Demand Forecasting For 300+
Stores Across All Product
Challenges and Industry
A national home amenities retailer that has been in business for 15 years, annually opening between 40 to 50 new stores reached out to the GFAIVE team. Having a unique business model of purchasing stock daily and renewing in store assortment frequently, the retailer faced the issue of having low accuracy of demand forecasts.
The challenge for GFAIVE was to build a demand forecasting tool that would provide highly accurate sales predictions for each existing category of products, 1 3 months in advance for every store. These predictions would then be used in the retailer’s warehouse and store stock planning as well as purchasing decision-making.
As a first step GFAIVE’s team collected internal transaction and inventory data and enhanced it with external data seasonality and economic trends, as well as accounted for any planned price changes and marketing initiatives. This was done in order to ensure that the forecasts provided by the resulting model would be of the highest accuracy.
After the machine learning algorithms identified patterns and provided predictive models, GFAIVE team was able to integrate the solution into the clients’ existing daily inventory planning tools. Before rolling out the update across all stores, the team first tested the demand forecasting tool in 3 stores in order to make sure the levels of prediction accuracy were to the satisfaction of the client.
After GFAIVE’s demand forecasting tool was integrated, the retailer observed its demand predictions reach the accuracy of up to 80%. This led to an improved stockage planning which reduced logistics costs by 11% and increased revenue by 13% as customers were able to acquire the products they needed and weren’t disappointed by out of stocks.