Mapping headphone brand positioning from e-commerce reviews with large language models.

Authors

Jian Li, Hengshan Cao

Published in

PloS one. Volume 21. Issue 11. Pages e0361055. Epub Nov 11, 2026.

Abstract

Online consumer reviews encode rich signals about brands, yet converting this unstructured text into actionable positioning insight remains challenging. This study proposes an LLM-driven framework that extracts structured brand perception variables from e-commerce reviews and demonstrates it in the headphone category. Using an analytic file of 5,895 review-level observations covering Sony, Bose, JBL, and Beats, we employ a structured GPT-5.4 prompt to extract multidimensional brand perception indicators from each review, capturing associations, judgments, emotional attachment, advocacy, and attribute focus. A manual validation exercise based on 400 stratified reviews showed meaningful agreement between the LLM-coded outputs and an adjudicated human benchmark. The results reveal clear brand separation. Bose leads on perceived quality (4.20/5), recommendation rate (78.3%), and strong emotional connection (39.9%), whereas Sony achieves the highest value perception (3.97/5). Beats and JBL occupy weaker quality positions, though Beats retains a comparatively stronger emotional profile. Patterns in attribute emphasis also help distinguish the brands: Bose reviews are distributed across sound quality, active noise cancellation, and comfort; Sony reviews emphasize sound quality, comfort, and value; JBL reviews are heavily concentrated around sound quality; and Beats reviews emphasize sound quality, value, and design. Negative reviews provide additional diagnostic insight, revealing durability and fit issues for JBL and perceived value concerns for Beats. The study contributes primarily to LLM-assisted marketing analytics and applied brand diagnosis by showing how constrained and human-validated LLM coding can transform review text into a scalable, multidimensional brand monitoring instrument. Brand equity and positioning theory are used as conceptual foundations for organizing the extracted perception variables rather than as targets of theoretical extension.

PMID:
43013602
Bibliographic data and abstract were imported from PubMed on 12 Nov 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.