The Impact of Discrepancies between Offerors’ Self-Disclosure and Customers’ Reviews on Online Sales of Experiences in Sharing Economy

Abstract

Experience sharing is becoming popular as customers increasingly respond to the rapid platform technology development. However, because of format diversity and quality variation, customers refer to multiple information sources before booking. Two of the most important information sources are the hosts’ self-disclosure texts about the experience project and customers’ online reviews. Using natural language processing (NLP) techniques, we analyze the data from Airbnb experience projects. We find that the information discrepancies in hosts’ self-disclosure texts of the experience project and customers’ online reviews, in terms of their focus on the attributes of products and services and the linguistic styles, exist, and these discrepancies affect sales. Hosts elaborate more on the descriptive attributes, whereas customers focus mainly on individual perceptions in their reviews. Customers also write online reviews in a more concise, diverse, and relaxed fashion, conveying positive emotion and a more subjective tone than expressed by hosts’ project descriptions. Additionally, a large topic difference, reflected by customers’ more details about various attributes elaborated in their online reviews compared with the attributes described by the hosts in the project description, increases sales. Further, a larger discrepancy in length and diversity increases sales, whereas a larger discrepancy in subjectivity reduces sales. Compared with the online mode, the in-person mode strengthens the impact of content and linguistic discrepancies on sales. This study’s findings will help hosts and sharing economy platforms use a relative approach to optimize their information provision and exploit the electronic word-of-mouth effect to improve customers’ online purchase intention and behavior.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.