Seizing your market share: Deciphering the role of visual branding with deep residual networks

Yijing Li, Eric T.K. Lim, Hefu Liu, Yong Liu

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review


Drawing on a well-accepted classification of brand experience, this study attempts to unravel how service vendors could harness images to profile their service experience and in turn, differentiate themselves from competitors. Specifically, we advance four distinct visual cues as focal disseminators of brand experience consumers would come to expect during service consumption. We then employ deep learning techniques to extract these experience-related stimuli from the portal images of over 282,000 service offerings from a leading service e-tailing platform by training various deep residual networks on a variety of annotated image datasets. We further explore the impact of the derived image features in amassing market share through multinomial logit market-share modeling. By attesting to the power of visual cues in branding service experience, this study not only adds to contemporary knowledge on the criticality of communicating service experience, but it also yields actionable prescriptions for achieving service branding online.

Original languageEnglish
Title of host publication40th International Conference on Information Systems, ICIS 2019
PublisherAssociation for Information Systems
ISBN (Electronic)9780996683197
Publication statusPublished - Dec 2019
Externally publishedYes
Event40th International Conference on Information Systems, ICIS 2019 - Munich, Germany
Duration: 15 Dec 201918 Dec 2019

Publication series

Name40th International Conference on Information Systems, ICIS 2019


Conference40th International Conference on Information Systems, ICIS 2019


  • Branding
  • Deep learning
  • Market share
  • Service E-tailing
  • Service experience

ASJC Scopus subject areas

  • Computer Science Applications
  • Information Systems

Cite this