Work

Essays on trust and reputation in the platform economy: how complementors leverage data portability and fake reviews to address the cold-start problem

Year: 2026
Type: dissertation
Abstract: Digital platforms have become a central mode of exchange for goods and services, yet their functioning hinges on trust under pronounced information asymmetries.Reputation systems mitigate these asymmetries by translating past transactions into ratings and reviews, thereby signaling complementors' trustworthiness to prospective customers.At the same time, reputation is typically platform-bound, creating a structural disadvantage for newcomers: without a track record, complementors face reduced demand and limited market access -the cold-start problem.This dissertation examines how complementors attempt to overcome the cold-start problem by leveraging (i) cross-platform reputation portability and (ii) fake reviews, and how users respond to discrepancies in online reputation in terms of trust and decision processes.It applies methodological pluralism by combining both laboratory and online experiments, as well as platform-data analytics along three research projects.The first project investigates "mixed signals" in reputation portability and analyzes how users process numerical reputation cues when platform-native and imported star ratings coexist.Using a laboratory experiment with eye-tracking, the study examines how rating discrepancy shapes visual attention, cognitive effort, and trust judgments.The results highlight that inconsistencies between signals are not merely noise: they change the way users allocate attention and evaluate trust, implying that interface design and the presentation logic of imported reputation can systematically influence trust formation.The second project extends reputation portability from numerical ratings to textual reviews by focusing on cross-platform signaling and perceived intentionality.It combines a pre-study that elicits and validates review sentiment with a scenario-based online experiment that varies the origin (native vs. imported) and sentiment configuration of text reviews.Findings show that imported text reviews can support trust when they are consistent with platform-native information, while configurations that suggest strategic selection of imported content can undermine trust -illustrating that users evaluate not only what is communicated, but also why it is communicated.The third project addresses reputation manipulation as an alternative pathway to overcoming the cold-start problem when portability is absent or constrained.Focusing on closed reputation systems in which writing a review requires a completed transaction, it analyzes Airbnb rating-score distributions and estimates the proportion of potentially fake early-stage five-star ratings using distributional divergence measures, while accounting for identification risks such as survivorship bias.The analysis suggests that even in closed systems, early reputation is vulnerable to distortion and that a non-trivial share of reviews may be fake in the studied context.Collectively, this dissertation advances research on trust, reputation systems, and platform governance by explaining how cross-platform reputation signals are integrated under inconsistency and by quantifying strategic reputation construction in closed systems.It provides actionable implications for complementors, platform operators, and policymakers: enabling reputation portability can facilitate market entry and competition, but it should be paired with interface designs and governance mechanisms that reduce ambiguity, deter manipulation, and preserve the credibility of reputation systems.
Language: English
Open Access status: green
OpenAlex ID: W7214511742
Cites:
Cited by:
Related to: 9