I think the author has it wrong, what the author sees as a "barrage of angry or ecstatic reviews" is actually a rich array of personal experiences waiting to be

I think the author has it wrong, what the author sees as a "barrage of angry or ecstatic reviews" is actually a rich array of personal experiences waiting to be interpreted. Traditionally, this interpretation has been done wholly by humans, and was a long and laborious task. However, we can now do much of the manual work with NLP [1] - discovering topics of concern and measuring the prevalence and sentiment. This lets us ask the dataset useful questions, like "What issue is angering customers most?", and "What features should we be talking more about?" It's a move towards a much, much wider sampling of customer experiences (compared to focus groups), and I think it should be celebrated for empowering companies to make products/services that consumers really, really like. [1] https://taggit.io https://taggit.io