Blog

Case Study: Top Clothing Brand + WebPurify

June 13, 2017 | Image Moderation, Profanity Filter, Case Studies

several multi-colored shirts on hangers

Background

Our client, a top clothing brand whose name is withheld due to NDA, launched a website allowing users to customize apparel with their own personal images and text. As with many brands, said client recognized the power of customer engagement implicit in customization. They aimed to increase sales while allowing buyers to express themselves.

The Challenge

The client approached WebPurify to develop a custom moderation plan that would mitigate the significant risks associated with printing UGC on their apparel. A fast turnaround on moderation times was especially critical for this use case, since customers uploaded their images and text during the checkout UX. If any content violated the website’s terms of use, the customer needed to be notified immediately, and prompted to try another submission, else the client risked a bottleneck and would lose potential sales.

The Solution

The WebPurify team helped our client devise and implement a hybrid moderation approach, one that made use of both our profanity filter AI and dedicated live moderators. Our profanity filter screened all submissions for obscene or hateful speech in 15 languages, along with any terms or words the client added to a custom block list, including brand-bashing and competitor names. The custom live moderation team was trained on the client’s exact image submission criteria, and provided 24/7 coverage.

The Results

  • 100,000 images and 500,000 text messages processed within days of website’s launch
  • 5+ years of ongoing content moderation
  • Average review time per image is under 1.5 minutes