Developing a macro space recommendations system

Profile

One of the UK’s leading supermarkets.

Situation

The supermarket originally used a series of large Excel VBA-powered models to collect bay space and sales data for Food and Non-Food in its Supermarket and Convenience stores and to optimise space based on historic sales curves. Over a number of years the models had become slow and difficult to understand. This solution was difficult to maintain, audit, and run with analysis sometimes taking weeks to complete.

Action

Working alongside internal data science and engineering teams, QuantSpark embarked on a four-stage process to transform macro space recommendations:

1. Optimise existing Excel tools to increase running speed, improve User Interface, and improve accuracy of recommendations

2. Refine the modelling logic based on store-specific customer behaviour

3. Develop a Proof of Concept Python-based model to calculation recommendations based on location-specific Missions, Need States, and Sales Curves

4. Develop bespoke Analytics Platform to support macro space workflow and accelerate time from scenario planning to space implementation

Impact

The bespoke macro space analytics platform developed by QuantSpark is now the backbone of macro space decisions at the supermarket. The tool vastly augments the capability of the internal team to carry out strategic and operational improvements to macro space across the estate and includes Supermarkets and Convenience stores.