We used machine learning to understand the different purchasing rhythms on a store by store basis, in order to increase the accuracy of sales predictions.
This increased accuracy, working in tandem with a machine learning product order algorithm, ensured the business orders, and only makes, the correct number of specific products each day to prevent the over-ordering of perishable ingredients.
Orders are generated for store managers, based on a host of complex data points, such as day of the week, weather forecast, external events etc, ensuring we both reduce
the manager's time pressure, as well as food wastage.
Increasingly consumers are looking to brands to step up and become more sustainable in their working business practices, and Pret identified an area for improvement within their kitchens and food stocking efforts.
Information, on a store level, was tracked daily in a live dashboard, to ensure a constant flow of data and insight back into the business to aid rapid decision making.
This work drove efficiencies in the stock ordering, and operational, processes, and resulted in significant cost savings across their stores, along with the sustainable benefits too.