Challenge

The sales force of a leading manufacturer of cooking oil, needs a statistical model to forecast daily sales at retail outlets – for 200 sales personnel serving 6000 plus retail outlets. With sales targets needed to be set at weekly and monthly levels, the daily sales forecasting at retail outlets is done using simple excel extrapolation techniques and the robustness of the forecast is questionable. Forecasts have to be aggregated at various levels for target setting and any approximation has a cascading effect through the hierarchy.

Approach

Forecasting models are used on current sales data to generate accurate daily forecasts for the retail outlets. These daily level forecasts are then leveraged for better target setting at weekly and monthly levels. Other inherent patterns extracted from the sales data (such as best days of week to call, best product mix for up-sell, best discount promotion period and many more) provide key insights for the sales team to target their retailers better. Sales team is educated on the insights at the retailer level that can help improve their overall performance and sales incentives.

Outcome

Intelligent Sales Forecasting (Daily, Weekly, Monthly) at Retailer level from historical sales data – that could positively impact potential supply chain and marketing decisions. Insights for planning promotions – timing, depth of discounts, understanding what SKUs to promote. Insights for managing salespeople - determining incentives for Salesman by Route, realigning Salesman by Retail outlets for better customer relationships. Product Insights – identifying Retailer/Product combinations to target for up-sell and cross-sell, product sales decline predictions, identifying SKUs to discontinue.