Purpose – Considering the importance of the content created by the host for Airbnb consumers while making purchasing decisions, this study aims to analyze how the Airbnb hosts promote their properties by revealing the predominant attributes considered by hosts when advertising them. Design/methodology/approach – The unstructured textual content of online Airbnb accommodations advertisements (property descriptions) is analyzed through a longitudinal text mining approach. This study defines a pipeline based on a topic modeling approach that allows not only to identity the most prevalent text attributes but also its distribution through time. Findings – This research identifies and characterizes the attributes most advertised over time, on about 30,000 accommodations posted monthly over two years, between 2018 and 2020. Five main topics were identified in the data reflecting only pull motivations. Noteworthy is the slight changes in properties’ descriptions topics along the two years, suggesting that ‘‘service’’ is increasingly being perceived by hosts as an important attribute of Airbnb guest experience. Originality/value – Through a text analysis, this study provides an insight into peer-to-peer accommodation on the key attributes that hosts consider in the description of their properties to leverage the attractiveness of Airbnb. In the light of existing research, which has predominantly focused on the trustworthiness and attractiveness of the Airbnb advertisement, this research differentiates by analyzing the main attributes in text over time. Given the Airbnb’s changes since its inception, a longitudinal view is relevant to clarify how hosts advertise their properties and how it evolves in the light of these changes.

--

Topic modeling,Airbnb,Hosts advertisements,Push and pull factors,Sharing economy

• Computer and Information Sciences - Natural Sciences
• Electrical Engineering, Electronic Engineering, Information Engineering - Engineering and Technology