CHALLENGES Initially, product attachment was a relatively manual process, in which the online merchandising team would manually set up bundles for each individual product and each individual add-on product to that. They identified the need to not only automate that process, but also add a layer of personalization. The team had created a natural attach model (similar to market basket analysis), which looked at combinations of products naturally purchased together by all customers (still, not personalized), and had been unable to put them into production on their website. Furthermore, when they looked at their Adobe Analytics data, they realized that it was very structured and not in a format that could be used right away for analysis. Without an extensive data engineering team,

Join for free to read