AI Case Study
Danone increased product demand forecast accuracy to 92% with a 55% improvement in net uplift from promotional events using machine learning
Danone has improved the reliability of forecasting increased demand for products as a result of its promotional efforts (uplift) through implementing the ToolsGroup system which analyses demand variability. The system draws in information from point-of-sale feeds in hypermarkets
Consumer Goods And Services
Food Beverage And Drugs
The ToolsGroup system "Employs demand sensing to incorporate POS feeds from Danone’s largest customers, hypermarkets....For Sales and Marketing, the new system provides reliable detailed modeling of trade promotion uplift. It allows campaign operational planning in which the promotion and media event is allocated to the account and provides support to customer plan definition for key accounts."
"Danone improved performance across a range of business metrics from front-end forecasting to back-end production planning, including:
* A quick and dramatic 20% reduction in forecast error, increasing forecast accuracy to 92%
* A 30% reduction in lost sales, increasing service levels to 98.6%
* A 30% reduction in product obsolescence
* A 6% increase in Net ROI in 2011, improved further to 8% in 2012
* A 36% improvement in net uplift from promotions in 2011, improved further to 55% in 2012
* Demand planners workload reduced by 50% and refocused to higher value-added activities."
"The ToolsGroup system uses machine learning technology to analyze demand variability and improve demand planning."
Budgeting And Forecasting
"Danone offers a wide range of fresh healthy food products impacted by trade promotions and media events. More than 30% of their volume is sold on promotional offers (discounts, leaflets, displays, hostesses, etc.), accounting for nearly 70% of forecast inaccuracy. Danone also spends a lot on media advertising, which significantly impacts sales and forecasts. Danone’s forecasting, planning and execution for these promotional and media events was somewhat ad hoc. Four functional departments: sales, demand forecasting, account planning and finance had trouble properly coordinating and communicating their goals, plans and expected outcomes, producing uneven results and unpredictable emergencies."
POS data feeds from hypermarkets concerning Danone products