Target, one of the largest retailers in the U.S., needed to leverage multiple data inputs to build a useful product recommendation engine and ensure product listing quality for its e-commerce business. Through its partnership with super.AI, the company was able to automatically categorize product images and score product listing quality using artificial intelligence (AI). This information was then used to flag product listings that do not satisfy quality standards as well as make relevant product recommendations.
Our customer is Target, one of the largest American retailers. The company partners with hundreds of vendors that are constantly uploading new product listing content to their website. They realized that they had a lot of product content at their disposal, but no insight into the actual quality of data or how this can be used to gain insights into conversion improvement or creation of additional sources of revenue.
Target wanted to automate quality assurance inspection for the content (images, descriptions, and reviews) in their product listings and create a scorecard for vendors that rates content after it is uploaded. Vendors would then receive guidance on how to best optimize their submitted product content to maximize sales and ultimately prevent users from encountering low quality product content.
The customer reached out to us to help with improving the categorization of their vendor uploaded content and to use product metadata to build a recommendation engine. To simplify this process for the customer and eliminate the need for IT resources on their side, super.AI scraped product content from their website and then:
This information was then combined with transaction data and used to compare products to create a recommendation engine and calculate lift from content improvement.
Target is now able to automate 96% of product listing inspection with up to 99% accuracy, drastically reducing the time and resources required to maintain quality levels while also increasing the scope of products inspected. Additionally, the company had granular product content metadata, which was used to build product recommendation models. For the first time, the company was able to launch product recommendation campaigns via email and their website, improving the bottom line for themselves and their third-party vendors.
“We got insights into our product content that we’ve never been able to get before. As soon as I saw the first report during the POC I realized I would make a very compelling presentation to leadership 3 levels above me”