If nearly all Airbnb reviews are positive, does that make them meaningless?

Abstract

Peer-to-peer business models rely on interpersonal communication for their success. In this article, we focus on Airbnb – an exemplar of the so-called ‘sharing economy’ – and more specifically, on Airbnb’s reciprocal reviewing system, which enables both hosts and guests to review one another. Our study takes a computer-assisted, qualitative approach to explore linguistic patterns of evaluation in Airbnb reviews. Our findings indicate that Airbnb reviews tend to comprise a very restricted set of linguistic resources, establishing the site’s norm of highly positive commentary, which in turn makes Airbnb reviews, on the surface, appear to be quite similar to one another. However, a micro-analytic comparison of positive reviews reveals that less-than-positive experiences are sometimes communicated using more nuanced, subtle cues. This study contributes to existing literature on electronic word of mouth in the tourism industry by highlighting how evaluation is communicated, while simultaneously responding to hospitality scholars’ calls for analyses which extend beyond the star ratings and also take into account consumers’ constructions of experience in the review texts themselves.

Keywords:

• Airbnb
• online consumer reviews
• eWOM
• sharing economy
• evaluation
• discourse analysis

Notes

1 A representative from Airbnb verified that they censor reviews only if they violate the guidelines, and indicated that this happens ‘very rarely.’

2 Airbnb.com’s review guidelines state:

When creating a review or review response, we ask that you stick to the facts. The best reviews provide constructive information that helps the community make better decisions and is educational for the host or guest in question. We strongly discourage personal insults, opinion that’s not backed up by examples, or generally unsociable behavior. (Airbnb, 2015)

3 Throughout this article, we have use the term ‘positivity bias’ following prior research in this area (e.g. Zervas et al., 2015). However, as one reviewer has pointed out, from a statistical viewpoint, this would suggest that there is a true mean value for guest rating – and such a value has not yet been established.