It is no surprise that the consumer purchase environment has changed with the rise of the internet. Next to the fact that consumers don’t need to leave their house, can choose between a much larger (international) assortment of products, they are also exposed to significantly greater levels of information; consumer reviews, in particular. Nowadays, consumers have access to a large amount of word-of-mouth (WOM) – expressed in online consumer ratings – to make better informed purchase decisions (He & Bond, 2015).
Consumer Review Characteristics
He & Bond (2015) in their paper named ‘Why is the Crowd Divided? Attribution for Dispersion in Online Word of Mouth’, identified this phenomenon and they chose to focus not only average rating score, but also rating distribution. They authors built a framework that argues that when consumers are exposed to dispersed WOM, they attribute the dispersion on either product performance or reviewer characteristics (He & Bond, 2015). To extend existing literature further, He & Bond (2015) attempt to predict the attribution by looking at taste similarity.
Figure 1: High and low distribution example (Source: He & Bond, 2015)
Attributions for Dispersion
He & Bond (2015) argue that when consumers are aware of WOM dispersion, they will attempt to find an explanation. In principle, explanations behind dispersion can vary greatly. However, overall people’s attributions of the cause of WOM can be categorized in two ways: (1) product sources, or (2) reviewer characteristics.
Consumers could one one had attribute the cause of WOM to product sources – performance, in particular (He & Bond, 2015). However, different consumers might simply value product attributes differently, resulting in different reviews for the same product (He & Bond, 2015). Folkes (1988) indeed found that negative WOM is oftentimes perceived by consumers as incorrect product usage by the reviewer, rather than the product performance.
I hear you asking: when do consumers attribute WOM to a product source and when to reviewer characteristics?
The authors raise the principle of taste similarity. They argue that when consumers read WOM, they judge how similar the reviewers are to each other (He & Bond, 2015). When there is a high consensus (low dispersion) consumers are more likely to attribute the WOM to product sources (He & Bond, 2015). Alternatively, when consensus is low, consumers might be more likely to attribute WOM to reviewers (He & Bond, 2015).
Consequences of Attribution
When consumers attribute high-dispersion WOM to product performance, this might induce performance uncertainty, possibly resulting in a negative customer response (He & Bond, 2015). This effect might be less prominent with WOM attributed to reviewer characteristics and might even be beneficial to consumers to learn about their own preferences (He & Bond, 2015).
Based on the above, the authors formulated three hypotheses:
H1: The negative influence of WOM dispersion on product evaluations is stronger for
taste-similar domains than for taste-dissimilar domains.
H2: The moderating influence of taste similarity on product evaluations is mediated by attributions for WOM dispersion.
H3: The moderating influence of taste similarity on effects of WOM dispersion will be stronger among consumers with greater openness to experience.
Figure 2: Conceptual Model (Source: He & Bond, 2015)
Method
The authors conducted four studies. Study 1 examined the effect of dispersion and product domain on consumer choice (He & Bond, 2015). Study 2 measured the effect of dispersion on attribution and purchase intention. Study 3 investigated the effect of different taste similarities within a product domain. Lastly, study 4 examined whether people that are open to new experiences react more positively in cases of high dispersion and taste dissimilarity.
Figure 3: Experimental Conditions Example (Source: He & Bond, 2015)
Findings
All hypotheses were supported. Consumers were more willing to accept dispersed WOM when the product domain was represented by dissimilar tastes (He & Bond, 2015). Also, the second experiment revealed that participants preferred low dispersion, but were much more tolerant of dispersion in product environments with dissimilar tastes. Furthermore, study 2 proved taste similarity was indeed responsible for the attribution process by which consumers attribute WOM to either product or reviewer. Study 3 revealed that when consumers know tastes are dissimilar in a product domain, negative attitude towards the product will drop, whereas negative attitude will increase when they perceive the tastes to be similar (He & Bond, 2015). Finally, study 4 revealed that participants with a low level of openness had lower product evaluations on average. Participants high in openness were more positive about WOM dispersion, given dispersion was attributed to reviewer characteristics (He & Bond, 2015).
Strengths & Weaknesses
One strength of this paper is that it takes a new approach to explaining consumer reaction on WOM dispersion. Where previous literature focused on reference dependence as explanation of negative consumer response (uncertainty increase), this research takes it back a step and provides understanding of how consumers actually perceive the WOM by explaining their attribution process. Also, taste similarity was not included in any prior research, while He & Bond (2015) proved this to be an important mediator. Furthermore, the authors highlighted that where reference dependence argues dispersed WOM is mostly negative, this is not always the case (He & Bond, 2015).
A limitation is that all experimental conditions showed consumer reviews on a 10-point scale. The authors do not test whether the found effects are still present in different length scales. For example, consumers might perceive responses on a 5-point scale as dispersed more easily, as the values are closer together. If the majority of people is very high up in the 10-point scale and then 1 individual gives a 1-star rating, this might have less effect than with 5-point scales. The authors themselves identify the possible effect of different WOM dispersion perception with different physical appearances of the scales (He & Bond, 2015; Graham, 1937). The authors could create more experimental conditions where they also alter scale length, to better the generalizability of their study, as not every company uses the same scale length.
Furthermore, the authors took into consideration similarity between reviewers, but not similarity between the consumer and the reviewers. The results might possibly differ if the consumer finds out the reviewers are drastically different from him/herself (or very similar). An improvement could be to include experimental conditions where the ‘average reviewer profile’ is described, so people could judge how similar they are to the reviewers – and of course explicitly state this to make it measurable.
So, now that you know more about review scales; on a scale from 1 to 10, how likely are you to interpret the reviews the same way as you did before reading this post?
References
Folkes, V. S. (1988). Recent Attribution Research in Consumer Behavior: A Review and New Directions. Journal of Consumer Research, 14(4), 548 – 565.
Graham, J. L. (1937). Illusory Trends in the Observations of Bar Graphs. Journal of Experimental Psychology, 20(6): 597 – 608.
He, S. X. & Bond, S. D. (2015). Why is the crowd divided? Attribution for dispersion in online word of mouth. Journal of Consumer Research, 41(6), 1509 – 1527.
Consumer Value Creation 