Retail supply chain teams will be able to use the Blue Yonder Replenishment Optimisation solution to overcome the challenges they face in ordering and distributing stock.
Blue Yonder’s Replenishment Optimisation for Distribution Centers uses artificial intelligence to optimise and automate these decisions to better balance oversupply in the warehouse with out-of-stock and waste situations in the store.
The new solution ensures that the distribution center has exactly the right amount of products available that the stores will need (not more, or less). Currently, most supply chain managers use a mix of manual processes and rule-based logic from their fulfillment systems to decide what to order and where to send available stock.
With this new solution from Blue Yonder, machine learning enables grocery retailers to calculate the likely demand from each store based on individual customer purchasing patterns, including the potential profitability of the decision.
Customer demand for individual products are calculated at the store level and harmonized across the demand chain to optimise order intake. This allows stock ordering to be based on the most granular level of information; knowing true customer demand ensures increased accuracy and optimal stock investment, synchronising the demand between the store and the distribution center. It can automatically include events like promotions into its calculations.
Professor Michael Feindt, Chief Data Scientist and founder of Blue Yonder says: “Our Replenishment Optimisation for Distribution Centers (DC) moves the Blue Yonder solution further along the supply chain into the warehouse and forms part of a complete end-to-end solution for demand planning and replenishment, synchronising the whole process. The solution is based on two optimisations. Our first optimisation uses actual customer demand at the store level to predict the purchase orders for the DC, removing “gut feeling” and work-around spreadsheet calculations from the process. The second optimisation applies ordering constraints such as minimum order quantities, rounding requirements and shelf life inputs.
“For too long warehouse managers have lacked an automated process for dealing with oversupply and undersupply of stock. Processes used today are either outdated, based on inflexible rules or non-existent. We can now predict which stores will better utilise the stock that is available at the DC and use this to determine how to distribute the stock. This balances oversupply in the warehouse with out-of-stock situations in stores, ultimately reducing costs and improving the customer experience.”