Consumer Loop is an in-house built tool that aggregates ratings and reviews of beauty products written by consumers on e-retailer websites (Amazon, Sephora, Nocibé, Target…).
Based on what consumers like or dislike, people in marketing and research & development improve or create new beauty products thanks to consumer feedback and inspiration from competitors.
Algorithmic data collection and categorization resulted in output errors. Incorrect product information causing user confusion and distrust. For example, users saw duplicate perfumes with different ratings and reviews when in reality it should have been a single product page.
Power users alerted these issues, overwhelming the product team with messages. They spent their time manually correcting data instead of enhancing the app's value.
This was not a sustainable situation, especially since the number of products on the app was growing. At the time, we had 30,000 beauty products; today, we have over 300,000 references.
They found unreliable information on data based tool. Failed first trials led to distrust.
They had no feedback and immediate correction of data errors from the product team.
They spent of lot of time on correcting manually data errors.
They had no tracking system to monitor the corrections.