• No results found

The effect of consumer reviews on the consumer’s physical store choice

N/A
N/A
Protected

Academic year: 2021

Share "The effect of consumer reviews on the consumer’s physical store choice"

Copied!
57
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The effect of consumer reviews on the

consumer’s physical store choice

Master Thesis Business Studies

University of Amsterdam, Amsterdam Business School Executive Programme Business Studies

Marketing Track

July 2014

Erwin de Haan (10282629) Supervisor: F. Situmeang

(2)

Colophon

Faculty: University of Amsterdam, Amsterdam Business School Executive Programme Business Studies

Specialization: Marketing Track

Title: The effect of consumer reviews on the consumer store choice

Student name: Erwin (W.E.) de Haan Student ID no.: 10282629

Supervisor: F. Situmeang

(3)

Abstract

The objective of this research is to examine the relation between consumer reviews about physical stores and the store choice decision process. By means of a literature study and a survey the following main question is answered: ‘What is the effect of consumer reviews on the consumer’s physical store choice?’ Furthermore the research was set up to discover to what extent the effect of reviews are affected by familiarity with the store, involvement with product purchase, type of buyer and the demographic gender and age.

First of all, the literature study showed the effect of reviews in other industries and online shopping. This gives insight in the search and selection process of the store choice of consumers. In addition, this gives an overview of the difference between the effect of reviews on familiarity and level of involvement and finally the differences between type of buyers, age and gender. The outcomes of the literature are tested with a pre-test (among 888 consumers) and a survey amongst 1032 consumers. A 2 (familiarity) x3 (involvement) between subjects experimental design was applied to investigate the effect of consumer reviews on the store choice decision. The results of the survey are analysed by means of SPSS.

The conclusion is drawn that, although reviews had no significant effect on the store choice, 33% of the consumers used reviews in their store choice decision process.

Furthermore, the use of reviews significantly differ in the amount of involvement with the purchase, gender and age. Whereby high involvement consumers, younger consumers and female consumers use reviews the most. The most important use of reviews is for imaging the store. This is especially the case for familiarity with the store, type of buyer, age and gender. In the case of unfamiliarity with the store, price/product/location buyers, females and older consumers get the most out of reviews.

(4)

Content of the thesis

1. INTRODUCTION 6 1.2 PROBLEM STATEMENT 7 2. LITERATURE REVIEW 8 2.1 IMPACT OF REVIEWS 8 FUNCTION OF REVIEWS 9

2.2 STORE CHOICE PROCESS 10

2.3 FAMILIARITY 11 2.4 INVOLVEMENT 13 2.5 GENDER 14 2.6 AGE 14 2.7 CONCEPTUAL MODEL 15 3. METHOD 15

3.1RESEARCH STRATEGY AND DESIGN 15

3.2 PROCEDURE 16

3.3 RESPONDENTS 17

3.4 PRE-TEST 17

3.5 UNDERLYING CONSTRUCT 19

3.6 MEASUREMENT 21

3.6.1 STORE CHOICE ASPECTS 21

3.6.2 KIND OF BUYERS 21 3.6.3 WAY OF USE OF REVIEWS 21 3.6.4 INVOLVEMENT 21 3.6.5 DEMOGRAPHICS 22 4. RESULTS 22 4.1 ANALYSIS 22 4.2TESTING HYPOTHESIS 23 4.2.1HYPOTHESIS 1 23 4.2.2HYPOTHESIS 2 23 4.2.3HYPOTHESIS 3 25 4.2.4HYPOTHESIS 4 27 4.2.5HYPOTHESIS 5 29 4.2.6HYPOTHESIS 6 30

(5)

5. DISCUSSION 33

5.2MANAGERIAL IMPLICATIONS 34

5.1LIMITATIONS AND FUTURE RESEARCH 35

REFERENCES 37

APPENDIX 41

1. RESULTS FACTOR ANALYSE PRE-TEST 41

2. RESULTS FACTOR ANALYSE STORE CHOICE ASPECTS 41

3. QUESTIONNAIRE PRE-TEST 42

(6)

1. Introduction

People have shared experiences with each other since time immemorial. We share experiences everywhere, on school playgrounds, at the coffee machine, in the café, at birthdays. We share experiences about everything, from products to services, but also about stores. We call these experiences word-of-mouth communication: "oral, person-to-person communication between a receiver and a communicator whom the receiver perceives as noncommercial, regarding a brand, a product, a service or a provider" (Arndt, 1967).

Times change and technologies evolve. This influenced the way we communicate with each other and has ensured that we now also share our experiences online. This happens on blogs, social media but also increasingly via review platforms. These appearances of

consumer-generated content are growing in importance and use (Gretzel & Yoo, 2008). If we focus on sharing online experiences, we call it reviews. The electronic version of traditional word of mouth (e-WOM). One of the functions of reviews is acting as a free ‘sales assistant’ for consumers, to help them identify the best products that fit their preferences. While seller-created information is mostly focused on attribute information (Chen & Xie, 2008), a

consumer review delivers user-orientated information which is generated and written from a consumer’s perspective. Therefore, it is generally perceived with more credibility and as influential, seller-created information. (Bickart & Schindler, 2001).

The evolution of the internet has also influenced consumer spending and with that their shopping behavior. More and more people buy their products online. Online spending has been growing for years, although there is a little stagnation in growth in the last years.

Regarding previous period 2005 2006 2007 2008 2009 2010 2011 2012 2013

Growth online shopping 32% 28% 38% 27% 17% 11% 9% 9% 8,5%

Obviously this is to the detriment of physical stores. An industry expert predicts a necessary closure of 30,000 store (Quix, 2013) of the total of 106.100 store in the Netherlands (data from HBD, 2012). On the whole this creates the key question for retailers: ‘how shall I attract people to my store’. Part of this question is: ‘why should consumers choose for my store instead of my competitor’. Throughout the whole purchase-decision process a lot of elements play a role, one of the elements is experience of other consumers.

(7)

1.2 Problem statement

Consumer reviews are a well-known and recognized medium in online sales. It has proven its worth many times over. This is also evident from prior research. There is, among others, research done into the function of reviews (Park & Lee, 2008) and the difference between consumer/user-generated and professional/seller-created reviews (Bickart and

Schindler, 2001; Chen and Xie, 2008). But also into the influence of an increase in star-scores on the growth of revenues (Ye et al. 2009). Furthermore, there is also research done into the difference in degree of impact between only stars and arguments/explanation (Chevalier and Mayzlin, 2006) or the influence of negative reviews (Berger et al. 2010). For example, for travellers, 25% of infrequent leisure travellers and 33% of frequent travellers changed a hotel stay based on reviews by other consumers (Gretzel and Yoo, 2008). Or an improvement of 1 point or 10% in stars of the reviews can lead to 4% increase in sales for videogames (Ye et al. 2009). Most of the research focused on the influence of reviews on consumer behavior, their purchase decision and the impact on revenue for only online sales or products. Unfortunately at this moment there is no clarity about the influence of reviews in the offline world in the choice of a store. The same applies for research about the influences of consumer shared experience on the decision making process of consumers in selecting or choosing of a

physical store. Hereby only one study is done into the traditional forms of word of mouth and their impact (Rigopoulou et al., 2008). The fact that there are more and more websites that focus on consumer reviews for physical stores could indicate that there is a need by consumers. Of course these positive effects of reviews on online sales and products by retailers create the following questions:

• What could the influence/impact of reviews be on the store choice of consumers? • What is the relation between the store selection aspects of consumers in combination

with the use of online consumer reviews about stores on their final store choice? • How could consumer store reviews help retailers to reach and to attract consumers to

their shop?

The aim of this research is to examine the relation between consumer reviews about physical stores and the store choice decision process. In addition the sub aims are to examine if this relation is affected by kind of buyer, familiarity with the store, the involvement with the purchase; gender and age. Based on the problem statement, the following research question has been formulated:

(8)

What is the effect of consumer reviews on the consumer’s physical store choice?

To what extent is this affected by the familiarity with the store and involvement of the purchase, type of buyer, age and gender?

The paper is divided into different sections. The paper starts with the relevant theoretical background in which previous research conducted is presented. In the second section the methodology is described, followed by the results and the related discussion. Finally, the managerial implications as well as the limitations of the study accompanied by recommendations for further research are included in the closing section.

2. Literature review

2.1 Impact of reviews

The world wide web has radically transformed the way we shop and thereby the way we share product consumption experiences. Word-of-mouth (WOM) is considered to be one of the most influential resources of information exchange. However, the impact has changed. When traditional Word of Mouth was limited to a local social network, online reviews can be used by everybody via the internet. Where a satisfied customer may tell some people about his experience with a company, a dissatisfied one will tell everybody he meets (Chatterjee, 2001). E-WOM (electronic Word of Mouth) is not a recent hype. The recommendations and discouragement of products and shops become common.

Reviews are already used frequently, very successfully, in various industries. For several years now reviews have been one of the most important influencers of the decision for a holiday destination. In online travel 78% of travelers actively read reviews of other travelers while planning their trip (Gretzel & Yoo, 2008). Also, the use and impact of reviews by purchasing products could no longer be ignored, it has become an indispensable tool for consumers (Pan & Zhang, 2011). Chevalier & Mayzlin (2006) found there is a relation between the use of reviews and the sales of books. While Ye, Law & Gu (2009) even showed that an improvement of 1 point or 10% in stars of the reviews can lead to a 4% increase in sales for videogames. Furthermore, online consumer reviews are also common for many other

(9)

product categories such as electronics, movies, music, beverages and wine (Zhu and Zhang, 2006; Chen and Xie, 2008).

However, there is discussion about what elements are responsible for the effect of reviews. According to Chevalier & Mayzlin (2006) the number and average star score of reviews is responsible for the effect. While Duan, Gu and Whinston (2008) claimed that consumers are not influenced by the persuasive effect (the rating) of online word of mouth, but that they are affected by the awareness (number of reviews) effect of reviews. Park, Lee and Han (2007) confirmed that the purchasing intention of consumers increases along with the number of reviews, because it indicates the popularity of a product. However, they also found that the quality of consumer reviews has a positive effect on consumers purchasing intentions. A qualitative ‘good’ review is logical and persuasive, with sufficient reasons based on

specific facts about the product. Unfortunately, not always all reviews are positive. Although it does not necessarily mean that negative publicity is bad. Sometimes it can be positive, because it generates awareness and accessibility about an unknown product (Berger, Sorensen, and Rasmussen 2010). However, the impact of negative reviews hurts more than positive reviews help to increase revenue. (Basuroy, Chatterjee and Ravid 2003).

Chen and Xie (2008) state that displaying consumer reviews can benefit products with a sufficient number of novice consumers, but can hurt the seller if the segment of expert consumers is relatively large. Thereby reviews are especially important for unsophisticated consumers, because it is less likely that they will buy a product if there is only seller-created information available (Chen & Xie, 2008).

Although some findings and views of researches are diverged, the common conclusion is that reviews have a positive influence on the choice of consumers. Based on that I expect that the influence of reviews have a positive correlation with store choice of consumers

H1: Reviews influenced the store choice decision

Function of reviews

To gain insights in the influence of reviews on the store choice process of consumers it is important to know what the function of reviews is. As mentioned earlier in this research, one of the functions of reviews is it acts works as a free ‘sales assistant’ for consumers. The research of Chen and Xie (2008) showed that it assists consumers in identifying the best

(10)

products that fit their preferences. (Chen & Xie,). It delivers additional user-oriented information and it provides a positive or negative signal on product popularity, with as purpose to learn about it and reduce uncertainty (Park & Lee, 2008). Herewith reviews have two functions, they work as an informant and a recommender.

Gretzel and Yoo (2008) found reviews were used for different reasons in different stages of holiday trip planning. The most important function was to narrow down the choices, but many also used reviews to get inspired. Travellers’ reviews are also used to confirm decisions. Which is reflected in the use of reviews, namely: to learn about the destinations, look for alternatives, provide them with ideas and warn for bad places or services (Gretzel & Yoo, 2008).

2.2 Store choice process

The store choice process of consumers is unfortunately not a straightforward process. There is already a lot of research done into the motivations of consumers regarding their shopping behavior. In order to understand this behavior it is necessary to identify and target ‘how consumers most likely purchase’ (Pan and Zinkhan, 2005). However, this is a dynamic process in which, for the consumer, the central question is: ‘where and when to

shop’(Leszczyc, Sinha, and Timmermans, 2000). However, answering this question is not easy, because of the changing consumer shopping behavior which is influenced by a lot of facets. According to Simonson and Rosen (2014) the purchase decisions is affected by a combination of three sources: the prior preferences, beliefs and experiences of consumers, marketing activities and information from other people. The greater the reliance with one of the sources, the lower the need of other sources. Arnold et al. (1996) went one step further and found, based on different researches, that the store choice is modeled on the importance and perception of store attributes, -attitudes and shopping patterns. Supplemented with personal factors, like buyer demographic, socio-economic characteristics and situational influences. When we focused on importance and perception of store attributes, this had a lot to do with the perceived store image of consumers. To gain more insight in this field Pan and Zinkhan (2005) did a meta-analysis and the results are used frequently in other researches. According to their analysis the most important factors for consumer retail choice are (in order of importance): assortment, service, product quality, store atmosphere, store location, price level, checkout speed, hours of operation, friendliness of salespeople, and parking facilities.

(11)

Of course the importance of these factors will differ among consumers depending on their shopping style and orientation (Rigopoulou, Tsiotsou & Kehagias, 2008). The setup and the interpretation of these aspects define the attractiveness of a store. This pallet of physical aspects in combination with the service and helpful and knowledgeable salespeople create the shopping experience for consumers. In addition, store name also influenced the perceived image of a store in the mind of consumers (Grewal et al., 1998). In the end, consumers will choose the store with the most favorable perceived image (Erdem, Oumlil and Tuncalp, 1999).

Rigopoulou, Tsiotsou & Kehagias (2008) suggest that for a reliable and thus useful approach of understanding consumers’ values, attitudes and reflection on store choice criteria, it is necessary to use shopping orientation-based segmentation. They segmented the aspects based on the difference in importance of store choice aspects for consumers. One group consists of product/price and the other on servicescape/personnel aspect are critical for the store choice. Based on these aspects preferences consumers are driven to the desired loyalty and patronage behavior. In addition, Grase and Cass (2004) substantiate that the key driver in the choice for a discount store is the perceived value for money, while for department stores these are consumer feelings. Consumers place greater importance on the quality of the service than on the costs associated acquisition of it (Cronin et al. ,2000). This is supplemented by the findings of Sirohi et al. (1998) that service quality has the most impact on store loyalty

intentions, while there is no direct impact of price on store loyalty intentions. This means that the willingness to recommend the store to others depends on the service quality. However, at the moment there is a lack in the literature of a link between the two segment (product/price and servicescape/personnel) and the use and impact of store reviews. Nevertheless I expect, based on the interpretation of the findings, that reviews are more often used by service

consumers and thereby it influences the store choice more than product/price consumers. This is because they attach more importance to the different aspects than just price and product.

H2: The effect of consumer reviews on the store choice will be higher when consumers

were service buyers than when they are product/price buyers

2.3 Familiarity

In shopping decision making, like any other option, we want to ensure that we make the right choice, hereby it is important to minimize the risk of making a wrong decision as

(12)

much as possible. One of the possibilities to minimalize the perceived risk is to choose for something familiar. Familiarity reflects the amount of information the consumer already has or knows. When a consumer is more confident they are more likely to purchase (Baltas, 1997). That is one of the reasons that when we like something we tend to return to what is familiar. This also applies for choosing and booking a holiday destination. When first-timers start collecting information for travel planning, they will more easily use and will also rely more on their friends, family and professionals. While repeaters in the search for a travel destination fall back on their own experiences instead of searching for other information sources. (Li et al., 2008). This is comparable to product familiarity. Especially for inexpensive and frequently bought products, familiarity is enough to base a choice on (Baltas, 1997). These earlier experiences affected the decision process and search behavior. The higher the level of familiarity the more the search for information will decrease. The cause for the limited external search could be explained by the knowledge of important attributes and brand-specific facts that experienced consumers already know (Johnson & Russo, 1981). On the other hand, Woodside and Trappey (1999) found that if you don’t have this evocation with a brand, the opportunity that consumers will choose for that brand is almost zero.

The online purchases intentions are strongly influenced by the influence of familiarity and trust. Although both are two different constructs, trust affected familiarity significantly (Gefen, 2000). Hereby is the impact of perceived trust on purchase decisions is stronger influence for potential customers as compared to that of repeat customers (Kim, Xu & Gupta, 2010). The same applies for visiting a new store or buying a brand for the first time, in these circumstances case we showed a higher level of information search. Meanwhile repeaters don’t start extensive information search (Sinhaa & Uniyal, 2005).

Vermeulen and Seegers (2009) found out that familiarity with a hotel makes

consumers resilient to the effects of online hotel reviews. The effects of reviews are stronger for lesser-known hotels than for well-known hotels. It affects more the awareness of a hotel and the persuasive effect of online reviews is bigger for lesser-known hotels. The same applies to games according to Zhu & Zhang (2010). They found that that online reviews are more influential for less popular games. They suggest when information offer are relatively scarce, the informational role of reviews becomes more salient. This suggest that Word-of-mouth information search is greater when a consumer is unfamiliar with a service or product or shop (Chatterjee, 2001). Based on this fact I expect that consumers which are familiar with

(13)

a store are lesser influenced and used in lesser extent consumer store reviews than unfamiliar consumers.

H3: The effect of consumer reviews on the store choice will be higher when consumers

are unfamiliar with the store than when they are familiar

2.4 Involvement

The average consumer makes many decisions a day. We always want to ensure that me make the right decision, therefor we attempt to minimize the risk of making a wrong decision. Of course not every choice will have the same importance or impact on the life of consumers, only a few of the decisions they make per day may be important (Zaichkowsky, 1985) . This depends on their needs, values, and interests. The individual differences explain why involvement with a choice affected the decision process (Josiam et al, 2004). Kenhove et al. (1999) found that the store choice decision depends on the task the shoppers intended to do. This varies between a shopping trip to get ideas to high urgent purchases. For

high-involvement purchases, which are considered important or personally relevant, the consumers seek for information about the choice and alternatives (Josiam et al, 2004). The store selection process is ultimately influenced by the perceived risk, which is coherent for expensive and infrequently purchased items (Dash et al., 1976). While on the other hand, consumers seek little or no evaluative information for low-involvement purchases. The greater the degree of risk aversion is, the greater the degree of information gathering will be (Evanschitzky and Woisetschläger, 2008). Thus the involvement in the shopping choice is influenced by the types of searching, our information-processing and decision-making process (Josiam et al, 2004). The shopping intentions have significant impact on the store choice, the satisfaction and perceived service quality of a consumer. Therewith also on the store loyalty and switching patterns. As an example, for specialty goods (a high involvement purchase) consumers are more likely to rely on salespeople’s expertise to help them compare among alternatives (Pan & Zinkhan, 2005). The conclusion is that the involvement-character of a particular product category explains the difference in importance of source of information (Rigopoulou et al. (2008). Based on that fact I expect that consumer reviews have significant higher impact on store choice for high involvement products.

H4: The effect of consumer reviews on consumer’ store choice will be higher by high

(14)

2.5 Gender

Males and females have different information needs. Although males buy more online than females, this does not mean that they use reviews in greater extent (Li et al.,2008). Females used a wider variety of online and offline information sources and they are also more involved in online information search (Kim et al., 2007). Furthermore, they are more likely to engage in deeper information processing by searching all available media for the required information (Okazaki and Hirose,2009). This underpinned and supported that females are high involvement with shopping and selecting restaurants (Josiam et al, 2004). When we focus on reviews, Gretzel and Yoo (2003) found that females get greater benefits from using reviews than males. Females see the benefit, especially in terms of enjoyment and idea generation, of reviews. Based on these results I expect that females use reviews in greater extent and are more influenced by reviews than males.

H5: The effect of consumer reviews on the store choice will be higher for females than

for males

2.6 Age

There were differences between ages in the way the consumers gather, process and use information. Younger consumers were more likely to be influenced by WOM (Gretzel and Yoo, 2003), they were more likely to seek information from friends and relatives. This aspect shows the biggest difference between the behaviour of the various groups. The use of this source is by consumers under-40s, 76%; 64% of 40-49 years and 49% of over-50s (Simcock et al., 2010). Nevertheless, traditional word-of-mouth communication through satisfied customers is the most important aspect that influences elderly people’s decision (Patterson, 2007). When this information came, not from familiar people or relatives, but from other consumers, we see differences with age. 50% of the younger group consumers and 33% of the older group used information from consumer reports (Simcock et al., 2010). When we focus on reviews, Gretzel and Yoo (2003) supported these findings. They found younger travellers use reviews more than older respondents (65 years or older) and they also found reviews are more important for deciding where to go. One of the reasons why older people reduce their information gathering and use already known alternatives, is when they want to averse the risk of a choice (Evanschitzky and Woisetschläger, 2008). Based on these facts I expect that

(15)

younger consumers use reviews in greater extent and are also more influenced by reviews than elder consumers.

H6: The use of consumer reviews on the store choice will be higher for younger

consumers than for older ones

2.7 Conceptual model

The variables described above are covered in this research. The relationships between these variables and the hypotheses concerning these variables are shown in the following conceptual model:

3. Method

3.1 Research strategy and design

To answer the research question and the hypothesises an online consumer survey is conducted. The research is exploratory and descriptive. For answering the research questions I used a qualitative data approach. The questions in the survey are partly derived from existing questionnaires in earlier researches, this causes a higher validity because the questions are tested earlier.

A 2x3 between subjects experimental design was applied to investigate the effect of consumer reviews on the store choice decision. Two factors are manipulated (involvement

Consumer store reviews Store choice

Kind of buyer Familiarity Involvement

Age Gender

H1

H5 H6

(16)

and familiarity), with two levels of each variable (high/low involvement and familiar/non familiar). A control group is added to measure the involvement under ‘medium’ involvement. This leads to six experimental conditions (table 1).

Table 1: Experimental design, N (%)

Familiair Non-familiair High involvement (specialty good) 632 (60.77%) 408 (39.23%) Control group (shopping good) 632 (60.77%) 408 (39.23%) Low involvement (convenience good) 632 (60.77%) 408 (39.23%)

3.2 Procedure

Respondents were asked via e-mail to participate in this scientific experiment. For participating in the research they filled out a self-administered questionnaire in Dutch. On the introduction page respondents were informed about the scientific design of this research and the design of the questionnaire. The questionnaire started with a question about the familiarity with selected stores in the three different industries. When they were, in each industry, familiar with three stores they were assigned to the ‘familiar’ part of the research. Otherwise they participated on the ‘unfamiliar’ version of the research. The questionnaire was designed as follows: respondent had to read three times, a story about the different stores and after reading the stories they had to answer some questions.

For the ‘familiar group’ we used the stories about the stores they had selected as ‘most familiar with’. For the ‘unfamiliar’ group the same stories were used but only with standardized names (retailer A, retailer B, etc.). The stories were based on the aspects which consumers used to choose a store and were supplemented with two consumer reviews per store. The stories about the stores where based on real information of the websites of the stores and adapted with some information to make equal images of the stores. Furthermore to all of the stories we added two real reviews about these stores from reviewplatform Wugly.nl (Dutch reviewplatform about physical stores). After reading the three stories about the stores in one industry, the respondents where asked which store they would choose. The purpose was simulating the orientation and decision making process of the respondents. After their choice they were asked why they chose for that specific store. Therefore they had to answer what influence the antecedents had on their store choice. Furthermore, they had to rate in which way they were influenced by the reviews. Here, the respondents had to answer to what extent they agreed with the 11 propositions about the use of reviews. This process was

(17)

repeated three times after each industry. Finally, we asked the respondents about their age, gender, education and income level.

3.3 Respondents

The sample in this study consisted of Dutch consumers. For the research on the consumer I used the panel of Q&A Research & Consultancy, a Dutch Research company specialized in online research for retail. Respondents were selected randomly and asked via e-mail to participate in this scientific experiment. For the research they filled out a

self-administered questionnaire. Participation was voluntary and participants did not receive any compensation for their participation. A total of 1.080 participants completed the full

questionnaire of which 1.032 are included in the future analysis. In total 47.6% of the

respondents were male and 52.4% female. 10.2% of them were aged between 16 and 30 years, 18.4 % between 31 and 45 years, 28.0% between 46 and 60 years and 21.6% was older than 60 years. The results were weighted and corrected for gender and age of the Dutch population, based on Eurostat data (statistical office of the European Union). The most important levels of highest education were HBO, MBO and WO, with respectively 34.6%, 25.9% and 12.6% of the respondents. The other results for education level were highly fragmented and low in interest. The income was more or less equal distributed between an income of €0 to more than €50.000.

3.4 Pre-test

In order to validate the scale for shopping instead of travelling, to reduce the number of items and to identify if there are underlying constructs for the way of use of consumer reviews by the consumer store choice process we set up a pre-test. For the test we used the results from Gretzel & Yoo (2003) research of the impact of reviews on travel planning. They asked: ‘in which ways do reviews of other travellers influence your travel planning?’ In this research, consumers had to indicate to what extent the 15 aspects were applicable to their preference and ultimate choice. They rated the items on a 5-point likert scale anchored ‘strongly agree’ to ‘strongly disagree’. For the pre-test these aspects were translated so they could be applied and used for ‘shopping’ instead of ‘travelling’.

The research was a short, self-administered, questionnaire with only two key questions in Dutch. The first question was to indicate if consumers have used consumer reviews for

(18)

their store choice decision in the past. This question was answered by 888 respondents. If they agreed they proceeded to the second question with the 15 aspects in which ways they have used reviews? This second question was answered by 421 respondents, 49,6% of them were male and 50,4% were female. The distribution of age was 22% between 16 and 30 years, 27% between 31 and 45 years, 33% between 46 and 60 year and 11% older than 60 years. The results where weighted and corrected for gender and age of the Dutch population based on Eurostat data (statistical office of the European Union).

A principal component analysis was used to reduce the number of aspects to

investigate the underlying structure of the items. Variables with factor loadings less than 0.55 were excluded from further analysis(see results in appendix X). Furthermore we calculated the reliability of the constructs with a Cronbach’s alpha per factor. The Cronbach’s alpha’s of the constructs needs to be above 0,60 to be reliable. All the factors have a Cronbach’s alpha higher than 0,6 and the removal of aspects had no positive impact on the factors. Factor 1, summarized as ‘imagining the store’, consists of 4 items (α=0.79) (see table 1 and 2). Factor 2, summarized as ‘supporting in the decision’ consisted of 4 items (α=0.74) (see table 3 and 4). Factor 3, ‘supporting by orientation’, with 3 items (α=0.65) (see table 5 and 6). The 11 aspects that are components of the three constructs will be used in the further research.

Table 1: Reliability Statistics

Cronbach's Alpha N of Items

0,791 4

Table 2: Item-Total Statistics

Scale Mean if Item Deleted Scale Variance if Item Deleted Total Correlation Corrected Item- Cronbach's Alpha if Item Deleted Are a good way to learn about a store 10,58 6,779 0,66 0,713 Make it easier to imagine what a store will be like 11,07 6,243 0,583 0,753 Make shopping planning more enjoyable 10,47 6,849 0,612 0,734 Help me imagine my shopping trips more vividly 10,76 6,865 0,561 0,759

Table 3: Reliability Statistics

Cronbach's Alpha N of Items

(19)

Table 4: Item-Total Statistics

Scale Mean if Item Deleted Scale Variance if Item Deleted Total Correlation Corrected Item- Cronbach's Alpha if Item Deleted Increase my confidence in the decisions I make 8,92 6,8 0,493 0,701 Reduce the risk/uncertainty involved in making

store choice decisions 9,17 6,318 0,565 0,662

Make it easier to reach decisions 9,7 6,116 0,514 0,692 Reduce the likelihood that I will later regret a

decision 9,86 6,062 0,559 0,664

Table 5: Reliability Statistics

Cronbach's Alpha N of Items

0,654 3

Table 6: Item-Total Statistics

Scale Mean if Item Deleted Scale Variance if Item Deleted Total Correlation Corrected Item- Cronbach's Alpha if Item Deleted

Provide me with ideas 6,66 3,614 0,427 0,607

Help me plan my shopping trips more efficiently 6,51 3,359 0,49 0,523 Help me save time in the shopping planning

process 6,73 3,272 0,479 0,538

3.5 Underlying construct

To test the relation between store choice aspects and the use of consumer reviews we used the aspects Rigopoulou et al (2010) compiled from other researches by Pan & Zinkhan (2006). Their analysis showed 20 antecedents which reflected the aspects of store choice criteria of consumers. These antecedents were categorized in to three groups: “Product/Price”, “Source of Information”, and “ServiceScape/Personnel”. In order to apply this to the Dutch market we removed four items, three of them because there is no difference between sales people and cashiers. The other one was removed because in most stores payment options in terms of number and time of instalments are not possible. Due to the focus of this research on consumer reviews, three antecedents of “Source of Information” have been adapted in two items for consumer reviews.

To these 14 antecedents we add three antecedents which are missing but are likely to be very important in the research of Pan & Zinkhan (2006). This aspect relates to a convenient location and openings hours. For testing the scale, to control the categorization and identity if review aspects form constructs, a Principal Component Analysis was carried out. Variables with factor loadings less than 0.50 were excluded from further analysis. This resulted in three

(20)

underlying constructs (see results in appendix 2). Furthermore, the reliability of the constructs was calculated with a Cornbach’s alpha per factor. Factor 1, summarized as ‘Personnel’, consists of 4 items (α=0,834) (see table 7 and 8). Factor 2, summarized as

‘product/price/location’ consisted of 7 items (α=0,795) (see table 9 and 10). Factor 3,

‘reviews’, with 2 items (α=0,845) (see table 11 and 12). This last factor consisted out of three aspects after the analysis, but when one of the items would be deleted the Cornbach’s alpha rose from 0,735 to 0,845, therefore the aspect is deleted out of the construct. All factors had a Cornbach’s alpha of more than 0.6 and are therefore used in the future analysis.

Table 7: Reliability Statistics

Cronbach's Alpha N of Items

0,834 4

Table 8: Item-Total Statistics

Scale Mean if Item Deleted Scale Variance if Item Deleted Total Correlation Corrected Item- Cronbach's Alpha if Item Deleted To have personnel that serve me fast 28,62 30,968 0,53 0,835 To have salespeople polite and friendly 27,46 32,009 0,675 0,791 To have salespeople that will understand my

needs 28,14 29,381 0,703 0,78

To be informed by skilled salespeople 27,91 29,893 0,709 0,779 To have a pleasant atmosphere and environment 27,73 33,675 0,582 0,815

Table 9: Reliability Statistics

Cronbach's Alpha N of Items

0,795 7

Table10: Item-Total Statistics

Scale Mean if Item Deleted Scale Variance if Item Deleted Total Correlation Corrected Item- Cronbach's Alpha if Item Deleted To have many and new models by brand 43,92 57,091 0,536 0,769 To have products ready to deliver 43,94 55,451 0,586 0,76 To have the lowest in cash prices 44,1 57,761 0,406 0,79 To have short delay time until I am served by a

salesperson 44,64 54,874 0,512 0,771

To have convenient parking facilities 44,61 49,448 0,504 0,781

To have convenient location 44,05 52,72 0,63 0,75

To have convenient opening hours 44,24 53,614 0,562 0,762

Table 11: Reliability Statistics

Cronbach's Alpha N of Items

(21)

Table 12: Item-Total Statistics

Scale Mean if Item Deleted Scale Variance if Item Deleted Total Correlation Corrected Item- Cronbach's Alpha if Item Deleted To have a variety of payment methods 9,78 21,959 0,372 0,845

To have been recommended 10,87 15,461 0,681 0,492

To have reviews from other consumers 11,12 15,996 0,651 0,532

3.6 Measurement

3.6.1 Store choice aspects

To test the relation between store choice aspects and the use of consumer reviews the results of the factor analysis on the store choice aspects will be used. The aspects were compiled from earlier research of Rigopoulou et al (2010) and Pan & Zinkhan (2006). In total 3 constructs consisting of 14 aspects were used for answering the scale. Items were rated on a 10-point scale anchored at “completely unimportant” and “extremely important”. An Example of such an item is: ‘to have salespeople that will understand my needs’.

3.6.2 Kind of buyers

For the segmentations of kinds of buyers the results of the factor analysis were used on the store choice aspects. This analysis was based on research of Rigopoulou et al. (2008) for the categorization of store choice criteria. In total two groups are formed, namely ‘Personnel’, consisting of 4 items (α=0,834) and ‘product/price/location’ consisting of 7 items (α=0,795).

3.6.3 Way of use of reviews

In order to identify in what way consumers use reviews in shopping and their store choice decision, in the research a scale consisting of 11 aspects was used. The results are derived from a research of Gretzel & Yoo (2003) about ‘in which ways reviews of other travellers influenced their travel planning’, but translated and tested for shopping instead of travelling. Items were rated on a 5-point scale anchored ‘strongly agree’ to ‘strongly

disagree’. An example of an item is: ‘reduce the risk/uncertainty involved in making store choice decisions’.

3.6.4 Involvement

In order to find out the influence of involvement on the consumer store choice we asked consumers to choose for three different types of products. The level of involvement refers to the amount of importance or interest which the consumer attaches to the type of product (Stell et al., 1996). For the classification of goods we used the definitions of

(22)

American Marketing Association's Committee. For convenience goods, ‘which the customer purchases frequently, immediately, and with minimal effort’, we selected grocery stores. For shopping goods, ‘which the customer, in the process of selection and purchase,

characteristically compares on such bases as suitability, quality, price and style’, we selected fashion stores. Finally, we choose for specialty goods: furniture, because this is a good example ‘on which a significant group of buyers are habitually willing to make a special purchasing effort’ (Bucklin, 1963).

3.6.5 Demographics

Demographics can also influence the dependent variable, store choice. For this reason age and gender are part of the research. Furthermore, we checked if education and income influence the effect of reviews on the store choice decision.

4. Results

4.1 Analysis

For the analysis of the data we used the SPSS program. The first step was to eliminate unreliable results. Therefore we checked whether specific outcomes deviated extremely from normal results. In total we removed 48 cases.

As mentioned above, the results of the six different groups participating in the study had to be compared on the mean of the influence of the aspects on the store choice decision and the use of reviews. For analyzing the influence of each aspect on the store choice decision, the results were divided into three groups: unimportant (1-4), neutral (5-6) and important (7-10). The same method was applied to the results of the aspects about the use of reviews. Because I used a 5 point scale, I divided the results as follow: disagree (1/2), neutral (3) and agree (4/5).

(23)

4.2 Testing hypothesis

4.2.1 Hypothesis 1

As hypothesis 1 states, consumers use reviews in their store choice decision. For the hypothesis I analyzed the results about the importance of store choice aspect. In order to answer the questions I searched for the mean score per store choice aspect. The aspects with the highest mean were assortment and the availability of the products. Followed by customer-friendly personnel and convenient location, the top five is completed with openings hours. The review aspects, ‘to have been recommended’ and ‘to have been reviewed’, scored respectively a mean of 5,02 (SD=2,526) and 4,77 (SD=2,511) on a 10-point scale on

importance in consumer’s store choice decision. The review aspects together form the reviews construct, which was rated with a mean of 4,89 (SD=2,343). In total 32,4% of the consumers rated the review aspect and rated the review constructs as important (7-10). The mean is lower than the important score. This means a negative score and therefore hypothesis 1 is not

supported.

To gain deeper insights in the use and impact of reviews on the store choice I analyzed the results of the question about the use of reviews. The results from the survey show that the top three of most important aspects consisted out of the factor ‘imaging the store’ with a mean of 2,72 (SD=1,027) on a 5-point scale. While the factor ‘supporting by orientation’ scored a mean of 2.53 (SD=1,063 ). This applies to the factor ‘supporting in the decision’ for 2,46 (SD=1,038). This also means a lower score than agree or totally agree (4/5).

4.2.2 Hypothesis 2

For testing the differences between personnel/service buyers (hereafter: service) and product/price/location (hereafter: product) buyers I used the results from the Principal Component Analysis. My expectation in hypothesis 2 was that the service buyers use

consumer reviews to a greater extent than product buyers. For the analyse I carried out a GLM (General Linear Model) Between-Subject effects test. For composing the groups I bundled the consumers that rated the product construct as important (6-10) and the service construct as not important (<6) in one group, called product buyers. The other group, consisting of consumers from the opposite, is called service buyers. The price group consist of 353 respondents, while 679 respondents were assigned to the service group. Although there is no normal distribution (34,2% vs 35,8%), it was necessary to use a non-parametric tests, because it could be assumed

(24)

that it is close to the population mean. The analysis showed that the mean is lower for service buyers (M=4,15, SD=2,13) in comparison to product buyers (4,43, SD=2,29). This means that the results are the opposite of the hypothesis. However the difference is not significant. All together I reach the conclusion hypothesis 2 is rejected.

Table 13: Tests of Between-Subjects Effects

Dependent Variable: impact of reviews

Source Type III Sum of Squares df Mean Square F Sig.

Corrected Model 10,946a 1 10,946 2,296 ,130 Intercept 17048,417 1 17048,417 3575,581 0,000 price _vs_service 10,946 1 10,946 2,296 ,130 Error 4911,054 1030 4,768 Total 23558,250 1032 Corrected Total 4922,000 1031

a. R Squared = ,002 (Adjusted R Squared = ,001)

To test if this contrast also applies to the underlying constructs of the way in which reviews are used, I replicated the test we used above. The results show that for all three constructs a significant difference between the service group and the product group(see table 14-16). However, the difference is in the favor of service instead of product buyers. This also means that on the underlying construct service buyers attach more value to reviews than product buyers.

Table 14: Tests of Between-Subjects Effects

Dependent Variable: Imagining_the_store

Source Type III Sum of Squares df Mean Square F Sig.

Corrected Model 7,472a 1 7,472 8,288 ,004 Intercept 6073,760 1 6073,760 6736,383 0,000 price _vs_service 7,472 1 7,472 8,288 ,004 Error 928,684 1030 ,902 Total 7534,500 1032 Corrected Total 936,157 1031

(25)

Table 15: Tests of Between-Subjects Effects

Dependent Variable: Supporting_by_orientation

Source Type III Sum of Squares df Mean Square F Sig.

Corrected Model 8,617a 1 8,617 8,799 ,003 Intercept 5156,761 1 5156,761 5265,892 0,000 price _vs_service 8,617 1 8,617 8,799 ,003 Error 1008,654 1030 ,979 Total 6598,667 1032 Corrected Total 1017,271 1031

a. R Squared = ,008 (Adjusted R Squared = ,008)

Table 16: Tests of Between-Subjects Effects

Dependent Variable: Supporting_by_orientation

Source Type III Sum of Squares df Mean Square F Sig.

Corrected Model 11,531a 1 11,531 12,303 ,000 Intercept 4929,927 1 4929,927 5259,983 0,000 price _vs_service 11,531 1 11,531 12,303 ,000 Error 965,369 1030 ,937 Total 6287,250 1032 Corrected Total 976,900 1031

a. R Squared = ,012 (Adjusted R Squared = ,011)

4.2.3 Hypothesis 3

To test for differences in the effect of reviews on the store choice decision between familiarity and unfamiliarity with the store, we compared the results of the two conditions mutually. The hypothesis suggests that the effect of reviews under unfamiliar conditions will be higher than under familiar conditions. For the analysis I carried out a GLM (General Linear Model) Between-Subject effects test. Although there is no normal distribution of the familiar/unfamiliar (60.77% vs. 39.23%) condition, it was necessary to use non-parametric tests. Because of the high N (632 vs. 408) we could assume that it is close to the population mean. The analysis showed that the mean under unfamiliar conditions (M=4.96, SD=2,28) was higher than under familiar conditions (M=4.85, SD=2,38). However, the differences between the groups was not significant (see table 4) and therefore hypothesis 3 is rejected.

(26)

Table 17: Tests of Between-Subjects Effects

Dependent Variable: Impact of reviews

Source Type III Sum of Squares df Mean Square F Sig.

Corrected Model 3,827a 1 3,827 ,691 ,406 Intercept 85168,227 1 85168,227 15386,810 0,000 Familiar_vs_Unfamiliar 3,827 1 3,827 ,691 ,406 Error 20236,490 3656 5,535 Total 108493,750 3658 Corrected Total 20240,317 3657

a. R Squared = ,000 (Adjusted R Squared = ,000) Although familiarity with the store does not affect the review construct it can

certainly affect the underlying constructs about the way reviews are used. For this analysis we replicated the test we used above for this hypothesis. The results show a significant difference between the unfamiliar group and the familiar group for all three constructs. For ‘imaging the store’ (M=2,88, SD=0,96 vs. M=2,62, SD=1,05) (see table 18), ‘supporting by orientation’ (M=2,61, SD=1,01 vs. M=2,47, SD=1,09) (see table 19) and ‘supporting the decision’ (M=2,57, SD=0,99 vs. M=2,38, SD=1,06) (see table 20) the unfamiliar group had a higher mean than familiar. This means that although there is no different influence between the groups, the way of use of reviews certainly differs between consumers who are unfamiliar or familiar with a store.

Table 18: Tests of Between-Subjects Effects

Dependent Variable: Imagining_the_store

Source Type III Sum of Squares df Mean Square F Sig. Corrected Model 53,904a 1 53,904 51,197 ,000 Intercept 26719,257 1 26719,257 25377,435 0,000 Familiair_vs_unfamiliair 53,904 1 53,904 51,197 ,000 Error 3849,310 3656 1,053 Total 31181,813 3658 Corrected Total 3903,213 3657

a. R Squared = ,014 (Adjusted R Squared = ,014)

Table 19: Tests of Between-Subjects Effects

Dependent Variable: Supporting_by_orientation

Source Type III Sum of Squares df Mean Square F Sig. Corrected Model 14,479a 1 14,479 12,692 ,000 Intercept 23043,424 1 23043,424 20199,278 0,000 Familiair_vs_unfamiliair 14,479 1 14,479 12,692 ,000 Error 4170,781 3656 1,141 Total 27893,778 3658 Corrected Total 4185,259 3657

(27)

Table 20: Tests of Between-Subjects Effects

Dependent Variable: Supporting_the_decision

Source Type III Sum of Squares df Mean Square F Sig.

rrected Model 27,271a 1 27,271 25,095 ,000 Intercept 21731,667 1 21731,667 19997,979 0,000 Familiair_vs_unfamiliair 27,271 1 27,271 25,095 ,000 Error 3972,950 3656 1,087 Total 26269,125 3658 Corrected Total 4000,221 3657

a. R Squared = ,007 (Adjusted R Squared = ,007)

4.2.4 Hypothesis 4

Presented in hypothesis 4, the expectation is that consumers who are occupied with purchasing a product requiring higher involvement will use reviews in their store choice decision to a larger extent, compared to a purchase of a low involvement product. For the test we used the three levels of involvement. First of all we carried out a GLM (General Linear Model) Between-Subject effects test to discover if there are significant differences between the involvement levels. The results show a mean for high involvement of 5.21 (SD=2.37), for medium M=4.81 (SD=2.31)and for low involvement M=4.67 (SD=2.32).

This results were significant (F(3, 3653) = 16.129, p<0.001). However, the hypothesis states that there is a difference between high and low involvement levels. To analyze this difference between the group, a (post hoc) Tukey’s HSD was carried out. This shows significant differences between low and high and medium and high involvement (see table 21). This means that the influence of reviews is higher in the purchase of high involvement products than in low involvement purchases and this means that hypothesis 4 is accepted.

(28)

Table 21: Multiple Comparisons

Dependent Variable: Impact reviews

Tukey HSD

(I) Involvement Mean

Difference (I-J) Std. Error Sig. 95% Confidence Interval

Lower Bound Upper Bound

Low involvement Medium involvement -0,14 0,09446 0,299 -0,3615 0,0814 High involvement -,5225* 0,09507 0 -0,7454 -0,2996 Medium involvement Low involvement 0,14 0,09446 0,299 -0,0814 0,3615 High involvement -,3825* 0,09516 0 -0,6056 -0,1594 High involvement Low involvement ,5225* 0,09507 0 0,2996 0,7454

Medium involvement ,3825* 0,09516 0 0,1594 0,6056

Based on observed means. The error term is Mean Square(Error) = ,489.

*The mean difference is significant at the ,05 level

To comprehend in which ways the difference between the groups could be explained, an analysis was performed to find out whether there also are differences between the groups in the way they use reviews. For this analysis we replicate the GLM (General Linear Model) Between-Subject effects test on the three underlying constructs of reviews. The results from this test showed there were differences between the involvement groups.

However, the differences are not significant on the constructs ‘imagining the store’ (F(3, 3653) = 1.391, p=0.249) and ‘supporting the decision’ (F(3, 3653) = 0.114, p=0.891). The construct ‘supporting by orientation’ (F(3, 3653) = 4.322, p=0.013) it showed to be significant. However only on low involvement the difference is significantly with medium involvement and not high involvement (see table 22).

(29)

Table 22: Multiple Comparisons

Dependent Variable: Supporting_by_orientation Tukey HSD

(I) Involvement Mean

Difference (I-J) Std. Error Sig. 95% Confidence Interval

Lower Bound Upper Bound

Low involvement Medium involvement ,1216* ,04309 ,013 ,0206 ,2227

High involvement ,0297 ,04337 ,773 -,0720 ,1314 Medium involvement Low involvement -,1216* ,04309 ,013 -,2227 -,0206

High involvement -,0919 ,04341 ,086 -,1937 ,0098 High involvement Low involvement -,0297 ,04337 ,773 -,1314 ,0720 Medium involvement ,0919 ,04341 ,086 -,0098 ,1937 Based on observed means. The error term is Mean Square(Error) = 1,142. *. The mean difference is significant at the ,05 level.

4.2.5 Hypothesis 5

In hypothesis 5 is assumed that females use reviews to a greater extent than males. For the analysis we carried out a GLM (General Linear Model) Between-Subject effects test on the results of gender. The results show a mean for females of 5.06 (SD=2.30), while for males it was: M=4.73 (SD=2.37). These results were significant (see table 23). The results show that females use reviews to a greater extent than males and in addition the difference is also

significant, which means that the hypothesis is accepted.

Table 23: Tests of Between-Subjects Effects

Dependent Variable: Reviews

Source Type III Sum of Squares df Mean Square F Sig. Corrected Model 117,916a 1 117,916 21,424 ,000 Intercept 88364,738 1 88364,738 16054,818 0,000 Gender 117,916 1 117,916 21,424 ,000 Error 20122,401 3656 5,504 Total 108493,750 3658 Corrected Total 20240,317 3657

a. R Squared = ,006 (Adjusted R Squared = ,006)

In order to find out why females use reviews to a greater extent, I carried out an analysis on the underlying contrast of reviews. For this analysis we replicate the GLM (General Linear Model) Between-Subject effects test on the three underlying constructs of reviews. The difference between females (M=2.81, SD=1.00) and males (M=2.64, SD=1.04) is the most at the construct ‘imaging the store’, these results were significant (see table 24). Followed by ‘supporting by orientation’ where the difference between females (M=2.58,

SD=1.05) and males (M=2.47, SD=1.08) was also significant (see table 25). The difference

(30)

construct ‘supporting the decision’, however the results were significant (see table 26). This means that, in addition of the acceptation of the hypothesis, it is established that females use reviews to a greater extent, for all intents and purposes.

Table 24: Tests of Between-Subjects Effects

Dependent Variable: Imagining_the_store

Source Type III Sum of Squares df Mean Square F Sig. Corrected Model 26,256a 1 26,256 24,760 ,000 Intercept 27304,588 1 27304,588 25748,434 0,000 Geslacht 26,256 1 26,256 24,760 ,000 Error 3876,957 3656 1,060 Total 31181,813 3658 Corrected Total 3903,213 3657

a. R Squared = ,007 (Adjusted R Squared = ,006)

Table 25: Tests of Between-Subjects Effects

Dependent Variable: Supporting_by_orientation

Source Type III Sum of Squares df Mean Square F Sig. Corrected Model 14,348a 1 14,348 12,577 ,000 Intercept 23722,610 1 23722,610 20793,984 0,000 Geslacht 14,348 1 14,348 12,577 ,000 Error 4170,911 3656 1,141 Total 27893,778 3658 Corrected Total 4185,259 3657

a. R Squared = ,003 (Adjusted R Squared = ,003)

Table 26: Tests of Between-Subjects Effects

Dependent Variable: Supporting_the_decision

Source Type III Sum of Squares df Mean Square F Sig.

Corrected Model 4,636a 1 4,636 4,242 ,040 Intercept 22269,504 1 22269,504 20376,814 0,000 Geslacht 4,636 1 4,636 4,242 ,040 Error 3995,586 3656 1,093 Total 26269,125 3658 Corrected Total 4000,221 3657

a. R Squared = ,001 (Adjusted R Squared = ,001)

4.2.6 Hypothesis 6

In hypothesis 6, I described that based on the results from other research I expected that younger consumers will use reviews to a greater extent than older people. For the test we used four age groups. While there is no normal distribution of groups (16-30 years: 10.2%; 31-45: 18.4 %; 46-60: 28.0%; >60: 21.6%) condition, it was necessary to use a

non-parametric test, because it could be assumed that it is close to the population mean. To discover if there are significant differences between the age groups, I carried out a (post hoc) Tukey’s HSD. The results show differences between the oldest (M=5.09, SD=2.38) and the

(31)

youngest groups (M=5.36, SD=2.21). However, the mean is the lowest for the group 31-45 years (M=4.40, SD=2.40), followed by the group from 46-60 years (M=4.75, SD=2.26). The difference is significant (F(4, 3653) = 29.203, p<0.001).

The results (see table 27) show that all differences between groups were on a level 0.05 significant. This means that younger people used reviews to a greater extent in their store choice decision than older people and thereby that hypothesis 6 is accepted.

Table 27: Multiple Comparisons

Dependent Variable: Imapct of reviews

Tukey HSD

(I) Age groups Mean

Difference (I-J) Std. Error Sig. 95% Confidence Interval

Lower Bound Upper Bound

15-29 years 30-44 years 1,0378* 0,11909 0 0,7317 1,3439 45-59 years ,6047* 0,11698 0 0,304 0,9053 60+ years ,2862* 0,10645 0,036 0,0126 0,5599 30-44 years 15-29 years -1,0378* 0,11909 0 -1,3439 -0,7317 45-59 years -,4331* 0,11626 0,001 -0,7319 -0,1343 60+ years -,7515* 0,10567 0 -1,0231 -0,4799 45-59 years 15-29 years -,6047* 0,11698 0 -0,9053 -0,304 30-44 years ,4331* 0,11626 0,001 0,1343 0,7319 45-59 years ,3184* 0,10328 0,011 0,053 0,5839

Based on observed means. The error term is Mean Square(Error) = 5,410. *. The mean difference is significant at the ,05 level.

When focusing on the underlying construct of the way reviews are used shows that there are differences between the younger and the older consumers. The results show that the two older groups had a higher mean than the younger groups.

(32)

Furthermore, the difference between the youngest and the oldest group was significant (Mdif=-.1748, p=0.001). The same applies for ‘supporting the orientation’ (Mdif=-.2822,

p<0.001) and ‘supporting the decision’ (Mdif=-.1924, p<0.001). The results show that about

the way of contracts with results about the extent of use. This means when older people use reviews they are more influenced by it than younger people. Furthermore, the least influenced group are the consumers from 31-45 years old, followed by the youngest group.

(33)

5. Discussion

The objective of this research was to answer the main question: ‘What is the effect of consumer reviews on the consumer’s physical store choice?’ Furthermore, the research was set up to discover to what extent the effect of reviews are affected by familiarity with the store, involvement with product purchase, type of buyer and the demographics gender and age.

Based on a quick glance at the results on the first hypothesis you could conclude reviews do not play a role in the store choice decision of consumers for a physical store choice. Because the results show no positive and important weight for the review aspects by store choice. Although the hypothesis was rejected, one third of the respondents stated that consumer reviews played an important role in their store choice decision. This means that in one out of the three store choice decision processes consumer reviews are used. Which leads to the expectation that reviews should a not be underestimated in many store choices.

Furthermore, the results show that consumer store reviews are mainly used as informer by the store choice for consumer, not as recommender. The primary function is to imaging the store. Followed by supporting the orientation and in the last place it acts in supporting the decision.

Although reviews are more important in the decision process for unfamiliar consumers than consumers that already know the store, the effect is not significant. It is not surprising that the main difference between two groups is on the construct ‘imaging the store’. This is because reviews will help consumers learn more about the store without visiting the store, while when you already are familiar with the store imaging the store is no longer necessary. Also, for supporting the orientation and decision process unfamiliar consumers used reviews to a greater extent than familiar consumers. Based on these results it can be inferred that when unfamiliar consumers use reviews they will use these to a greater extent than familiar

consumers.

The type of product that consumers buy and the degree of involvement with it, is essential in the extent of use of reviews. The more involved consumers are with their

purchase, the more they use reviews. Especially in the purchase of high involvement products, this is in line with earlier research. Consumers want to minimalize the risk to make a wrong decision and therefore use, amongst other things, reviews. Meanwhile is not the case that the way of use of reviews shows the same pattern as the extent of use of reviews. It can be

(34)

concluded that the way of use of reviews is independent on level of involvement. This means that although the extent of use of reviews increases by a growing involvement, the way of use of reviews remains almost the same under all conditions.

Striking is that, in contrast to the expectation and earlier research, reviews are more used by product/price/location buyers in comparison to consumers who attach great

importance to service. This is may be due to the fact that imaging the store choice is the most important use of reviews. Product, price and location of course play a more important role in that compared to service and personnel.

In line with research about the search process of females our results confirm that females used reviews to a greater extent than males. This is not only confirmed by the use of reviews, but also in the way of use of reviews. Especially on imaging the store there is difference between females and males, this underpinned that females see the benefits, especially in terms of enjoyment and idea generation, of reviews (Gretzel and Yoo, 2003). The fact that females on all the constructs gave higher scores on importance showed that females are more engaged in the searching process and use all available sources for their information (Kim et al., 2007).

As expected, the younger consumers used reviews to the greatest extent, even the highest score on importance of all analyses. It is obvious they are followed by the oldest groups who also scored above the mean. Consumers 30-45 years old used reviews in their store choice decision the least. Furthermore, the results of the underlying review constructs show that the two oldest groups, all the consumers older than 45 years, agreed the most positively on the way reviews influenced their decision. Where older people formerly used traditional word-of-mouth, it now seems that the older consumers trust reviews and frequently use reviews in their store choice. The difference with the people under 30 years is small and this confirms the acceptance of use of reviews by younger people.

5.2 Managerial implications

The answer to the question if reviews will help retailers attract people to their store, is highly dependent on the type of store and the visitors of the store. The results have shown that reviews play an important part in the store choice decision of consumers in 33% of the cases. Especially to help consumers to imaging the store. The interest however could grow in importance depending on various aspects. First of all, the type of products that the store sells

(35)

is essential in the use of reviews. Stores that sell products that are high involvement purchases for consumers could get a lot of benefit out of reviews and therewith reviews will support retailers to attract consumers to their store. On the other hand, for example for grocery stores, the effect reviews will be significantly smaller and for the most consumers not essential in their store choice. Meanwhile when a store is new in a city or attracts relatively more new (first time) consumers it could be interesting for a store to use reviews. Mainly for new customers, reviews will help them learn more about the store, because it helps them to imagine the store.

In addition a retailer must wonder which consumers are their potential customers. If females are the most important group, it is a very interesting option to consider deploying reviews. Females used reviews to a greater extent in their store choice decision than males and the way of use is that not different. On the other hand, this difference is not so obvious amongst the different age groups. For both older and younger consumers reviews make sense. The results showed that the younger consumers used reviews the most. When older

consumers used reviews, they used them in all facets more than youth. The only group that really used reviews in lesser extent are 30-45 year olds.

5.1 Limitations and future research

The most important limitation of this research is that it is a simulation of the store choice decision of consumers. The disadvantage of a simulation compared an actual situation is the lack of real conditions and thus the results are only an approximation of the real store choice decision process. Furthermore, it is a virtual setting with a forced choice for a store. Under normal conditions for example it also plays a role if you have time pressure or you are going for fun shopping. Further research needs to establish if under real conditions the results will be the same.

Although the results are tested for different industries to find out the differences between involvement, it provides only an indication of the use of reviews. First of all, each consumer has a different involvement level in the purchase of products. What may be a special purchase for one consumer will only be a routine purchase for another one, for example the purchase of fashion products. Further research should show if instead of an approximation of involvement, consumers make the same decision on their real involvement level.

(36)

The results are tested for familiarity however the degree of familiarity with the store is not tested. Indeed, there is a difference between people who ‘know’ the store, have visited a long time ago and are frequent consumers. Further research should confirm if the results of familiarity are different with the degree of familiarity with the store.

The research is carried out under Dutch consumers with Dutch stores. This means that generalization to global effect of reviews will be difficult. Therefore the research must be replicated in other countries spread over the world. The same must be concluded for

generalization through all industries. Now I selected three ‘extreme’ industries, which means the extremes are very different industries. For generalization for all industries the research should be replicated through more industries. Finally the research is carried out online, which means that consumers without internet, are excluded in the research. Therefor it is not a possible to generalize the results for the complete Dutch population. For future research should also include the group without internet in the research, by doing the research also offline.

Referenties

GERELATEERDE DOCUMENTEN

Three mayor conclusions were drawn: (1) review quantity has a positive effect on sales, (2) review variance has a negative effect on sales and (3) review valence has a positive

In this study we expected the mediators product involvement and number of connections to be mediating the effect of consumer innovativeness on the level of ingoing

Moreover, I expect that those people exposed to high economic inequality prefer to buy high effort goods, as a result of a reduced sense of control.. In total 143

Central question: How do review valence and spelling mistakes in a review affect webroomers’ choice confidence, in-store, and is this effect mediated by the

Additionally interaction variables were created between all the independent variables, with the main interaction variable called Inequality * Endorsement * SJS, this

While this study builds on previous literature on online consumer reviews by studying real name exposure, spelling errors, homophily and expert status (Schindler

This study investigates the effect of positive emotional expressions in online consumer reviews on the buying intention and product evaluation towards shampoo and a digital

[r]