A deep understanding of influencer marketing in the tourism industry: a structural analysis of unstructured text

Hyunsang Son and Young Eun Park

Current Issues in Tourism, 2025, vol. 28, issue 14, 2231-2241

Abstract: Using both a word frequency approach and a cutting-edge transfer learning technique for natural language processing with BERTopic, the present study analysed the entire texts from the top 40 travel influencers’ Instagram posts (n = 23,223). Among the 256 features that we initially extracted, we ranked the top 19 features using the machine learning algorithm XGBoost and estimated the effects of these features on consumer engagement using Negative Binomial regression. The results show that seasonal trips, travel destination recommendations, recommendations for fashion during the trip, and emphasising travel-related emotion generate a higher level of engagement. For message strategy, specifically focusing on linguistic features, it is recommended that influencers use analytic, authentic, want-related, and space-related words in the caption but should avoid using too many hashtags. Also, overall, influencers should avoid sending messages during the night, with messages that are too long or with too many emojis.

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/13683500.2024.2368152 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:rcitxx:v:28:y:2025:i:14:p:2231-2241

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/rcit20

DOI: 10.1080/13683500.2024.2368152

Access Statistics for this article

Current Issues in Tourism is currently edited by Jennifer Tunstall

More articles in Current Issues in Tourism from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().