Demand Forecasting For Food Industry

Challenge

One of the key challenges faced by a company in the fresh food industry, which supplies freshly prepared products to its customers, is accurately predicting food demand in its network of stores and restaurants.

Restaurant managers are tasked with placing product orders twice a week for next-day delivery, which can be a complex and uncertain process. Anticipating customer needs is difficult, often leading to issues such as food waste or shortages.

The company’s goal is to assist managers with demand forecasting, ensuring a balance between minimizing waste and avoiding stockouts.

Solution

To better predict required stocks in stores and restaurants, we’re using Machine Learning (ML) to learn from sales tickets data.

In collaboration with IDLabs, a deep learning neural network was developed. This was used to predict the quantity per product, that individual restaurant should purchase for the coming days.

The overall solution was fully integrated and automated within a cloud environment (Microsoft Azure).

Impact

Thanks to the fully automated solutions, more accurate forecasting on food demand is possible.

The result is more than just a tool to assist restaurant managers, as our solution allows for food waste reduction and increase in revenues.