CALL THE DOCTORS
Over the last couple years, the client was experiencing sharply increasing inventory levels and relatively low inventory turns. Although overall inventory levels were rising, there was too little inventory available at the right place and at the right time for certain items, resulting in lost sales. The client’s demand forecasting methods were proving to be too basic to meet the needs of its growing business. The forecasts did not incorporate variables such as seasonality and promotions. The replenishment decision-making process was a time-consuming process yielding sub-optimal results because of the manual adjustments necessary for seasonality, promotions, and location-specific insights.
Additionally, the client had specific inventory management requirements given its operational realities that were not fully accounted for in its replenishment processes. For example, a hub-and-spoke system existed among the stores whereby smaller ‘spoke’ stores without backrooms would receive stock from larger ‘hub’ stores instead of the distribution centers. As a second example, certain items were received from suppliers as a given stock set. Some of the items in that standard stock set were demanded much more quickly than others but it was not possible to order those items independently from the entire stock set. The decision of when to replenish stock given the fixed stock set was not optimized.
Looking to gain control over increasing inventory levels and improve its forecasting and replenishment methods, the client approached Invent.
Using historical data Invent conducted a system assessment. A retrospective forecast was built to quantify and analyze lost sales. Sensitivities to various market and seasonality factors were also calculated and built into an in-season forecasting model. Using its unique algorithms that are able to dynamically profit-optimize the location of every unit of stock as market conditions are changing and select the corresponding service levels, Invent Replenishment provided the client with daily replenishment recommendations for every SKU-store combination via its web interface. These recommendations incorporated in-season forecasted demand, actual stock levels, and promotion information. They also optimized replenishment considering hub-and-spoke store relationships and shipped stock-set constraints for certain items.
A 2-month pilot was conducted using the solution with a group of comparable test and control stores for selected categories. The pilot sales categories accounted for 10% of the client’s sales revenues. By the end of the test period, inventory decreased by more than 10% while gross margin increased by 1 percentage point. A 20% increase in demand forecast accuracy also helped to drive more accurate replenishment decisions.
The solution has been rolled out to all categories and stores and is currently being extended to distribution centers.