Leroy Merlin wanted to deliver a tailored experience to users who were entering their site from email, or their account profile. This was a good starting point for Leroy Merlin South Africa’s personalization program because it utilized context from known users’ browsing history. By focusing on users who are already engaged and have signed up for an account, they could see early success before turning their attention to low intent and new users.
Using stacks of recommendation blocks and dynamic content, they were able to compile personalized pages that show users affinity-based recommendations. These recommendations are served based on customer data for past preferences, offline purchase history, and real-time interactions with the brand. Affinity scores are calculated using select attributes such as category, brand affinity, price range, topic, color, and more.
New users can also visit the “Just for You” page, including by navigating from the lower page welcome banner mentioned previously. These users are shown current trending products as well as deep learning recommendations based on the NextML algorithm.
NextML is an algorithm within AdaptML, a centralized AI system that automates decision making and adapts experiences according to real-time user behavior across channels, predicting their next move with unparalleled accuracy.
NextML is a self-training deep learning AI system that adapts the digital experience to each user individually, by extrapolating buying intent from user data and predicting which products they might be interested in. NextML recommends the next best series of products based, adapted to user behavior as they navigate the site. As a self-learning model, NextML continuously optimizes recommendations during the session based on buying patterns, visitor behavior, location, and more.
The team used NextML to specifically promote first-party products where possible, driving their brand reputation for top quality products in the region.