Abstract: Background and Objective: A recent study of what is known about online word of mouth identified that there was a need to understand more about the way in which it influenced the consumer decision making journey. The objective of this study was to contribute to fulfilling this need by examining the influence of online word of mouth (WOM) on the perceptions of quality for a service in the presence of different price acceptability levels. The study was conducted in the context of a service, a group package tour, which embodies high levels of experience attributes where it is known that potential customers use word-of-mouth advice as well as price to gauge quality in a purchase decision. Methodology: The study was conducted through a series of 12 online experimental settings using typical consumers in realistic information settings. Three valences of online WOM were systematically examined for their effects: Positive, negative and inconsistent. The way that WOM may influence the relationship between advertised price of a service and perceptions of its quality was examined. Three hypotheses were developed and subsequently examined using 2-way ANOVA and t-tests of results from the various conditions of WOM and price in the experiments. Results: The findings are that for all of the price acceptability levels, online WOM was found to positively relate to consumer’s quality perceptions. Under all price conditions, the level of perceived quality was not found to differ significantly when online WOM was inconsistent, that is, both positive and negative online WOM were present, than when WOM was absent. Conclusion: It is concluded that online WOM as a cue to service quality, moderates the price effect on perceptions of quality for a service with high levels of experience attributes, like a Group Package Tour (GPT). An implication is that online WOM, facilitated by the growth of Web 2.0 applications, such as social networking sites, can weaken the ability of service providers to influence consumer’s purchase choices by relying on price alone to signal quality.
INTRODUCTION
In their study and synthesis of the literature on online word-of-mouth (eWOM), King et al.1 identified a series of research and managerial questions that remained to be addressed. In their study on the consequences of eWOM on the receiver of it they identify 5 research questions to be addressed in subsequent research by scholars. In this paper we seek to address one facet of one of these questions, namely ‘How does eWOM change the consumer decision journey?’
Many studies have examined and reviewed the apparent direct relationship between WOM receipt by consumers and their choice of products or services1-3. The presence of online consumer review, a form of online WOM-has been found to help consumers judge products before purchase4 and it has been further found that WOM has a greater effect on the purchase of experience goods, rather than search goods, regardless of whether it is online or offline WOM4,5. Intermediary variables could be expected in the process through which online WOM influences the choice outcome. Offline WOM is said to shape, for example, product awareness6, product evaluation7,8 and purchase intention9,10, all of which can be considered as intermediary variables and it follows that perceived quality can be considered as another intermediary factor consumers would regularly take into account in a purchase choice process. Research has shown that advertised price has a signaling effect on quality perceptions11,12. In view of the above considerations, the role of price in influencing consumers’ perceptions of quality was investigated as influenced by online WOM.
It is well established that advertised price can also be an indicator of quality11 though Rao and Monroe13 concluded that the quality indicating effect of advertised price is only supported for medium-priced, frequently purchased goods as opposed to expensive, infrequently purchased ones. Consumers who process information comprehensively consider factors other than just advertised price in assessing quality14. Alpert et al.15 find that the quality indicating effect of advertised price is more significant when consistent with other quality cues such as brand name and in-store display. Akir and Othman16 find consumers are less attentive to price in their choice decisions when: (1) Other influential attributes are present and (2) They ask for other’s comment. Therefore, it is proposed that the presence of online WOM can influence consumer’s quality assessment of a service offer, lessening the quality signaling effect of advertised price. It follows that online WOM will be more important than advertised price as the major signal of quality when both online WOM and price information are present. The perceived price of an offer is influenced by whether the advertised price is above, within or below the consumer’s acceptable range17. Therefore, the way WOM might influence perceived quality was examined taking into account the possible influence of consumers’ price acceptability levels.
In their study of the online WOM literature King et al.1 accept a definition of WOM as involving positive or negative statements, seemingly ignoring the situation where both valences can be present, that is, WOM is inconsistent. Nowlis et al.18 argue that the attitudes of people are not affected by inconsistent information because positive information evaluations are offset by negative information. In addition, individuals who are exposed to inconsistent WOM about a brand they are not familiar with remain neutral in their attitudes when compared with individuals exposed to consistent WOM19. Nguyen and Romaniuk20 reveal that when prior probabilities for choosing to see a film are equal, positive WOM has the same effect as negative WOM. In this study we examine all three possibilities of valences of WOM.
From the above, it could be expected that: Quality perceptions will be influenced positively by the valence of WOM and when both positive and negative online WOM are present simultaneously, their collective effects will be neutral.
From consideration of the discussion above, two research questions were derived:
| • | How is the effect of the acceptability level of price on quality perceptions influenced by online WOM, for an experience, service product? |
| • | How does online WOM, in the presence of different price acceptability levels, influence perceptions of quality for an experience, service product when WOM is: (i) Positive, (ii) Negative and (iii) Both positive and negative are present simultaneously? |
The following three hypotheses were framed to help answer these questions:
| • | Hypothesis 1 (H1): For all the valences of online WOM, the influence of online WOM on perceived quality will be greater than that exerted by the acceptability level of price, when both are present |
| • | Hypothesis 2 (H2): In the presence of all acceptability levels of price, the strength of the online WOM perceived by consumers about an online offer by a service brand will relate positively to the consumers’ quality perceptions of the offer |
| • | Hypothesis 3 (H3): Online WOM, when both positive and negative WOM are present, will not affect the level of perceived quality in the presence of all acceptability levels of price |
MATERIALS AND METHODS
The hypotheses were examined by collecting data from subjects via a 4 x 4 between-subjects full factorial design as in Table 1. Subjects were asked to participate in an online exercise. They were presented with a series of virtual reality scenarios21 in which, through using information acceleration, they were given information about potential package tour holidays, including the travel agent it was available through, the price and simulated online WOM messages. Letters in each cell of Table 1 represents an individual, fictitious but real-looking Travel Agent. The definitions adopted in this study-online WOM, perceived price acceptability and perceived quality are given in Appendix 1.
Price manipulations came from the acceptable price range reported by the respondents (Appendix 2). The mean level between the lower and upper price limits reported was used for "Prices within respondent’s acceptable range"; prices of 50% below the lower price limit were used for "Prices below respondent’s acceptable range" and 50% higher were used for "Prices beyond respondent’s acceptable range".
Participants were then told that they might like to read (the simulated) online consumer reviews of the GPT by visiting (a fictitious but real-looking) travel advice web site-TravelAdvisor.com, where past users of the travel agents posted their recent experience with them for the same tours. Compared with non-user’s WOM, WOM communicated by past users appear to be viewed as more legitimate and convincing22-24. These messages were based upon the most acceptable and unacceptable travel experiences as reported by Wang et al.25 and these were associated with three fictitious travel agents who were offering the holidays.
The number of WOM messages was the same for the "All positive" and the "All negative" manipulations, whereas half of the input WOM was positive and half negative for the "Both positive and negative present" manipulation. The context and focus of the messages were the same for the three categories of manipulations.
Next, respondents were asked to complete a questionnaire (Appendix 3) that asked about their perceptions of the following key constructs, using 7 point, bi-polar scales for: WOM (3 items), service quality (5 items) and acceptability of price (1 item).
The questionnaire was pre-tested on members of an online travel interest community. The amended questionnaire was then incorporated in the design of twelve interactive web sites for the twelve experimental groups. Testing of the web sites was then conducted before posting online invitation messages in twelve online travel interest communities. Invitation emails were also sent out to leisure travellers with the help of a direct marketing company. A total of 15,000 invitation emails (1,000 emails for each comparison group) were sent out and 264 complete responses were finally obtained. On average, there were 22 complete responses for each comparison group.
Response measurement: To measure the attitudes of the respondents towards the online WOM received, the scale used by Stafford26 and Day and Stafford27 were adapted. Three seven-point bipolar adjectives (good/bad, favourable/ unfavorable, positive/negative) were used.
The SERVPERF was adapted to measure respondent’s service quality perceptions of the travel agents28-30. One question was taken from each of the original five dimensions28 and perceptions of the travel agents were measured using a seven-point likert scale.
The respondents were screened so that those who participated in this study had little knowledge of prices of alternative, similar tours and such price uncertainty is said to lead to their consciously establishing an aspirational goal or target price to pay31,32 for their holiday package tour.
| Table 1: | Full factorial experimental design |
| *The letter in each cell represents an individual, fictitious but real-looking travel agent | |
| Table 2: | Construct metrics and reliabilities |
| Table 3: | Square root of AVE and cross-correlations of constructs |
| *Single-item measure, NA: Not available | |
| Table 4: | t-tests comparing test and retest data of final perceived price in comparison group 4, 7 and 10 |
| *Not significant | |
The use of a single-item measure is appropriate for concrete attributes33 and price perception can be taken as a concrete attribute34. The respondents were required to indicate whether they thought the price charged by the travel agent was acceptable, more than acceptable or less than acceptable on a seven-point likert scale.
All the measurement scales were subjected to face validity tests by a panel of three marketing academics familiar with research on this topic. The scales were then incorporated into the interactive web sites. A Chinese language version of the web sites was also made available.
Comprehensive pilot testing of the interactive websites was conducted on samples of potential respondents and improvements made where appropriate.
Measurement adequacy: The adequacy of the measurement scales can be seen from Table 2, showing the constructs of WOM and quality perceptions exhibit sufficient internal consistency and reliability.
Table 3 indicates that the items of each construct are distinct and that they exhibit sufficient discriminant validity35.
For the construct of final price perception, the scale item was retested three months after the actual data collection with an additional round of data collection for the manipulation of positive online WOM in different price acceptability levels (i.e., comparison group 4, 7 and 10). The t-tests show that the mean final perceived price in the retests for all the three Comparison Groups does not differ significantly from the test data (Table 4): The measurement scale has good test-retest reliability.
| Table 5: | Results of the two-way ANOVA for comparison group 4-12 |
| *Not significant | |
RESULTS
Influences of online WOM versus advertised price on quality perceptions: A two-way ANOVA was performed for comparison groups 4-12 to examine the effects of online WOM and acceptability level of price on perceived quality. This shows that the interaction of online WOM and price acceptability level is not significant: F (4, 193) = 0.62, p = 0.65 (>0.05) (Table 5). Most notably, there is a significant main effect of online WOM, F (2,193) = 2,416.84, p<0.001 and an insignificant main effect of acceptability level of price, F (2, 193) = 1.13, p = 0.33 (>0.05). This means that when both online WOM and price information were present, online WOM was the major source of quality information for the holiday tour, instead of the acceptability level of price.
Figure 1 plots the mean quality perception against participant’s acceptability level of price. It shows that the mean quality perception increases only slightly as the acceptability level of price increases from below the participant’s expectations to beyond their expectations; whereas it drops sharply across the three online WOM conditions (Fig. 2). This coincides with the earlier finding that the effect of the acceptability level of price on quality is insignificant. Notably, the lines are fairly parallel to each other in Fig. 2, meaning that the insignificant main effect of price acceptability level on perceived quality is fairly consistent irrespective of the valence of online WOM. From the above discussion, support is given to H1.
Further support to the above findings comes from examining the quality perception data from comparison group 4 and 11. If the effect of the acceptability level of price on quality perceptions was not moderated by online WOM, then low quality would have been perceived when online WOM was positive and the acceptability level of price was lower than expected (Group 4).
| Fig. 1: | Comparing mean quality perceptions for various price acceptability levels |
| Fig. 2: | Comparing mean quality perceptions in different WOM and price conditions |
| Fig. 3: | Comparing mean quality perceptions (positive vs. WOM absent condition) in comparison group 4 |
High quality would be perceived when online WOM was negative and the acceptability level of price was higher than expected (Group 11). A t-test performed on comparison group 4 shows that the mean quality perception was significantly higher when positive online WOM was present (M = 5.3, SD = 0.16), than when positive online WOM was absent (M = 3.1, SD = 0.22), t (42) = 39.29, p<0.001, d = 11.85, even when the acceptability level of price was lower than expected for both circumstances (Fig. 3).
| Fig. 4: | Comparing mean quality perceptions (negative vs. WOM absent condition) in comparison group 11 |
A similar result is obtained for comparison group 11. The results of the t-test show that the mean quality perceptions was significantly lower when negative online WOM was present (M = 1.8, SD = 0.38), than when negative online WOM was absent (M = 3.7, SD = 0.37), t (46) = -17.19, p<0.001, d = 4.96, even when the acceptability level of price was higher than expected in both circumstances (Fig. 4).
Influences of WOM valences on quality perceptions in different price acceptability levels: Results of t-tests performed on data obtained from comparison group 1 and 2 show that the GPT was perceived to be of higher quality in the positive online WOM condition (M = 6.9, SD = 0.17) than in the negative online WOM condition (M = 2.2, SD = 0.18), t (44) = 89.38, p<0.001 with a large effect size (d = 26.38). There is also a significant positive relationship between online WOM and consumer’s quality perception scores (r = 0.98, df = 44, p<0.001).
Results of t-tests on data from comparison groups 4-12 confirm that the service quality of the tour was perceived to be lower in the negative online WOM condition (M = 1.6, SD = 0.33) than when both positive and negative online WOM were present (M = 3.3, SD = 0.33), t (132) = -28.87, p<0.001, d = 4.99 and when positive online WOM was present (M = 5.6, SD = 0.32), t (128) = -70.01, p<0.001, d = 12.29 (Fig. 5). There is also a significant positive relationship between online WOM and consumers’ quality perception scores (r = 0.95, df = 194, p<0.001). From the above, respondents with a positive perception of online WOM perceived the tour to be of high quality under all price conditions. Support is therefore given to H2.
| Fig. 5: | Comparing mean quality perceptions for various WOM valences |
The t-test on comparison group 6, 9 and 12 showed that the mean quality perceptions did not differ significantly when both positive and negative online WOM were present (M = 3.3, SD = 0.33) compared to when online WOM was absent (M = 3.4, SD = 0.38), t (130) = -1.79, p = 0.08 (>0.05). Participants’ quality perceptions were virtually neutral when both positive and negative online WOM were present, supporting H3.
DISCUSSION
The current finding that the strength of online WOM perceived by consumers for an online service brand relates to the degree of consumer’s quality perceptions is supported by one of those of Mortimer and Pressey36 who studied the information search activities of consumers when purchasing a credence service. They found that, among other sources, the opinion or the sought WOM of friends influenced the assessment of the quality of a service. We also found that online WOM moderates the price effect on perceptions of quality for a service with high levels of experience attributes, like a Group Package Tour (GPT) and so this study largely supports the stream of behavioural pricing research15,16,37, that the quality-indicating effect of advertised price is contingent on the presence of alternative information sources or quality cues.
Most studies of online WOM only examine the effects of positive or negative valences whereas we included situations where both are present, which we deemed meant that WOM was inconsistent. We found that both positive and negative WOM being present together will not affect the level of perceived quality for all the acceptability levels of price. This result substantiates the findings of studies by Lim and Beatty19 and Nguyen and Romaniuk20 and further reveals that regardless of whether the objective price falls below, within or beyond their expectations, the quality perceptions of individuals towards an unfamiliar brand become neutral when they are exposed to inconsistent WOM.
CONCLUSION
The major conclusions of the study can be summarised as the following:
| • | As one might expect, the more consumers interpret the online WOM they might receive about a service to be positive, the more this influences their positive beliefs about the quality of the service |
| • | Online WOM moderates the effect of price on the perception of quality for a service with high levels of experience attributes, like a GPT |
| • | Inconsistent WOM about a service, where both positive and negative views may be received, has the same influence on consumers’ perceptions of the quality of the service as the absence of WOM. Hence, if a provider of a service were seeking to try to use online WOM as a way of moderating the effect of price on quality perceptions it would be very important to try to avoid there being inconsistent online WOM. Equally, when seeking an explanation as to why price perceptions were seeming to influence quality perceptions managers should ascertain whether there is inconsistent online WOM present |
| • | Increasingly, consumers are relying on online word-of-mouth to evaluate experience-focussed services. The growth of Web 2.0 applications, such as social networking sites, is thought to weaken the ability of service providers to influence consumers’ purchase choices via traditional marketing tools, such as pricing. The present findings indicate that online WOM has much greater influence on perceptions of quality than price. More effective marketing resource allocation could be achieved by use of the insights we provide on how online eWOM compares to other extrinsic cues such as advertised price on quality perception for a service with high levels of experience attributes |
LIMITATIONS AND FUTURE RECOMMENDATIONS
Since the level of the perception of the quality offered by a service provider is an important intermediary factor that consumers consider in a purchase choice process38, future research on consumer choice process for services would be advised to include considerations of the relationship between online WOM for a service and the subsequent perceived quality of it.
In this study, the IA procedures and the VR scenarios were incorporated in an experimental research design to assess the effects of online WOM and advertised price on quality perception. In this regard, consumer’s reactions to the manipulations could be captured with both internal and external validity. Nevertheless, the findings of this research are limited to services with high levels of experience attributes, like a GPT. Future research could replicate our study for a service that is more credence-based in nature as research has shown that consumers generally use a different approach to assess experience and credence services36,39. In addition, product type (i.e., service versus search goods) moderates the effect of WOM in the consumer choice process5. Thus, it would be also desirable to replicate this study in the context of other kinds of products or services to test the generalizability of the present findings.
ACKNOWLEDGMENTS
The researchers would like to acknowledge the following individuals for their assistance in this study:
| • | Participants who have contributed to the pilot tests of the interactive web site. They have provided suggestions for further improvement of the prototype site preceding the actual experiments |
| • | Respondents of the actual experiments who spent time participating in the study |
| Appendix 1 | |
| Appendix 2 | |
| Appendix 3 | |