Online Consumer Reviews
How does the Reac-on of a Service Provider to an Online Consumer Review impact Consumer’s Willingness to Pay and Purchase Inten-on?
Master thesis
Annemieke Doornbos
Studentnumber: s1975811
Course Program: Marke<ng Management
1st Supervisor: Dr. J. van Doorn
2nd Supervisor: Dr. J.A. Voerman
Annemieke Doornbos Marwixstraat 19a 9726 CA Groningen
Email: a.doornbos.3@student.rug.nl
Preface
Wri<ng this preface makes me suddenly aware that this thesis and therefore my studies have come to an end! While love living in Groningen and being as free as a bird, I also admit that it is <me to look for new challenges and leaving my student life behind.
The topic for this paper was easily found, I love to travel myself and when I heard about the opportunity to do this project for my supervisor I was immediately excited. Besides this, you know you have chosen a good subject when it turns up in a sketch of the well-‐known English comedian Micheal McIntyre.
(...) Booking a hotel is quite difficult now with all hotels being reviewed. You know what I am talking about, if you have ever been on TripAdvisor.com. This site has reviews from every single hotel in the world! Which is a good thing, I suppose. But what is not that posi-ve is that all hotels in the world have received at least one bad terrible review and it is only those reviews that you believe! (..) And then you find one which you like. It looks like paradise, heaven, the best hotel you’ll ever stay in! ‘Oh, it was just the most miraculous two weeks of our lives. We were picked up from the airport on a unicorn, which flew us to our des-na-on, which was so wonderfully beau-ful, the beds were so comfortable, the fish would just come up and sacrifice themselves on the plate' . . . And you're siSng there at home, and you think, 'This is it, darling. This is the one we should go to – everybody loves this hotel.' But you keep searching, and there you'll find it..page 36, one star... 'The waiter slapped my wife in the face’ (...)
Michael McIntyre, Comedy Roadshow 2009 I do know that I have learned a lot during this final phase of my studies, not only academically but also about myself. I have come to realize that I am not a person who likes to sit by herself working individually on a assignment, I prefer to work in groups and brainstorm on ideas. Although I am proud to deliver this academic research paper, I am also glad that it was the last thing on the list before gradua<ng. I am excited to look for an inspiring job where I can combine the prac<cal tools learned during my Bachelor Interna<onal Business & Languages with the academic skills provided to me during this Marke<ng Management Master.
I would like to thank my friends and family who have supported and encouraged me during my studies. In par<cular, I would like to thank my supervisor Dr. Jenny van Doorn, who has been very helpful in providing construc<ve feedback and assistance during the whole process. Also, a word of thanks to Dr. Liane Voerman, for reviewing and providing helpful <ps during the last phase of this thesis.
Management summary
This research paper looks into Online Consumer Reviews (herea_er OCR’s) and in par<cular if a response of a service provider to an online review influences consumer’s willingness to pay and purchase inten<on. As OCR’s grow in importance and popularity, this special type of electronic WOM is found to have a large influen<al role in the consumer decision process and on consumer behavior.
What has not been researched yet in literature, is if service providers should respond to reviews and if so, how service providers should go about this. This raises the ques<on, can responding to OCR’s serve as a new element in the marke<ng communica<on mix?
This thesis has focused on the way in which a company can respond to an OCR; responding with a personalized and customized reac<on, simply responding with a vague, non-‐personalized (almost automa<cally generated) reply, as well as not responding to an OCR.
The above men<oned responses and their effects on purchase inten<on and willingness to pay were tested by means of an online survey. The experiment developed consisted out of a 3X2 between-‐par<cipants design. Par<cipants were shown one out of six scenario’s and were ques<oned on their willingness to pay and purchase inten<ons.
The results show that responding with a personalized response is more effec<ve in realizing a higher purchase inten<on than responding with a vague response. An interes<ng finding is, however, that a vague response results in a lower purchase inten<on than not responding at all. When looking at the nega<vely valenced reviews, one can also conclude that a personalized response results in the highest purchase inten<on. But in contrast to the posi<vely valenced reviews, a vague response to a nega<ve review results in higher purchase inten<ons than not responding.
When looking at the other dependent variable, willingness to pay, the outcomes differ per valence group. In the posi<vely valenced scenarios, willingness to pay did not differ significantly between the scenarios. The nega<vely valenced group did differ in their willingness to pay. A personalized response scores highest on WTP, herea_er a vague response, and people were willing to spend the least for just a nega<ve review with no response.
The overall findings from this thesis thus offer strong support that responding to online reviews can be a useful tool for managers to increase WTP and purchase inten<on.
Table of Content
Chapter 1: IntroducAon
1.1 Background... 6
1.2 Theore<cal and Managerial Relevance... 8
1.3 Research ques<on... 8
1.4 Structure of thesis... 8
Chapter 2: Conceptual background 2.1 Online Consumer Reviews... 9
2.2 Posi<ve Word-‐of-‐Mouth/Nega<ve Word-‐of-‐Mouth... 10
2.3 Responding to Online Consumer Reviews... 12
2.4 How to respond to Online Consumer Reviews... 16
2.4. 1 A personalized reac<on and a vague reac<on to a posi<ve review... 16
2.4. 2 A personalized reac<on and a vague reac<on to a nega<ve review... 17
2.5 Conceptual model... 19 Chapter 3: Methodology 3.1 Research design... 20 3.2 Data collec<on... 20 3.3 Descrip<ves... 20 3.4 Randomiza<on... 21 3.5 Manipula<ons... 21 3.5.1 Valence... 21
3.5.2 Type of response... 22
3.6 Dependent variables... 24
3.6.1 Willingness to pay... 24
3.6.2 Purchase inten<ons... 25
3.7 Scales... 25
3.7.1 Cronbach’s alpha... 25
3.7.2 Factor analysis... 26
3.8 Plan of analysis... 26
Chapter 4: Results 4.1 Purchase Inten<on... 27
4.2 Willingness to pay... 29
Chapter 6: Conclusion 5.1 Discussion of the results... 32
5.2 Managerial implica<ons... 33
5.3 Research limita<ons & future research direc<ons... 34
References... 36
Appendices
Chapter 1 IntroducAon
Tradi<onal Word-‐of-‐Mouth (herea_er WOM) has been shown to play a major role in the consumer decision-‐making process (Hennig-‐Thurau et al. 2004; Lee et al. 2007). In services marke<ng, for example, authors describe WOM as ‘a dominant force in the marketplace’ and the ‘ul<mate test of the customer’s
rela<onship’ (Brown et al. 2005). WOM can be described as informal advice passed on between consumers.
It is usually interac<ve, swi_, and lacking in commercial bias (East, Hammond & Lomax, 2008). WOM may be posi<vely framed (herea_er PWOM), encouraging brand choice and/or posi<ve consumer aotudes, or nega<vely framed (herea_er NWOM), discouraging brand choice and/or nega<ve consumer aotudes (East, Hammond & Lomax, 2008).
With the rapid grow of the Internet, the op<ons for consumers to engage in electronic WOM (herea_er
eWOM) is much larger than the tradi<onal scope of WOM. Whereas tradi<onal WOM is passed on by
friends and/or family in the inner circle, eWOM on the Internet can be read/posted by anyone with access to the Internet around the globe. eWOM occurs on various online channels e.g.; discussion forums, blogs, social networking sites, news groups, consumer review websites, virtual communi<es etc. (Hennig-‐Thurau
et al, 2004; Chu & Kim, 2011). Besides the large spread, another difference between tradi<onal WOM and
eWOM is that eWOM can be found on the exact same <me that a customer is searching for the informa<on. Furthermore, WOM providers on the Internet can supplement their words with pictures, scanned documents, and suppor<ng comments by other consumers. The web thus appears to be magnifying the power of WOM in the marketplace (Bickart & Schindler, 2002).
In this study, the focus is on eWOM communicated via online consumer reviews (herea_er OCR’s) (which
can be found on sites such as tripadvisor.com, dine.com, iens.nl etc). The amount of prior literature on
OCR’s and the rela<onship between OCR’s and brand evalua<ons and/or decision-‐making shows the interest of marketers and researchers on this topic, this and different aspects of OCR’s have been studied and built upon.
1.1 Background
In 2009, 90% of all Dutch households had access to the Internet; the Dutch Central Bureau of Sta<s<cs (2009) showed that 87% of this group regularly looks for (non-‐company) informa<on on products and services on the Internet (CBS, 2009). People thus find it important to read about experiences of other people with regard to products or services that they might want to buy. Customers can find this informa<on in Online Consumer Reviews (herea_er OCR’s).
Chen and Xie, 2008; Zhu and Zhang 2010). An OCR is seen as a new product informa<on channel with growing popularity and importance (Chen and Xie, 2008). It subs<tutes and complements other forms of business-‐to-‐consumer and offline WOM communica<ons about product quality (Chevalier and Mayzlin, 2006).
Different from the tradi<onal face-‐to-‐face WOM, where the influence is typically limited to a local social network, the impact of OCR’s can reach far beyond the local community, because consumers all over the world can access a review via the Internet (Chen and Xie, 2008). OCR’s are quite common for many product categories such as books, electronics, games, videos, music, beverages, and wine (Chen & Xie, 2008). Nevertheless, OCR’s are important for many services as well, such as restaurants, hotels and flights. There are a number of reasons why people publish their experiences on opinion platorms. Hennig-‐Thurau et al. (2004) find that social benefits, economic incen<ves, concern for others, and extraversion/self-‐ enhancement are the primary reasons consumers publish their experiences on opinion platorms. Addi<onally, marketers need to be aware of the fact that recommenda<ons of other consumers are more influen<al than recommenda<ons put forward by experts of the company itself (Huang & Chen, 2006). This effec<veness can be explained by the fact that personal recommenda<ons by other consumers are usually seen as more reliable and credible than commercial informa<on provided by companies (Sjödin, 2008).
Online reviews o_en occur on separate (third-‐party) Internet platorms (such as tripadvisor.com, dine.com, iens.nl etc) and are, thus, mostly beyond any control of companies and service providers. However, companies are currently given more and more the possibility to respond to reviews that are posted on these websites. While this creates new opportuni<es for firms to directly facilitate and manage consumer social interac<ons, this also imposes new challenges because separate strategic ac<ons are o_en required to manage e-‐WOM (Chen, Wang and Xie, 2011).
Should, and if so ‘how’ should companies (re)act to OCR’s? To this date, this has not been a focal point of research, but as a nega<ve OCR can be seen as a cri<cal incident, literature on service failures and cri<cal incidents will be outlined to give direc<on on how a service provider could react to OCR’s. For example, research by Bitner et al. (1990) finds that customers were likely to have posi<ve reac<ons to encounters in which ini<al service failures were followed by effec<ve recoveries. Smith et al. (1999) have added that a recovery should match with the encounter faced and that it should match with the type and magnitude of the incident. It would be interes<ng to see whether this can be applied for the hospitality industry as well. This thesis will be build around an experiment with TripAdvisor, an online travel community network with
user-‐generated content. TripAdvisor claims to have over 10 million registered members and to feature over
1.2 Theore2cal & managerial relevance
Prior studies have inves<gated the effects of OCR/e-‐WOM on purchasing decisions and/or willingness to pay and have generated mixed results. Most research on WOM discusses the consequences for the consumer aotudes and/or buying inten<ons. Also, the antecedents of WOM are studied o_en, of which commitment is researched the most (E.g.: Matos and Rossi, 2008; Hennig-‐Thurau et al, 2004) In literature, an important part is thus le_ out of the equa<on, namely the service providers and what they can do in response to OCR’s. This lack of studies that focus on the side of service providers provides an exci<ng research opportunity, as this could give some direc<on as to how firms may poten<ally use and impact online reviews.
1.3 Research ques2on
Taking this prior research into considera<on, the following interes<ng ques<ons remain: Can OCR be used as a new marke<ng tool? Should companies react to OCR’s, and if so, in what way? Because of this shi_ in balance between consumer and marketer, there is a need for some changes or ‘updates’ in the current literature on OCR’s. Are consumers influenced by the reac<on of a service provider to an online review? So the main ques<on in this thesis is:
‘How does the reac2on of a service provider to an online consumer review impacht purchase inten2on and willingness to pay?’
1.4 Structure of this thesis
Chapter 2 Conceptual background
In order to answer the research ques<ons outlined above and to set up the research frame, a look will be given at the current research in this area. A_er this, the hypotheses will be formulated, which are drawn up in order to test the predic<ons that are under study in this thesis. Herea_er, the conceptual model will be outlined.
2.1 Online Consumer Reviews
In this sec<on, the main literature on one type of eWOM will be outlined, namely OCR’s. OCR’s -‐ defined as
any posi<ve or nega<ve statement made by poten<al, actual, or former customers about their experiences, evalua<ons, and opinions on products and services (Park & Park, 2008) -‐ are considered to be a major informa<onal source for consumer judgments and a cri<cal decision variable for online merchants (Chaverjee, 2001). Not only is the amount of consumers contribu<ng their opinions online growing, but poten<al buyers are also increasingly relying on the informa<on provided by others online (Moe and Trusov, 2011). This user-‐generated content -‐ content created by consumers themselves -‐ has gained much credibility in the eyes of the consumer as an unbiased and relevant input into their decision making process
(Sjödin, 2008). OCR’s differ from marketer-‐generated content in that they are more consumer-‐oriented
with product avributes described in terms of usage situa<ons, whereas the seller is usually more focused at
product avributes, technical specifica<ons and performance results (Lee et al, 2007). In addi<on to this,
Bickart and Schindler (2002) find that eWOM is more effec<ve in genera<ng interest in a product category than marketer-‐generated informa<on.
Recent research suggests that firms should pay aven<on to OCR’s; e.g., Godes and Mayzlin (2004) find that online pos<ngs have an impact on the ra<ngs of TV shows, and Liu (2006) and Duan, Gu & Whinston (2008) find that the volume of user reviews has a posi<ve impact on future box office revenues of movies. Several
other authors agree on this and also find that online reviews have a significant impact on sales of online
As was indicated in the first sec<on of this chapter, the Internet plays a magnifying role when it comes to eWOM. Social networking sites represent an ideal tool for eWOM, as consumers freely create and disseminate brand-‐related informa<on in their established social networks composed of friends, classmates and other acquaintances (Chu et al, 2011). Through these interac<ons, consumers voluntarily display their brand preference along with their persona (e.g. name and picture), which can engender eWOM communica<on (Chu et al, 2011). A friend pos<ng a message on his or her social network profile may
engender more trust than an anonymous person might do. Because social networks are usually formed
between consumers with similar interests, such opinions are perceived to be both relevant and unbiased and thus more likely to be believed by today’s skep<cal consumer than adver<sements or content generated by professionals (O’Connor, 2010).
2.2 Posi2ve Word-‐of-‐Mouth/Nega2ve Word-‐of-‐Mouth
Different from tradi<onal offline WOM informa<on, where a single piece of informa<on is either posi<ve or nega<ve in valence, eWOM on the Internet can include several reviews from mul<ple sources at the same
<me which can be framed both posi<vely and/or nega<vely. In literature, there is no uniform answer to the
ques<on whether posi<ve eWOM (herea_er PWOM) or nega<ve eWOM (herea_er NWOM) has more impact on the receiver’s aotude or on decision-‐making. There is a large stream that finds posi<ve WOM to have a greater effect on evalua<ons and purchase inten<ons. However, there is also a group of researchers that counter argue this by explaining why NWOM has more impact and therefore is more important in
decision making and evalua<ng products/services. The sec<on below will outline both streams.
According to East, Hammond & Lomax (2008) and Liu (2006) who have analyzed the impact of PWOM and NWOM on brand evalua<ons and purchase probability, the impact of PWOM is greater than NWOM. These authors find that PWOM tends to increase purchase probability and NWOM actually reduces purchase probability. In addi<on, Liu (2006) finds that PWOM typically gives either a direct or an indirect recommenda<on for product purchase while NWOM may involve product denigra<on, rumor, and private complaining. The reason why valence mavers according to Liu (2006) is that PWOM enhances expected quality, whereas NWOM reduces it. Despite the fact that posi<ve framing has a greater posi<ve impact on evalua<on and purchase inten<on, academic research suggests that it is more important to inves<gate nega<ve customer experiences rather than posi<ve experiences.
Prior research has derived that unfavorable informa<on about products tends to carry greater weight with prospec<ve buyers than favorable informa<on. This is also described in the prospect theory, which infers that a nega<ve, dissa<sfying customer experience may maver even more than a posi<ve, sa<sfying experience because ‘losses loom larger than gains’ (Luo, 2007). There is a quite large research stream in the
area of consumer behavior that supports this. According to this stream, NWOM has more value to the
than posi<ve informa<on, in both judgment and decision-‐making tasks (Chevalier & Mayzlin, 2006; Park & Lee, 2009). Here, nega<ve informa<on is considered more diagnos<c or informa<ve than posi<ve informa<on (Maheswaran & Meyers-‐Levy, 1990). This finding is called the nega-vity effect: people place more weight on nega<ve informa<on than to posi<ve informa<on in forming overall evalua<ons of a target. This effect has been found in person percep<on as well as product evalua<on contexts (e.g., Herr, Kardes, and Kim 1991; Ahluwalia, Burnkrant & Unnava, 2000; Sen & Lerman, 2007).
Involvement
This research stream also finds that commitment/involvement of the consumer towards the brand will moderate this effect. When commitment is lower, consumers are expected to process nega<ve informa<on in a rela<vely objec<ve manner. Conversely, highly commived consumers are likely to counter argue the nega<ve informa<on more extensively than posi<ve informa<on and therefore change their aotude in response to nega<ve informa<on (Ahluwalia et al, 2002). Another different viewpoint comes from a study
performed by Zhang, Craciun & Shin (2010). The results of this research show that consumers do not give
equal weights to posi<ve and nega<ve product reviews. Rather, the consump<on goals that consumers associate with the reviewed product trigger consumers' regulatory focus, which, in turn, bias consumers' evalua<ons of posi<vely and nega<vely valenced product reviews. For products associated with promo<on consump<on goals, consumers show a posi<vity bias, whereby they rate posi<ve reviews as more persuasive than nega<ve ones. Conversely, consumers show a nega<vity bias for products associated with preven<on consump<on goals.
Lastly, a recent study by Berger, Sorensen & Rasmussen (2010) finds that posi<ve reviews have a more posi<ve impact on book sales than nega<ve reviews. However, the study also finds that nega<ve reviews have a posi<ve impact on book sales. The authors explained the laver finding by referring to reviews as to be ‘informa<ve’. This component, informing readers of a book’s existence and characteris<cs, might en<ce readers to purchase a book, even when the persuasive component of the review advises the reader not to do so. Marke<ng theorists would relate this informa<ve component of a review to consumers’ product or brand awareness. In the considera<on set model of consumer decision-‐making, ‘awareness’ is a key
variable (Vermeulen & Seegers, 2009). In the research by the aforemen<oned authors, there was a posi<ve
main effect of review exposure on hotel considera<on, which was explained by the fact that all reviews – posi<ve or nega<ve-‐ made the consumers more aware of the reviewed hotel’s existence. Even though nega<ve reviews lower consumer aotudes toward the reviewed hotels, enhanced hotel awareness might compensate for this effect, yielding a near neutral net effect on considera<on.
Consumer goods vs. experience goods
avributes where complete informa<on about the goods can be acquired prior to purchase; experience goods are characterized by avributes that cannot be known un<l the purchase and a_er use of the product for which an informa<on search is more costly and/or difficult than direct product experience. In this light, Huang, Lurie & Mitra (2009) find that the presence of product reviews from other consumers and mul<media that enable consumers to interact with products before purchase have a greater effect on consumer search and purchase behavior for experience than for search goods. Park and Lee (2009) agree on this and argue that the nega<vity effect described above appears to be more significant when the eWOM is for experience goods rather than for search goods. In a nutshell, these studies illustrates that the eWOM effect is greater for experience goods than for search goods. Therefore, Internet marketers intending to u<lize eWOM strategically should make every effort to strengthen the perceived usefulness of online reviews, especially for experience goods (Park & Lee, 2009)
Conclusion
In conclusion, the different streams in literature do not study the same variables nor the same effects. However, what can be concluded is that PWOM is more effec<ve in genera<ng a posi<ve aotude, whereas NWOM is being seen as more informa<ve and is weighted more heavily by consumers in decision-‐making and evalua<ons. Thereby, product reviews for experience goods (e.g. services) seem to have more impact on consumer purchase decisions than consumer goods.
Prior research has thus indicated that both posi<ve and nega<ve OCR’s impact customer sa<sfac<on which ul<mately affect consumers’ willingness to pay and purchase inten<ons (Park, Lee & Han, 2007; Homburg
et al, 2005). Whereas posi<ve reviews seem to have a posi<ve effect on consumer sa<sfac<on, consumer
aotudes and purchase inten<ons (Hennig-‐Thurau et al 2004; Lee et al, 2007), consumers tend to put more weight on the nega<vely framed WOM (East et al, 2008; Liu 2006; Chevalier & Mayzlin, 2006; Park & Lee, 2009; East, Hammond & Lomax 2008). In sum, one can expect that both posi<ve and nega<ve OCR’s impact consumer’s willingness to pay and purchase inten<ons.
H1a: A posi-ve review, in comparison to a nega-ve review, will result in a higher WTP.
H1b: A posi-ve review, in comparison to a nega-ve review, will result in a higher purchase inten-on.
2.3 Responding to Online Consumer Reviews
reviews online (Ye, Gu & Chen, 2010). One possible cause of disuse is that businesses are uncertain about
the benefits of managerial response (Ye, Gu & Chen, 2010). Unfortunately, managerial responses to
reviews have not been researched in academic literature yet, there is only one preliminary study that measures the impact of managerial responses to consumer reviews, therefore this part of the literature review will also look into a few areas of consumer behavior that slightly touch upon the reac<on of a service provider to an OCR; service encounters/failures, cri<cal incidents and recovery ac<ons. Firstly, the study by Ye, Gu & Chen (2010) on managerial responses to OCR’s will be outlined, therea_er the literature on service encounters/failures and recovery ac<ons will be discussed.
Literature on service failures/recovery remedies is mostly developed in offline seongs; in such seongs the objec<ve of the service recovery is to reduce the avri<on of complaining customers and the distribu<on of nega<ve eWOM. Differing from offline service recovery, the objec<ve of online managerial responses is to react to nega<ve eWOM already posted by dissa<sfied customers (Ye et al, 2010). A working paper by Ye et al (2010) does show some preliminary interes<ng results on this topic. In this exploratory study that measures the influence of managerial responses on customer reviews, the authors have build a natural experiment provided by two online travel agencies. Both agents allowed customers to post reviews on hotels, but only one of the travel agents allowed hotel management to post managerial responses. Using a difference-‐in-‐difference approach, they find that managerial responses have a significant and posi<ve impact on the valence and volume of subsequent customer reviews. The average review ra<ng increased by 15% and average review volume increased by 48% a_er the provision of managerial responses. They also found evidence that the increase in review volume can be par<ally avributed to increases in hotel sales. This study thus provides a first insight into the possible posi<ve effects of responding to OCR’s.
Related studies
Utz, Matzat & Snijders (2009) have addressed the role of short text comments given in reac<on to nega<ve feedback in online auc<ons (e.g. Ebay). The result of their study shows that it is bever to acknowledge an incident than to deny it. Moreover, a plain apology (e.g. sorry, my fault) is more successful than an apology that offers an explana<on. According to these authors, a plain apology is a clear sign of regret, whereas explana<ons are not believed by every buyer. These effects were mediated by the perceived believability of the comments. The conclusion in this study is that operators of online marketplaces should encourage text feedback comments and reac<ons.
responsibility or apologizing in the form of a narra<ve that triggers consumers’ affec<ve reac<ons.
Only when consumers can empathize sufficiently with the company, does a narra<ve apology gain a clear advantage over a denial.
Related theories on service failures and service recoveries
As was indicated at the beginning of this sub-‐sec<on, there has not been a lot of academic research on this specific topic; therefore this lack of empirical results will be supplemented with theory from the field of service failure and service recovery. Grönroos (1988) defined service recovery as the ac<ons taken by an organiza<on in response to a service failure. It includes all the ac<vi<es and efforts employed to rec<fy, amend, and restore the loss(es) incurred a_er the failure (Dong, Evans & Zou, 2008).
Theory on service recovery indicates that consumer feedback can be used as a base for developing customer sa<sfac<on monitoring programs (Bitner, Booms & Tetreault, 1990). In this study by Bitner et al (1990), the authors show that providing customers with logical explana<ons for service failures and compensa<ng them in some way can mi<gate dissa<sfac<on and might even lead to a memorable, sa<sfying encounter. In addi<on, the authors find that customers are likely to have posi<ve reac<ons to encounters in which ini<al service failures, when the encounter is followed by effec<ve recoveries, e.g. provided with an explana<on as to why the service was unavailable, or in any way assisted in solving the problem. In contrast, failures to apologize, compensate, or explain the problem led to an unfavorable recollec<on of the encounter.
Smith, Bolton & Wagner (1999) have conducted an experiment in two different service seong (restaurants and hotels) in which customers evaluated various failure/recovery scenarios with respect to an organiza<on they recently had patronized. The results show that customers prefer to receive a recovery that ‘matches’ the type of failure they have experienced in ‘amounts’ that are commensurate with the magnitude of the failure that occured. Complainants who feel that jus<ce has been served are likely to patronize the retailer, whereas complainants who perceive a lack of jus<ce are likely to engage in NWOM (i.e. complain about the retailer to friends and family) (Blodgev, Granbois & Walters, 1993). Thus, this research indicates that service failures can lead sa<sfactory encounters if they are handled properly. Well-‐executed service recoveries are important for enhancing customer sa<sfac<on, deflec<ng the spread of damaging WOM, building customer rela<onships, and preven<ng customer defec<ons (Smith et al, 1999; Tax, Brown & Chandrashekaran, 1998).
Smith et al (1999) have determined the effects of various types of recovery efforts on customer evalua<ons
in a variety of service failure contexts. The authors treat service recovery as a ‘bundle of resources’ that an
speed, apology, recovery ini<a<on) influence customer evalua<ons through disconfirma<on and perceived jus<ce, thereby influencing sa<sfac<on with the service failure/recovery encounter.
Basically, there are two types of losses; economic losses and social losses. When outcome failures occur (e.g. a reserved hotel room is unavailable because of overbooking), customers experience an economic loss. Therefore, customers' percep<ons of distribu<ve jus<ce will be restored by recovery avributes that are economic resources, such as compensa<on (money). The impact of an apology (a social resource) on customers' percep<ons of distribu<ve jus<ce will be lower (i.e. have less u<lity) for outcome failures (an economic loss), because the resources are stored in separate mental accounts. In a similar fashion, customers' percep<ons of procedural jus<ce will be restored by recovery avributes, such as response speed (<me). When process failures occur (e.g. a front-‐desk clerk is rude), customers experience a social loss. In a nutshell, compensa<on has the greatest effect on percep<ons of distribu<ve jus<ce, whereas an apology has the greatest effect on percep<ons of interac<onal jus<ce. In the hotel context, the results show that both compensa<on and a speedy response have a greater incremental impact on customers' jus<ce evalua<ons when the failure is less severe.
Swanson & Kelley (2001) have inves<gated whether consumers’ verbal behavior related to service recoveries is impacted by their perceived avribu<ons for that recovery. A sa<sfactory recovery outcome provided through an acceptable process may result in favorable customer behaviors. One such favorable consumer behavior is posi<ve word-‐of-‐mouth. Although not sta<s<cally significant, employee-‐based recoveries had the highest mean level of posi<ve valence inten<ons in this research. An unsa<sfied recovery resulted in nega<vely valenced inten<ons.
Summary
Recent literature appears to advise to respond to online consumer reviews. The ques<on that arises is, what kind of response is appropriate? Past research has primarily focused on responding to nega<ve (cri<cal) feedback. To this author’s best knowledge, there has not been much wriven on how to respond to posi<ve comments or reviews for that maver. Up to now, academic literature has le_ responding to OCR’s out of the equa<on. It is interes<ng for both online review sites as well as companies who face nega<ve/ posi<ve OCR’s on these sites, to find out if responding has an effect on consumer decision-‐making. How to respond and if responding to posi<ve and nega<ve reviews changes purchase inten<ons and willingness to pay will be researched in the remainder of this thesis.
2.4 How to Respond to OCR’s.
poor complaint handling will damage the rela<onship most when prior experience is posi<ve, and damage it least when prior experiences are poor and expecta<ons are low (Tax et al, 1998) it is believed that responding to posi<ve and nega<ve reviews have different outcomes on purchase inten<ons and willingness to pay.
2.4.1 A personalized reac-on and a vague reac-on to a posi-ve review
A dis<nc<on is made between a vague reac<on and a personalized reac<on. A personalized response is defined as using a customer’s informa<on to deliver a targeted solu<on to the customer (Murthi & Sarkar, 2003). This is translated to a response which is build to respond directly to the comments pointed out in the review (e.g calling the guest by name, responding to specific elements in the review). In contrast, a vague response will not respond directly to any of the comments men<oned in the review (as if the review was an automated message that could have been posted in response to any review, posi<ve or nega<ve). A personalized reac<on to a posi<ve review as well as to a nega<ve review is expected to strengthen the posi<ve effect of a posi<ve review as well as to lessen the nega<ve effect on a nega<ve review. The reasoning behind this is that a personalized reac<on is part of providing service quality to customers, which is found in literature to have a posi<ve effect on purchase inten<on and WTP (Homburg, Koschate & Hoyer 2005, Zeithaml, 2000).
On the contrary, a vague reac<on towards an online review is expected to weaken the posi<ve effect of posi<ve reviews, as well as to strengthen the nega<ve effect of nega<ve reviews.
It is believed that a personalized reac<on of a service provider to an OCR will – in case of a posi<ve original review – strengthen the posi<ve impact of a posi<ve review on customer sa<sfac<on and ul<mately on purchase inten<on and WTP. The ra<onale behind this expecta<on is that a personalized response signals that a company is customer-‐oriented and focuses on a high service quality. This line of reasoning will be further outlined below.
In environments with livle informa<on, consumers tend to base their decisions on signals provided by businesses to evaluate their ability to produce (and deliver) quality products (signaling theory) (Gregg & Walczak, 2008). The less easy it is for customers to assess quality prior to purchase, the more likely they are to rely on signals to form expecta<ons about quality (Gregg & Walczak, 2008). Consumers thus seek informa<on that can ensure them that products or services are of a good quality. This is also one of the main reasons (as indicated in the literature review) why people read online reviews.
and WTP. The importance of focusing on service quality is underlined by Rust and Miu (2006) who indicate that customer sa<sfac<on can be achieved with customiza-on (which in turn increases greater WTP, frequency of purchase and probability of repurchase). In this line of reasoning, a personalized response (customiza<on) to a posi<ve OCR will be appreciated by customers, which is believed to strengthen the posi<ve rela<onship between a posi<ve review and purchase inten<on and WTP.
H2a: A personalized response will strengthen the posi-ve effect of a posi-ve online review on customer purchase inten-on
H2b: A personalized response will strengthen the posi-ve effect of a posi-ve online review on willingness to pay
It is expected that a vague response to a posi<ve review will weaken the posi<ve main effect of a posi<ve review on purchase decisions and WTP. As a posi<ve review intui<vely indicates that consumes are quite (up to highly) sa<sfied with the service provided, a vague reac<on to a posi<ve review is not likely to extend this posi<ve effect. In the sec<on below, reasons for this ra<onale will be outlined.
A theory that can help to further explain why it is expected that a vague response to a posi<ve review will tamper the posi<ve rela<onship between the posi<ve review on purchase inten<on and WTP is the disappointment theory, which is derived from the disconfirma<on paradigm used throughout literature to confirm that quality results from a comparison of perceived with expected performance (Homburg et al, 2005). This theory basically suggests that the greater the disparity between outcome and expecta<ons, the greater is a person’s disappointment or ela<on. Both these emo<ons generate addi<onal value (nega<ve or posi<ve) (Homburg et al, 2005). Thus, with respect to OCR’s, a vague response toward a posi<ve review will fall short on the expected service level that is created by the posi<ve reviews. This disconfirma<on will then result in less sa<sfied customers and purchase inten<on as well as a decreased WTP compared to the baseline.
H3a: A vague response will decrease the posi-ve effect of a posi-ve online review on purchase inten-on H3b: A vague response will decrease the posi-ve effect of a posi-ve online review on willingness to pay 2.4.2 A personalized reac-on and a vague reac-on to a nega-ve review
In contrast to posi<ve online reviews, it is expected that responding to nega<ve online reviews will yield different outcomes on purchase inten<on and WTP. A personalized reac<on to a nega<ve review is believed to weaken the nega<ve effect of nega<ve reviews on the dependent variables. The reasons for this expecta<on will be discussed in this following sec<on.
publicity has mainly focused on developing service recovery ac<ons to lessen the nega<ve impact on purchase inten<ons and consumer decision-‐making. The above-‐described equity theory and research on service recovery indicate that a firm’s service recovery efforts have important implica<ons for levels of sa<sfac<on, purchase inten<ons, and WOM: if a firm handles complaints effec<vely, this will reduce the perceived risk of occurrence (which improves trust) (Kim, Cooper, Ferrin & Dirks, 2004) In addi<on, acknowledging a failure and responding shows that a company is willing to take on responsibility and illustrates to consumers that it has a customer-‐oriented focus (Xie & Peng, 2009). Jus-ce theory sheds light on why making a statement is important in service recovery; an reac<on is viewed as a valuable reward that redistributes esteem in an exchange rela<onship, conveys empathy and concern to customers who have experienced the inconvenience (Liao, 2007).
Going a step further, there are researchers who claim that highly effec<ve recovery efforts can produce a
service recovery paradox in which secondary sa<sfac<on (i.e., sa<sfac<on a_er a failure and recovery effort) is higher than pre-‐failure levels (Dong, Evans & Zou, 2008). While other studies have reported mixed results with respect to the recovery paradox, a meta-‐analysis by de Matos & Rossi (2008) indicates that the effect was significant on sa<sfac<on (but not significant on repurchase inten<ons and WOM). In online seongs however, the recovery paradox manifests itself only for outstanding recovery efforts (Sousa & Voss, 2009). Despite the small chance of an outstanding recovery, it does provide evidence that a personalized response is very likely to achieve a higher customer sa<sfac<on, purchase inten<on and WTP than the baseline (which is nega<ve).
When responding personally, a service company can create a dialogue with its customers, which is an opportunity to listen, ask ques<ons, explain, apologize and provide an appropriate remedy. This proac<ve method can restore trust with current customers and prospec<ve customers, as well communicate the values that the company stands for (Berry, Parasuraman & Zeithaml, 1994). It is believed that in accordance with the signaling theory outlined above, a proac<ve and personalized response will lessen the nega<ve effects of nega<ve online reviews on purchase inten<on and WTP.
H4a: A personalized response will weaken the nega-ve effect of a nega-ve online review on purchase inten-on
H4b: A personalized response will weaken the nega-ve effect of a nega-ve online review on willingness to pay
In this light, a vague reac<on will further increase the nega<ve effect that the ini<al nega<ve review will have caused.
H5a: A vague response will enlarge the nega-ve effect of a nega-ve online review on purchase inten-on H5b: A vague response will enlarge the nega-ve effect of a nega-ve online review on willingness to pay
2.5 Conceptual model
Hereunder, one can find the conceptual model which follows from the hypotheses outlined in the above sec<on. As can be seen from the model, three types of responses (not responding, personalized response and a vague response) and their impact on willingness to pay and purchase inten<on will be measured for both valence groups (nega<ve review vs posi<ve review).
