Personalization
Offer 1 to1 personalized experience by detecting the shoppers preferences from their recent interactions
Dynamic ranking
Our Learning-to-Rank models optimize the ranking of products in real-time using contextual signals received along with the search query
Our advanced Learning-to-Rank (LTR) models automatically optimize the ranking of products on your website to increase conversions. The LTR models have ability to optimize ranking for 90% of the search queries (in-contrast to traditional clickstream based ranking models which only cover 20/30% of the head queries) with the help of more than 200 signals.
The signals are fed to the learning-to-rank model in real-time to rank the products in the result set.The LTR model is trained to identify shifts in trend and automatically rebalance the importance of the signals to rank products which convert.
We use a combination of optimized real-time processing techniques and precomputed cached metrics to generate the signals without affecting the performance of the search query. The signals are generated for each request by Unbxd using a variety of factors associated to the query, the user, the store and the catalog:
Dynamic ranking
Offer 1 to1 personalized experience by detecting the shoppers preferences from their recent interactions