business strategy, data science

How Sales Receipt Data Contributed to Review Customer Classification and Marketing Strategy


Our client, a retail group of hypermarkets, was faced with a challenge of increasing margins in a competitive market. Ironically, when they got in touch with us, they were losing money even when sales were increasing. Their business model was based on providing premium items at the most competitive price.

Making sales receipt data valuable was the mission we accepted.


When our client presented their challenges, we had the intuition that one of the opportunities lied in their marketing approach. In particular, they were segmenting their customers based on stereotypes rather than using data to analyse their shopping habits.

We used sales receipt data to understand customers’ habits.


To make predictions about customers’ shopping behaviour, we started by analysing a three-year record of sales receipts. After cleaning, sorting and categorising data, we, then, matched receipt data with customer classes.

Then, we used this sales receipt data and applied it to create simulation of customers' typical shopping baskets. Doing this, we were able to redefine customers’ profile and segmentation.


For now, our model was presented to the board of directors. What might be coming next could be automated data-based marketing, including modules of recommendations and discounts based on each customer’s buying behaviour.