THE PROBLEM
Client, a leading global FMCG, sells its products to end consumers in Switzerland through a supermarket chain with over 800 stores across the country. Sales to end customers are forecasted on a monthly basis and supply to different stores is based on these forecasts.However, it was observed that the accuracy of such forecasts was low, resulting in inventory issues (excess stock or stock outs) at the supermarket stores.
INXITE OUT APPROACH
Sales Diagnostic and Causality Analysis
Statistical analysis techniques were applied to understand the sales distribution and sales patterns (trends, seasonality etc.) across products, stores, and geographies. Additionally, the dependence of sales on various parameters like stocking pattern, weather, tourist behavior, promotion etc. were analyzed and quantified to understand the causal relationships for consequent incorporation in the model.
Demand Modelling
Sophisticated statistical and machine learning algorithms incorporating the causal parameters were utilized to develop the predictive models.
RESULT
Enhanced the sales forecasting accuracy by ~10% leading to a 3-5% uptick in sales from the convenience store chain
