Customer description

Millano Group is a company with more than 40 years of experience in the production of top quality chocolate. The company's mission is to provide the market with products that give customers and consumers satisfaction and pleasure, which at the same time are safe and environmentally friendly. The Millano Group offers its products on the domestic market as well as on many foreign markets. Currently, the company cooperates with distributors from all over the world, working with partners from more than 50 countries including the United States, the United Kingdom, China, Japan, Australia, Belgium or Mexico.

Purpose of the project

The aim of the Dature implementation project is to support the Demand Planning process through the use of machine learning in relation to finished goods forecasting. The expected business benefit is to improve the accuracy of forecasts and support the S&OP process from the point of view of optimising indicators such as Service Level or average stock levels.

Challenge

Demand forecasting is carried out both for branded products such as Hibbi, Pomorzanka and Choctopus and for private labels. Product sales are seasonal as well as event-driven, varying significantly across channels. Such a wide variety of demand patterns requires the use of demand forecasting based on machine learning, but which also allows the use of expert knowledge as well as forecasts provided by customers.

Solution

As part of the Dature implementation project at the Millano Group carried out by Smartstock, machine learning (ML) functionalities were activated. Forecasting takes place both at the lowest level of the hierarchy and at defined aggregates and is used in both the S&OP and operational planning process.

Business benefits

‘The use of machine learning in the Dature system for demand forecasting significantly supports the process of continuous improvement in our planning. To date, analyses of the accuracy of forecasts using the Dature system's ML-based models confirm a significant advantage over the previously used approach. By automating the process of generating a statistical forecast, we have a faster and more complete view of market potential, which allows us to serve our customers more effectively and achieve our mission.’

Ireneusz Kozber
Director of Logistics
Millano sp. z o.o. S.K.A.