Business case

A Polish distributor of construction chemicals needed support in a project related to sales forecasting based on sales data with key components of stores, products, dates and transaction costs.

The TIDK team’s main challenge was integrating transaction data with external data and extracting knowledge from them. Advanced Machine Learning and Deep Learning algorithms turned out to be necessary. Thanks to these actions, it was possible to determine the optimal path of the forecast horizon, adjust sales plans and revenue expectations to make decision-making more effective and planning shortened to the necessary minimum.

Solution:

• Analysis of data, process characteristics, business needs and potential additional data sources for forecasting,
• Establishing a workflow for the sales forecasting process,
• Development of a Machine Learning algorithm and verification of the potential of the new data sources,
• Workshops and training for the client team.

Outcomes: