Conference Agenda

11-06: Thijs Johannes Zwienenberg
Saturday, 20/Jul/2019:
3:15pm - 3:40pm

Seminar Room 2-6

Chair: Velitchka Kaltcheva


To Solicit or Not? Exploring the Effects of Soliciting Reviews in the Collaborative Economy on Review Content and Style

Authors: Thijs Johannes Zwienenberg (KU Leuven, Belgium), Tine Faseur (KU Leuven, Belgium), Yves van Vaerenbergh (KU Leuven, Belgium)

Organizations and consumers strongly rely on customer reviews and ratings. Consumers use reviews in their decision making processes, while organizations use the input of customer reviews to gather relevant insights and develop new or improve current products and services (Chevalier & Mayzlin, 2006; Ludwig et al., 2016). As organizations benefit from collecting as many reviews as possible, many organizations explicitly solicit reviews from customers. This practice is particularly prevalent in the collaborative economy, where consumers acquire temporary access to personal goods or services from peer-consumers via dedicated platforms. Apart from process and product improvement purposes, platform providers rely extensively on reviews to instill trust among the users. Platform providers explicitly ask customers to review the peer service provider. Any peer service provider who fails to reach a certain rating gets removed from the platform. This policy creates an abundance of reviews, yet researchers suggest that these reviews tend to lack informative value, underreport negative experiences and report inflated ratings.

To date, research on the consequences of explicitly soliciting reviews from customers is surprisingly scarce. Research shows a positive effect of explicitly asking consumers to spread traditional word-of-mouth, yet it remains unclear whether these results are also applicable in today’s digital economy (Wirtz and Chew, 2002; Söderlund & Mattsson, 2015). The purpose of this research is to expand our knowledge on the consequences of explicitly asking customers to write a review by testing whether this practice influences the review content (i.e. what and how much is being said) and the review style (i.e. the way the review is written).

We gained access to about 7,500 online reviews from two car- and ridesharing organizations, which were either solicited (N=6,524) or unsolicited (N=1,045). The initial results of an automated text analysis reveal that the content of solicited (unsolicited) reviews are highly positive (negative), with the majority of customers providing a 5(1) on a 5-point rating scale, that solicited reviews were much shorter then unsolicited reviews, and contained less disclosures than unsolicited reviews. In terms of review style, solicited reviews were much more impersonal and contained more signs of dishonesty than unsolicited reviews. Additional analysis will be carried out and include more in-depth analyses of both review content (i.e. which aspects of the service are customers writing about) and review style (is there a difference in linguistic style features). In a second step, we will use this input to build a predictive model that can estimate the probability that a review was explicitly solicited or not. This model may assist customers in deciding the weight they attach to a particular review they have read. This research contributes to literature as it fills an underexplored gap in current literature regarding online reviews and collaborative consumption.