Shopping basket

Case study


A top-3 UK supermarket


Grocery Retail


The client concerned has a database of c. 13m UK loyalty cardholders and around 10m customer transactions per week, with each basket containing on average around 30 items. Moreover, these transactions are split across 6 differing retail formats and, with 24 hour trading and home delivery at stores across the UK, cater for a wide range of shopping missions, including everything from weekly shops, to mid-week top-ups to lunch time snacks.

Given such a diverse range of activities the client needed a structured way to manage its direct communications with customers and allocate funds and resources between differing communication types and objectives.

The purpose of the project was to create a system for budget management that would enable the best message to be sent to the best customer at the best time, taking into account the customer profile, the business need, time of year, who was funding the communication and many other conflicting criteria.

Analyses and Answers

Historical results covering the sales uplifts from differing customer segments in response to differing types of direct marketing activities were modelled to provide forecasts for uplifts from all customers, for all types of campaigns for each week of the year.

A scenario and optimisation model, called Pyramid, was built using the SAS software system to bring together the marketing rules (such as the maximum number of communications that a customer could be sent within a specific number of weeks or the fact that a Christmas mailing should only be sent at Christmas etc) with the customer rules (such as opt-outs), and the range of available campaigns, budgets and expected profitability’s.

By allowing both manual choice of campaigns as well as the ability to see the maximum potential mathematical return (calculated using the SAS optimisation engine) the client was able to design a laydown of the marketing budget that met both their commercial objectives and allowed freedom to explore alternatives.


The Pyramid tool enabled the client to define promotional schedules in such a way that the marketing funds available (both from the retailer and its suppliers) could be used to generate the highest footfall, customer loyalty and share of wallet. It also enabled changes in budget and circumstance to be quickly re-planned. Within the first three months of operation over £150m of incremental sales and £9m of incremental profit was able to be generated as a result of using the system to optimise the allocation & timing of the customer loyalty marketing programme.