Case studies/Forecasting & Optimization Case-Study
Dynamic pricing to smooth demand.
Beverages / FMCG · 2024
Challenge
Delivery demand swung sharply between weekdays, low on Monday, high on Friday, and most orders arrived only 1 to 3 days in advance. That required high inventory and expensive external vehicles and drivers on peak days.
Solution
Over 500 time-series models were trained, tested, and compared, combined with customer segmentation by location, product mix, and ordering behaviour. An optimisation model calculated delivery cost per day, plus incentive structures to shift orders to quieter days.
Impact
Lower logistics costs, better planning reliability, and less dependence on external providers.