Predictive Segments: Discount Affinity

Offering discounts to potential customers on purchases is one of the quickest ways to draw their attention for them to complete their purchases. However, some of the customers prefer to keep their premium status. Targeting customers who have an affinity towards discounted products is precious for marketers as it enables them to target these users as discount seekers to offer cut-off prices.

Discount Affinity uses machine learning to predict the intent of a user to purchase the discounted products and segments users based on their discount preferences as High, Mid and Low. 

This algorithm examines the past user interactions consisting of the discount information in the last 6 months such as purchase, visited product, adding the product to the cart, etc. and it uses the random forest algorithm to calculate a discount affinity score for each purchase based on different features such as the number of items visited, items' prices when visited, mean of discount percentages of visited items, etc. The initial state of a user is medium. A user with high affinity, grouped as a discount seeker, score will be likely to purchase discounted products while a user with low affinity, grouped as a premium shopper, score will be likely to purchase premium products.  

Discount seekers, who have a strong affinity towards discount, are users who consistently engage, view and buy discounted products. On the other hand, premium shoppers have no affinity towards discounts and are usually interested in buying newer collection and expect premium service.

You can use this model to tailor your discounts and coupons to ensure high conversions at good margins.