The project was ordered by one of Russia's major retail chains selling electronics. The goal was to accurately assess consumer demand, determine factors affecting sales and optimize the product range in accordance with the chosen strategy.
In order to achieve its revenues, turnover and margin targets and strengthen its position in the household appliances and electronics market, the company decided to implement a set of measures, including several technological projects.
GoodsForecast was invited to develop mathematical models to forecast the home appliances market, build sales forecasts in the chain and balance the product mix, i.e., determine the optimal set of goods to be sold and the conditions for selling them.
The selection of GoodsForecast as a vendor was preceded by a pilot project that built monthly revenue forecast models for 15 large stores for 2021. They were 3% more accurate than the models used by the retail chain at that time. At the same time, the forecast error was reduced by 20–25%.
Next, a life R&D project was launched. The models built during the project allow forecasting demand both in the short term, up to a few months, and on a strategic horizon of several years. Also, to assess the situation in the market, GoodsForecast specialists have built a sale forecast for household appliances and electronics in Russia in units and money. The forecasting horizon was 24 months.
At the proof-of-concept stage, several dozens of models were tested using retrospective data, from classical time series forecasting models to neural networks. The models showed an average of 5% improvement in accuracy. At the same time, the baseline accuracy was above 80%.
Building demand forecasts is the starting point for solving a wide variety of business problems in retail, including complex ones such as inventory rationing, product mix management and pricing. The reliability and efficiency of business decisions depend on the accuracy of the forecasts made. The predictive models built by GoodsForecast have been successfully deployed in the customer's IT infrastructure and put into operation.