Market Basket Analysis
Our client needs to understand better customer purchase behavior. A structured need to know which products co-occur in the same purchase rises further due to increase in portfolio, customer dynamics and per-store differentiation.
Having such information will support them in: sorting their inventory, setting-up the store layout and managing promotions.
Association rule mining (a rule-based machine learning approach) was deployed to tackle the case of product co-occurence and items causality within the purchase.
The solution gives markers a set of rules in the manner “If product A is purchased” then “Product B is purchased” with the following info as a result:
- Frequency of the purchase;
- Strength of the relationship between in-basket items;
- Validation that the combined purchase is not due to chance;
- Direction of the in-basket items purchase – which product/products is the cause and which the consequence.
This information is coupled with the financial perspective – value and margin of the basket.