Text Mining Online Reviews: What Makes a Helpful Online Review?

Research output: Contribution to journalArticlepeer-review


One of the main strategies to facilitate the innovation diffusion of a technology product is displaying positive customer reviews. However, not all reviews matter to people. We examine more than 100 000 smartphone reviews on Amazon.com to investigate how sentiment, emotions, and length of online reviews might influence online review helpfulness. We perform a lexicon-based sentiment analysis to extract sentimental and emotional polarity. The findings of this study indicate certain patterns in online reviews that receive the most 'helpful' votes. First, negative reviews leave a bigger impact on people in terms of 'helpful' votes. Second, long reviews might leave a bigger impact on people compared to short reviews; however, the effect is moderated by both positive and negative sentiment polarity. People find long reviews helpful when they are 'unsentimental.' Emotional words, in general, are not helpful for improving the impact of online reviews; however, trust-triggering words improve online review helpfulness. To best introduce a technology, give practical information, let people be the judge, and use many trust-triggering words, such as 'assure' and 'believe.'

Original languageEnglish
Pages (from-to)145-156
Number of pages12
JournalIEEE Engineering Management Review
Issue number4
StatePublished - 2023


  • Fake reviews
  • online review helpfulness
  • online reviews
  • sentiment analysis
  • text mining


Dive into the research topics of 'Text Mining Online Reviews: What Makes a Helpful Online Review?'. Together they form a unique fingerprint.

Cite this