Product Recommendation System Based on Similar Profiles
The E-Commerce Director wants to increase the average basket value by recommending relevant items to shoppers based on the tastes of users with identical behaviour.
Where the money comes from
What to do now
- 1Calculate the contribution margin per product (price - cost - shipping - returns).
- 2Launch the recommender with an A/B test on the product page and the cart.
- 3Measure margin per session, not just average basket value.
For your committee
One page with the decision, the impact, the risks, the steps and the KPI to track.
The print dialog opens: choose "Save as PDF".
Watch out: more volume is not always more profit
Recommending the best sellers can push minimum-margin products, bulky items or items with many returns: the basket value goes up and profit goes down. Rank recommendations by contribution margin net of logistics and returns.
KPI for the committee
Contribution margin per session.
Detail for your data team
Proposed technique: K-Nearest Neighbours (KNN)
Why: KNN classifies and recommends by pure direct similarity ('tell me what your closest neighbours buy'). It is extremely intuitive for product and marketing teams.
Cost of getting it wrong: Recommending a poorly matched garment slightly reduces the click rate, with a low financial impact.
Data needed: Web Analytics: History of pages viewed, time on product and categories explored.; Transactions: Historical shopping basket, sizes, favourite brands and price range.; Demographic Data: Geographic location and purchase seasonality.