Using predictive analytics to make cycle counting better

While cycle counting the existing inventory in your warehouse is a necessary process, it eats up valuable employee time. And frankly, it's tedious and thankless work. 

To make the whole process more efficient and effective, third-party logistics provider GEODIS is calling on the power of predictive analytics. 

In a presentation at this week's Council of Supply Chain Management Professionals (CSCMP) EDGE Conference, company representatives outlined how they have developed a neural network model that identifies what items have the highest possibility of a mismatch between the actual inventory level and what is record in the company's system. Instead of counting all the items in the warehouse, inventory control workers focus specifically on those high-risk items.

This change has greatly reduced the amount of time needed to perform cycle counting, which frees up inventory control workers to conduct a root cause analysis of problems revealed during the count. Additionally inventory control is better able to unearth inventory discrepancies before they actually cause a problem, such as when a picker goes to pick an item and it is not available in the slotting location. As a result, picker productivity has improved as has employee retention. 

GEODIS has piloted the program for two clients, a medical supply company and a well-known toy brand. The company will be releasing a white paper with greater details at the end of this month.