Deep Learning engine for enhanced customer conversion using Propensity Models
Propensity model uses the historic transaction and activity data of the users for predicting the likelihood of a new customer buying images.
It helps to distinguish users those who are loyal customers who do not require discount from those users who have less chance of buying unless incentivised with a discount.
The deep learning predictive model relies on features like
pages visited by the users and time spent on each page
clicks on each link and number of clicks
location and other demographic data of users
successful transaction with discounts applied and discount amount provided
successful transactions without discounts, etc.
The machine learning model predict likelihood of the new user to buy an image with and without discount, thereby avoiding blanket discounts for users while persuading the undecided users to make a decision by giving them incentives.