Lenta hypermarket chain and Korus consulting group of companies have developed an automated system to forecast demand for regular and promotional products. It is designed to improve the accuracy of inventory planning of the retail network, not to overload warehouses and reduce possible product losses. The pilot project was launched in 10 retail outlets.
"Digital technologies have become an essential part of retail. Thanks to the pilot implementation of the new forecasting system, we were able to increase the flexibility and accuracy of forecasting demand in selected retail outlets, - said Andrey Tarasov, Director of IT systems management, Lenta company. - Moving the forecasting system from the Hadoop cluster located in the Lenta data center to the Microsoft Azure cloud allows us to accelerate the calculation of demand forecasts and this is an important advantage due to the limited time available for this task."
Migrating to the cloud allows to distribute investments over time: the customer pays only for the hours used. In addition, the tools of the Microsoft Azure cloud platform, in particular Azure Databricks, made it possible to speed up the development and support of data processing procedures and reduce the calculation time required for making daily forecasts of the assortment of stores by 30%. The migration of the system to the cloud was carried out by the joint efforts of the Lenta IT team, specialists of Korus consulting group of companies and Microsoft experts.
"Lenta always chooses advanced technologies that allow the network to meet the needs of customers. Using the Microsoft Azure cloud gave us the opportunity to offer our customers a modern, easily scalable solution based on Microsoft artificial intelligence, which will increase the accuracy of inventory planning in the retail network, and thus make it more competitive," - said Alexey Sverdlov, Director of analytical solutions department, Korus consulting group of companies.
Information about receipts and store parameters enriched with external data from the own storage of Lenta is copied to Azure Data Lake Storage. Analysis and forecasting is performed using the Azure Databricks dynamic computing cluster.
"Microsoft, together with its partners, is committed to helping its customers on their way to digitalization, - said Anna Kulashova, Director of work with large organizations, Microsoft in Russia. - Close interaction of all participants as the customer, partner and vendor is extremely important in such initiatives. Lenta's readiness to experiment and ability to quickly implement the latest technologies, the expertise of specialists of our company and Korus consulting group of companies, as well as artificial intelligence tools and big data analysis technologies in the Azure cloud formed the basis for the successful implementation of this project."