The Unali platform wishlists are personalized. The purpose of this personalization is to enhance the shopping experience suggesting and giving priority to items customers might like; nudging them with limited time events they might be receptive to.
This personalization is made possible by the influencer and customer data we collect. Both feed into our Recommendation Engine.
This is the first and most crucial step for building a recommendation engine. The central data points are a combination of influencers rating and customers behavioral data.
Customers behavioral data, or implicit data, is tied to the Events we track.
Following are the successive steps taken by Unali to feed its Recommendation Engine to determine items ranking and sequencing.
The key to item ranking is the development of an implicit rating function.
The goal of this function is to use the Event tracking in combination with a relative weight to determine a score of how likely is a user to "be interested" in an individual product.
We define an implicit rating function that outputs a number that shows how much user u will be interested in buying item i. To be more precise, we’re interested in knowing how close user u is to buying item i, so that we can use this knowledge to find similar things that the user might buy instead or buy also with item i.
With this in mind, implicit rating of item i can be calculated as:
Where