Three essays on information transparency strategies of online digital platforms and their implications

Abstract

In the first essay of my dissertation, information transparency is contextualized as user-generated content moderation transparency. This paper conceptualizes user-generated content (UGC) moderation transparency and investigates how it affects the nature of UGC in the context of online reviews. Synthesizing the principal-agent perspective and differential self-awareness theory, we theorize how review moderation transparency affects volume, length and negativity of reviews. A change to the Yelp platform in 2010 providing review moderation created a natural experiment. We used a panel dataset of online reviews from the same set of restaurants on both Yelp and TripAdvisor platforms in a difference-in-differences (DID) model to test our hypotheses. We find that increasing review moderation transparency positively affects review volume but negatively affects review negativity and review length. Our findings suggest that review moderation transparency has a facilitating impact on online user behavior by increasing review quantity but also has an inhibiting impact by decreasing review helpfulness. We discuss the theoretical and practical implications of these results as they relate to the design and consumer use of online review platforms.

In the second essay of my dissertation, information transparency is contextualized as providing visual cues about products in reviews. Some review platforms have recently proposed providing photos directly in-line with review text in reviews to better convey product information. Drawing from media synchronicity theory and expectation-confirmation theory, this study investigates whether and how providing visual cues about products in reviews affects review helpfulness and product ratings. The identification strategy is a natural experiment on Yelp.com in June 2013 in which the platform shifted from separate to integrated images in reviews. We examined 68,382 reviews of the same set of restaurants from Yelp and TripAdvisor and used a difference-in-differences approach to identify the causal effect. The results suggest providing visual cues in reviews leads to a reduction review helpfulness of extreme reviews. The results also suggest that providing visual cues through pictures in reviews improves consumers’ product understanding and as a result improves the rating for selected products, especially when rating variance of a product is high.

In the third essay, information transparency is contextualized as prediction process transparency. Data-driven recommendation agents (DDRA) analyze massive amounts of data and employ algorithm in order to give consumers recommendations for both products and the degree to which prices are likely to fluctuate up or down. While adoption of such agents has been extensive, it is difficult to understand the degree to which the DDRAs are acting in the interest of the consumers or the seller. One solution to this agency problem is to provide a detailed disclosure related to the prediction of the agent as a part of the interface which enables users to incorporate their own judgments. Our study draws from a classic expectancy view of trust (Gefen et al. 2003; Mayer et al. 1995) to specifically examine how the extent to which an agent transparently discloses data during the prediction-making process influence consumer outcomes. In addition, this paper investigates how a biased DDRA (reflected by low recommendation variability) and repeated usage over time relative (vs a one-time interaction) affects users’ recommendation bias perceptions and trusting beliefs. An experimental study was conducted to test the hypothesized model.

This dissertation consists of three distinct but related essays about information transparency strategies of online digital platforms and their implications. Current Internet and mobile technologies have brought consumers the benefit of reduced information asymmetry, creating users who feel entitled to be well informed. This creates pressure for upstream sellers on online platforms to disclose more information on products, even though doing so could ultimately lead to some negative consequences, such as information overload, the adjustment of users’ information sharing behaviors and so on. Given the importance of information to consumer decision making, firms’ strategies with respect of information transparency are not clear. Therefore, there is a compelling need to develop a set of theories and principles that guide the practice of information transparency strategy of digital platforms.