUi) + εi (3)

Third, a different approach will be used to test for differences in relative importance of the service clues for low- compared to high-frequent users. The sample will be split, in order to create two different subsets of travelers with regard to frequency of use. The top 52,2% (N=47) is named high-frequent user and the bottom 47,8% (N=45) is named low-frequent user. A model will be constructed for each subset, including the main effects of the service clues, in order to observe possible differences in relative importance:

(20)

4. Results

4.1 Main Effects

In order to analyze the proposed relationships, several multiple regressions were performed. Because the conceptual model presented in chapter 2 included both main effects and interaction effects, the first step is to estimate a model including only the main effects. The model shows sufficient explanatory power with an R2 of .458 and an R2adj of .426. The results of the regression analysis are presented in table 4.1.

Table 4.1 Main Effects

Variables Hypothesis Model 1

Functional Clues H1a ,180**

Mechanic Clues H1b ,317**

Humanic Employee Clues H1c ,067 Humanic Customer Clues H1d ,405

Past Experience H3 ,316**

*** p<.01, ** p<.05, *p<.10

The effect of functional clues on customer experience is found significant (β = .180, p

< .05). This would indicate that functional clues have a positive effect on customer

experience. H1a is therefore accepted. There also is a positive relationship between mechanic clues and customer experience (β = .317, p < .05). H1b is accepted, which means that more positive perceptions of mechanic clues will lead to higher levels of customer experience. Neither humanic employee clues nor humanic customer clues show significant effects on customer experience, which means that H1c andH1d are rejected, and a relationship between humanic clues and the experience of the customer cannot be assumed. Lastly, a relationship between past experience and customer experience was found (β = .316, p < .05). This would indicate a positive effect of past experience on current customer experience and thus H3 is accepted.

(21)

4.2 Interaction Effects

In order to check for the moderating effects and test hypotheses H4a, H4b, H4c, H4d, H5a, H5b, H5c and H5d, interaction terms were constructed and two separate models (2 and 3) were estimated. The R2 (R2adj) of the models is .524 (.471) and .502 (.447) respectively. Model 2 shows VIF-scores between 22 and 52, indicating that multicollinearity is an issue. This is due to the fact that the main effects are also present in the interaction effects. Model 3 only shows moderately high to high VIF-scores for the interaction terms (ranging from 15 to 32).

The results of the regression analysis are presented in table 4.2.

Table 4.2 Regression results for Customer Experience testing for moderating effects

Hypothesis Model 2 Model 3

Main effects

Functional clues -,46 ,185

Mechanic clues -,02 ,459***

Humanic employee clues 1,4989** -,127

Humanic customer clues -,223 ,10

Past Experience ,975* ,286**

Interaction effects

Past experience*functional clues H4a ,338

Past experience*mechanic clues H4b ,619

Past experience*humanic employee clues H4c -2.280** Past experience*humanic customer clues H4d ,490

Frequency of use*functional clues H5a ,255

Frequency of use*mechanic clues H5b _-1,027*

Frequency of use*humanic employee clues H5c _,739**

Frequency of use*humanic customer clues H5d ,083

*** p<.01, ** p<.05, *p<.10

(22)

While the results indicate that frequency of use does not have a significant effect on the relationship between functional/humanic customer clues and customer experience, rejecting H5a and H5d, they do show effects of frequency of use on the two remaining relationships. As hypothesized, mechanic clues become less important as frequency of use increases, which means that H5b is accepted. Additionally, the relationship between humanic employee clues and customer experience becomes stronger when frequency of use increases, indicating support for H5c. The implications of these results will be discussed in chapter 5.

4.3 Frequency of Use Reexamined

In order to further examine the effects of frequency of use, a second approach was used to examine differences in magnitudes of effects between low-frequent and high-frequent travelers. In order to do this, the sample was split. The top 52,2% (N=45) was named

high-frequent users, and the bottom 47,8% (N=47) was named low-high-frequent users. The means and

standard deviations of frequency of use for the three different groups are reported in table 4.3.

Table 4.3 Frequency of use

Group Mean SD

Low-frequent users 4,55 2,28

High-frequent users 31,69 19,22

For each sub-sample, a regression model was estimated. The R2 (R2adj) of model 4 is .463( .397), while for model 5, it is .535 (.475). The results are presented in table 4.4.

Table 4.3 Main effects of low- vs high-frequent users

Model 4 Low-frequent Model 5 High-frequent Main effects Functional clues ,177 ,248** Mechanic clues ,341** ,265**

Humanic employee clues -0,44 ,226*

Humanic customer clues ,055 ,205

Past Experience ,341** ,215*

*** p<.01, ** p<.05, *p<.10

(23)

First, while functional clues do not have a significant effect on customer experience for low-frequent users, this effect is significant for high-frequent users (β = .248, p < .05). This is counter to what H5a suggested.

Second, mechanic clues appear to have a significant positive effect on customer experience for both low-frequent ((β = .341, p < .05) and high-frequent (β = .265, p < .05). Since these are the standardized coefficients, they can be compared, demonstrating that low-frequent users do indeed find mechanic clues more important than high-low-frequent users. This is in line with H5b.

Third, as H5c suggested, humanic employee clues are indeed far more important for high-frequent users than for low frequent users. The effect of humanic employee clues is not significant for low-frequent users, while for high-frequent users, it is significant, albeit at the .1 level ((β = .226, p < .1).

Fourth, humanic customer clues do not seem to have a significant effect on customer experience for either sub-sample. This is contrary to H5d, but is in line with models 1, 2 and 3, where the humanic customer clues were also not found significant.

Lastly, the coefficient of past experience is positive and significant for both sub-samples, but it seems to be a better predictor for customer experience in the low-frequent group (β = .341, p < .05), than in the high-frequent group (β = .215, p < .1).

The results have shown that low-frequent users and high-frequent users do in fact have different priorities when it comes to their customer experience. The results of model 3 and a comparison between models 4 and 5 show that mechanic clues have a stronger effect on customer experience for low-frequent users than for high-frequent users. Contrastingly, both approaches showed that humanic employee clues have s strong effect on customer experience for high-frequent users, while this is not the case for low-frequent users.

Looking at the hierarchies of the effects for low- vs high-frequent users, the following can be said: For low frequent users, mechanic clues and past experience are seemingly the only predictors of their customer experience. They are equally important. For high-frequent users, mechanic clues are also the most important predictor of customer experience, followed by functional clues, then humanic employee clues and past experience.

(24)

5. Discussion and Conclusion

This study set out to answer the following research question: How do the different

components of service quality affect the customer experience and to what extent is this relationship influenced by past experience and frequency of use? This question has been

answered through empirical research in the context of a bus company in the Netherlands.

While the main effects of functional- and mechanic clues were found significant, no support could be found for the main effects of humanic employee- and customer clues in the initial model. Furthermore, there is indeed a significant relationship between past experience and current customer experience. Additionally, past experience negatively moderates the relationship between humanic employee clues (which were found significant after adding the interaction effect) and customer experience. When testing for effects of frequency of use, it was found that for low-frequent customers, only mechanic clues are significant, while for high-frequent customers, mechanic-, functional- and humanic employee clues have an effect on customer experience.

5.1 Discussion

(25)

Furthermore, it seems that past experience is indeed a reliable predictor of future customer experience. This is in line with the findings of Joby (1992) and Verhoef et al. (2009).

As stated, humanic employee clues do become important in models 2, 3, 4 and 5. The results have shown that when the overall quality of past experience increases, the effect of humanic employee clues on customer experience weakens. This could be explained by the motivator hygiene theory (Herzberg 1968). As humanic employee clues are used to exceed expectations (Berry, Wall & Carbone. 2006), they serve as motivators. When past experiences are positive, the customer wouldn’t have as much need for exceeded expectations as when their past experiences have been predominantly negative. One can also look at this relationship the other way: when a customer has had predominantly negative experiences in the past, their future customer experience could be improved by the attitude and behavior of the service employees.

The effects of frequency are visualized in figures 5.1 and 5.2.

Figure 5.1 low-frequent users Figure 5.2 high-frequent users

Figure 5.1 only includes mechanic clues. The results shows that this is the only important service clue for low-frequent customers. It is supported by past experience. Figure 5.2 shows that for high-frequent customers, more factors come into play. This could be explained by the fact that high-frequent customers are more demanding than low-frequent customers. They will have a large frame of reference, and just plainly expect higher levels of service than do low-frequent customers. The hierarchy is visualized by putting mechanic clues at the base, as this was found to be the most important service clue for high-frequent

Past Experience Past Experience Mechanic Clues Mechanic Clues Functional Clues Humanic Employee Clues

(26)

customers, followed by functional- and humanic employee clues. This pyramid too, is supported by past experience, although the effect on customer experience is smaller than for low-frequent users. As stated, the effect of the humanic employee clues in figure 5.1 is mitigated by past experience. Combining this with the finding that humanic employee clues are only important for high-frequent customers when it comes to customer experience, the following conclusion can be made: customers who frequently use the service but have had predominantly negative experiences in the past are the most susceptible to helpful and friendly service employees (in addition to functional and mechanic clues). This conclusion is part of figure 5.3. Past Experience Low High Frequency of use Low  Mechanic,  Employee  Mechanic High  Mechanic  Functional  Employee  Mechanic  Functional

Figure 5.3 Service clues per customer segment

Figure 5.3 shows a matrix, presenting four quadrants of customers. Customers who don´t use the service often will, as elaborated upon earlier, benefit mostly from improved mechanic clues. Of those low-frequent customers, the ones with negative past experiences will also pay some attention to the humanic employee clues, as the results indicated that these clues become more important, the lower the value of past experience is. Similarly, for the customers in the bottom quadrants, mechanic and functional clues are important. The humanic employee clues, again, come into play when the past experiences of these customers have been predominantly negative. Sections 5.2 and 5.3 will give theoretical and managerial implications of these findings and conclusions.

5.2 Theoretical Implications

(27)

two, resulting in humanic employee clues and humanic customer clues. Contrasting the work of Chua, Jin, Lee and Gog (2014), this research did not find significant relationships between the humanic clues and the customer experience, but did find a significant effect of mechanic and functional clues. Furthermore, the current study explains, at least in part, the relative importance of these clues in the mind of the consumer. This relative importance, or hierarchy, is at least somewhat affected by the frequency with which the customer uses the service. The frequent users put the most emphasis on mechanic-, functional and humanic employee clues, while the low- frequent users only value mechanic clues. This study is the first to examine such hierarchical differences while including usage frequencies.

Additionally, the concept of past experience has been proposed by Verhoef et al. (2009) in that customer experience t-1 would have an effect on the customer experience at time t. The current study has included this proposed effect and found it highly significant, indicating that dynamic experiences are indeed present.

5.3 Managerial Implications

Managers in service environments have a wide selection of marketing tools and instruments at their disposal. The implications from the current study can aid them in making decisions regarding budget allocation for these instruments. It has become apparent that for low-frequent users, the mechanic clues have the biggest impact on their customer experience. Firstly, when the goal is to keep new customers coming back, the manager thus has to focus his or her attention towards increasing the sensory presentation of the service.

Secondly, when the goal is to keep frequent customers from staying loyal and recommending the service to their friends and family, the attention of the manager should be focused towards improving essentially all aspects of the service. Since mechanic clues were found to be the most important for high-frequent customers as well as low-frequent customers, investments are perhaps most effective in aspects of the interior of the service environment. This could mean up dating furniture, getting better equipment or even implementing free Wi-Fi. Since high-frequent customers also value functional clues and humanic employee clues, managers should aim to subsequently invest in increasing the quality of the core service and enhancing the performance of the employees through training.

(28)

effects for time t+1. Making relatively large investments at once, seems to be preferable over investing smaller amounts over a longer period of time.

5.4 Limitations and Directions for Future Research

(29)

References

Ahluwalia, R. 2000. Examination of Psychological Processes underlying Resistance to Persuasion. Journal of Consumer Research. 27(2): 217-232.

Anderson. E & Sullivan M. (1993). “The antecedents and consequences of customer satisfaction for firms.” Marketing Science. 12(2): 125-143.

Berry, L.L., Wall, E., Carbone, L.P. (2006). Service Clues and Customer Assessment of the Service Experience. Academy of Management Perspectives. 20(2): 43-57.

Bitner, M. J. (1990). Evaluating service encounters: The effects of physical surroundings and employee responses. Journal of Marketing. 54(2): 69-82.

Bitner, M.J. (1992). Servicescape: the impact of physical surroundings on customers and employees. Journal of Marketing. 56(2): 57-71.

Bitran G., Ferrer J. & Oliveira P. 2008. Managing Customer Experiences: Perspectives on the Temporal Aspects of Service Encounters. Manufacturing and Service Operations

Management. 10(1): 61-83.

Brocato, E., Voorhees, C., Baker, J. (2012). Understanding the influence of cues from other customers in the service experience: a scale development and validation. Journal of Retailing. 88(3): 384-398.

Chauhan, V. & Manhas D. 2014. Dimensional Analysis of Customer Experience in Civil Aviation Sector. Journal of Services Research. 14(1): 75-98.

Chua B., Jin N., Lee S., Goh B. (2014). Influence of Mechanic, Functional, and Humanic Clues on Customers’ Experiential Values and Behavioral Intentions in Full-Service Restaurants. Journal of Foodservice Business Research. 17: 67-84.

Churchill G. A. 1979. A Paradigm for Developing Better Measures of Marketing Constructs.

Journal of Marketing Research. 16: 64-73.

Ettema, D., Gärling T., Eriksson, L., Friman, M., Olsson, L.E., Fujii, S. (2011). Satisfaction with travel and subjective well-being: development and test of a measurement tool.

(30)

Ezeh, C. & Harris, L. 2007. Servicescape Research; a review and a research agenda.

Marketing review. 7(1): 59-78.

Gentile, C., Spiller, N., Noci, G. (2007). How to sustain the customer experience: an overview of experience components that co-create value with the customer. European Management

Journal. 25(5): 395-410.

Herzberg, F. (1968). One more time: how do you motivate employees? Harvard Business

Review. 65(5): 109-120.

Huang, S., Hsu, C. (2009). Effects of travel motivation, past experience, perceived constraint, and attitude on revisit intention. Journal of travel research.

Jakobsson Bergstad, C., Gamble, A., Gärling, T., Hagman, O., Polk, M., Ettema, D., Friman, M., Olsson, L. (2011). Subjective well-being related to satisfaction with daily travel.

Transportation. 38(1): 1-15.

Joby, J. (1992). “Patient Satisfaction; the impact of past experience”. Journal of healthcare

marketing. 12(3): 56-64.

Johnston, R., Kong, K. (2011). The Customer Experience: A Road Map for Improvement.

Managing Service Quality. 21(1): 5-24.

Jung, H. (1991). How product usage influences consumer satisfaction. Marketing Letters. 2(4):403-411.

Keng C., Huang, T., Zheng, L. & Maxwell, K., 2007. Modeling Service Encounters and Customer Experiential Value in Retailing. International Journal of Service Industry

Management. 18(4): 349-367.

Klaus P. & Maklan S. (2012). Towards a better measure of customer experience.

International Journal of Market Research. 5(2): 227-246.

Knutson, B. J., Beck, J.A., Kim, S., Cha, J. (2010). Service Quality as a Component of the Hospitality Experience: Proposal of a Conceptual Model and Framework for Research.

Journal of Foodservice Business Research. 13: 15-23.

(31)

Mascarenhas, O.A., Kesavan, R., Bernacchi, M. (2006). Lasting Customer Loyalty: A Total Customer Experience Approach. Journal of Consumer Marketing. 23(7): 397-405.

Maslow, A. 1987. Motivation and Personality. Harper and Row: New York.

Mathwick, C., Malhotra, N. & Rigdon, E. (2001). Experiential value: Conceptualizing, measurement and application in the catalog and internet shopping environment. Journal of

Retailing. 77(1): 39-56.

Nicholls, R. (2010). New directions for customer-to-customer interaction research. Journal of

services marketing. 24(1): 63-82.

Olsson, L.E., Friman, M., Pareigis, J., Edvardsson, B. (2012). Measuring service experience: Applying the satisfaction with travel scale in public transport. Journal of retailing and

Consumer Services. 19: 413-418.

Parasuraman, A., Zeithaml, V.A., Berry, L.L. (1985). A Conceptual Model of Service Quality and Its Implications for Future Research. Journal of Marketing. 49: 41-50.

Rawson, A., Duncan, E. & Jones C. (2013). The Truth About Customer Experience. Harvard

Business Review. 91(9): 90-98.

Rust, R., Williams, D. (1994). How length of patronage affects the impact of customer satisfaction on repurchase intention. Journal of consumer satisfaction, dissatisfaction and

complaining behavior. 7: 107-113.

Severt D., Tesone D. & Murrmann S. (2008). Prior experience satisfaction and subsequent fairness perceptions within the service experience. Journal of hospitality and leisure

marketing. 13: 121-137.

Torres, E.N. (2014). Deconstructing Service Quality and Customer Satisfaction: Challenges and Directions for Future Research. Journal of Hospitality Marketing & Management. 23: 652-677.

Urban, W. (2010). Customers’ Experiences As A Factor Affecting Perceived Service Quality.

Economics and Management. 15: 820-826.

(32)

Van Pham, K., Simpson, M. (2006). The Impact of Frequency of Use on Service Quality Expectations: An Empirical Study of Trans-Atlantic Airline Passengers. The Journal of

American Academy of Business, Cambridge. 10(1).

Vázques-Casielles, R., Suárez Álvarez, L. & Díaz Martin, A. 2010. Perceived justice of service recovery strategies: impact on customer satisfaction and quality relationship.

Psychology & Marketing. 27(5); 487-509.

Velázquez B., Blasco M., Saura I. & Contri G. (2010). Causes for complaining behavior intentions: the moderator effect of previous customer experience of the restaurant. Journal of

Services Marketing. 24 (7): 532-545.

Wu, H.J. & Liang, R. D. (2009). Effect of experiential value on customer satisfaction with service encounters in luxury-hotel restaurants. International Journal of Hospitality

Management. 28(4): 586-593.

Zhang, L., Lam, L., Chow, C. (2009). Segmenting the customer base in a CRM program according to customer tolerance to inferiority – a moderator of the service failure-customer dissatisfaction link. Journal of Consumer Satisfaction, Dissatisfaction and Complaining

(33)

Appendix

A: Questionnaire

The questionnaire was developed in Dutch, since it was directed at the customers of a Dutch bus company. It was used to measure the constructs. It is referred to in chapter 3.

Reizen met de bus

Bedankt voor uw deelname aan dit onderzoek!

Voor deze vragenlijst is het van belang dat u (zo goed mogelijk) terugdenkt aan uw MEEST RECENTE busrit.

Er komen verschillende stellingen voorbij over verschillende aspecten van deze busrit. Daarnaast worden enkele vragen gesteld over uw ervaring met de bus in het algemeen.

Door het voltooien van deze vragenlijst maakt u kans op een gratis dagkaart voor de bussen van Qbuzz in Groningen en Drenthe. Meer informatie hierover vindt u aan het einde van de vragenlijst.

Reisfrequentie

Hoe vaak per maand maakt u gemiddeld een busrit?

Heen en terug zijn twee ritten. Kies 0 wanneer u minder dan 1 keer per maand met de bus reist.

………

Het voertuig

Hieronder volgen enkele stellingen omtrent het voertuig (de bus) tijdens uw meest recente busrit. Kies telkens in hoeverre u het eens of oneens bent met de stelling.

Het interieur van de bus was aantrekkelijk/zag er goed uit

Oneens 1 2 3 4 5 6 7 Eens

Het geluid van het voertuig was vervelend

Ik was tevreden met de informatievoorziening in de bus

(34)

De zitplaatsen in het voertuig waren comfortabel

De bus was toegankelijk

Ik was tevreden over het klimaat in de bus

Ik kon gemakkelijk een zitplaats vinden, er was voldoende ruimte

De rit

De vertrektijd van de bus was acceptabel

De bus vertrok op de afgesproken tijd

De bus bracht mij op tijd bij mijn bestemming

De bus kwam op de afgesproken tijd aan

Ik zou willen dat de vertrek- of aankomsttijd van de rit anders was geweest

De busrit was betrouwbaar

m.b.t. vertrek- en aankomsttijd

De buschauffeur

(35)

De buschauffeur was vriendelijk tegen mij

Ik stoorde me aan het gedrag van de buschauffeur

De buschauffeur zorgde ervoor dat ik minder genoot van de rit

De rijstijl van de buschauffeur was goed

Ik voelde me veilig tijdens de rit

Medepassagiers

Het gedrag van mijn medepassagiers was gepast

Mijn medepassagiers waren vriendelijk

Mijn medepassagiers gedroegen zich acceptabel

Het gedrag van mijn medepassagiers was plezierig

Klantbeleving

(36)

De reis was de beste die ik me kon indenken

De reis had een hoge kwaliteit

De reis verliep goed

De busrit maakte mij relaxed

De busrit maakte mij gestrest

Ik had vertrouwen in de reis

De busrit maakte mij enthousiast

Ik had genoeg van de reis

Klantbeleving

Hieronder kunt u een rapportcijfer geven voor de busrit

1 2 3 4 5 6 7 8 9 10

Ervaringen in het verleden

Mijn ervaringen met busritten in het verleden zijn positief

Mijn ervaringen met busritten in het verleden zijn gunstig

(37)

Ik ben over het algemeen tevreden met de busritten die ik in het verleden heb gemaakt

Ik ben over het algemeen blij met de busritten die ik in het verleden heb gemaakt

Over u

Tot slot verzoeken we u enkele vragen te beantwoorden die beschrijven wat voor soort reiziger u bent. Ik ben een Vrouw Man Mijn leeftijd ……….

Mijn meest recente busrit was met:

Qbuzz Groningen/Drenthe Qbuzz Zuidoost Friesland Arriva

Anders: …………

De busrit was:

In de stad In de streek

Van stad naar streek (of andersom)

Wat was het reisdoel van uw meest recente busrit?

Onderstaande antwoorden gelden ook voor de terugweg (van school naar huis wordt dus “school”) Werk School/studie Sport Winkelen Bezoek vrienden/familie Tandarts/dokter/ziekenhuis Anders:………

Had u deze reis ook op een andere manier kunnen maken?

(38)

Ja Nee

Einde van de vragenlijst

Onder de deelnemers worden 5 ‘vrij reizen’ kaartjes verloot, waarmee u een dag lang onbeperkt kunt reizen in de bussen van Qbuzz Groningen/Drenthe en Zuidoost Fryslân. U kunt de vragenlijst afsluiten door op ‘verzenden’ te klikken.

Ik wil kans maken op een ‘vrij reizen’ kaartje

Ja Nee

Ik wil per e-mail geïnformeerd worden over de resultaten van dit onderzoek

Ja Nee

E-mailadres

(39)

B. Factor loadings

Rotated component matrix

1 2 3 4 5

The interior of the vehicle was attractive _,100 _,717 _,026 _,066 _,076 The noise in the vehicle was bothersome _-,063 _,203 _,019 _,024 _-,397 The information supply in the bus was

appropriate

,072 ,727 ,168 ,045 -,037 The cleanliness of the vehicle was appropriate _,023 _,699 _,129 _,069 _,122 The seating in the vehicle was comfortable _,263 _,668 _,033 _,012 _-,180 The accessibility of the vehicle was appropriate _,024 _,648 _,291 _,199 _,085 The temperature in the vehicle was to my liking _,229 _,562 _,298 _,002 _,073 I could find a place to sit with ease _,210 _,354 _,130 _,114 _-,147 The timeliness of the drive was acceptable _,172 _,068 _,138 _,840 _,051 The bus drive brought me to my destination on

time

,271 ,128 -,016 ,797 ,085 I wish the departure and arrival time of the trip

had been different

,075 ,009 -,173 -,622 -,130 This bus drive was consistent and reliable _,245 _,276 _-,011 _,810 _,087 The bus driver created a positive atmosphere _,433 _,332 _,079 _-,046 _,422 The bus driver was friendly to me _,360 _,215 _,065 _,062 _,589 I took offense to the behavior of the bus driver _,043 _,154 _-,161 _-,175 _-,710 The bus driver made this drive less enjoyable _,068 _-,029 _-,022 _-,096 _-,764 The driving style of the bus driver is appropriate _-,006 _,374 _,064 _,198 _,369

I felt safe during this drive _,004 _,418 _-,050 _,074 _,587

The behavior of the other passengers was appropriate

,042 ,049 ,931 ,054 ,053 The other passengers were friendly _,122 _,209 _,758 _,121 _,094 I found that the other passengers behaved well _,076 _,189 _,852 _,052 _-,022 The other passengers’ behavior was pleasant _,053 _,284 _,832 _,103 _,069 Your overall evaluation on past bus drives is

positive

,925 ,040 ,051 ,021 ,135 Your overall evaluation on past bus drives is

favorable

The relative importance of the service clues and the influence of past experience and frequency of use

The Clues to the

Customer Experience