Let’s look at some of the traits of an optimised distribution plan. Below are some of the key ingredients:

• Service Level: The primary objective of optimising the distribution is to always meet the customer service level, in other words, avoid stock outs.
• Safety Stock Targets: An optimised distribution plan should meet the safety stock targets across all distribution centres in the network so that target service levels can be achieved. Having the safety stock targets calculated scientifically using Inventory Optimisation techniques is an added advantage.
• Fair-Share of Stock: If there is excess stock, it should be distributed to all Distribution Centres (DCs) equally (in a fair-share manner). Similarly, if stock is not enough to meet safety stock targets, all DCs must take an equal hit (must go equally below the safety stock targets). This is explained in more detail in the next section.
• Maximum Stock: Stock should be within the upper boundary or maximum stock level if maximum stock level is a constraint. This helps in reducing stock ageing.
• Transportation Costs: Transportation costs must be kept to a minimum by choosing preferred routes wherever possible.
• Push/Pull Rules: Plan must meet the push/pull rules to ensure that customer facing DCs are served according to the demand signal and the excess due to cycle stock is stored at the Primary Warehouse.

There could be many more rules or scenarios specific to a business.

An optimised distribution plan in the short term helps in ensuring that the input to Stock Transport Orders generation process is accurate and generated truck loads are optimal as well. In the mid to long term, it shows a realistic stock projection across the supply chain network.

Fair Share 

An optimiser should ensure that in the case of oversupply (excess stock) or undersupply (insufficient stock) in the network, the stock should be distributed in a fair-share manner between all the DCs such that the stock levels are equally above or below the safety stock level. Fair-share function in the optimiser helps in striking the right balance of stock within the network.

Either because the items in our inventory are costly, perishable, central to our business or for any other reason, a proper stock planning always pays-off.

The fair-share function in the Optimiser is an increasing cost function of storage (inventory holding) cost in case of oversupply. In other words, as the stock level goes higher, the resultant storage cost increases. Similarly, it becomes an increasing cost function of safety stock penalty in case of undersupply.

The stock levels are divided into different segments above and below the safety stock target. Resulting cost jumps to a higher level as the stock moves to a higher stock level segment (oversupply) or lower stock level segment(undersupply). As a result, the stock is evenly distributed to achieve the lowest total cost.

The Fair-share function does not come as standard in the APO SNP Optimiser and is applied as a parameter which enables the fair-share functionality in the optimiser engine. However, in SAP's Integrated Business Planning (IBP) optimiser, fair-share has been added as part of the standard functionality.

Case Study 

From a recent implementation at one of the world’s biggest consumer goods companies in its North American territory, the SNP Optimiser was implemented to plan distribution across its supply chain which was extensive and extremely complex in terms of number of sources and possible distribution routes.

Network: The network consists of several factories with a dedicated Forwarding Warehouses (FW) which could be stocked or non-stocked. The FWs supply to Distribution Centres (DC) which in turn supply to Secondary DCs (SDC). DCs as well as SDCs serve the end customers. Overall, there are about 50 locations in the network. The FW can also supply Direct to SDCs (called Direct Shipments).  SDCs have limited storage capacity so there is a limit to which the products could be sent directly to SDCs. The network schematic diagram shows the supply chain network.

Problem: Being a large and complex supply chain with 1000's of SKUs, the company faced a lot of challenges when trying to distribute the goods through its network and suffered frequent out of stocks, imbalanced inventory - stock in the wrong place leading to wastage while some DCs being understocked and incurring huge transportation costs. There was a complete lack of predictable outcome when using their originally implemented solution and so the optimiser results were not trustworthy.

Project: The project involved designing their complete optimiser solution with the aim of achieving a predictable outcome according to the business rules. Key targets were as follows:

• Higher Service Levels
• Higher Direct shipments (Factory to SDC) to reduce transportation costs
• Stock Fair-share
• Adhere to Maximum inventory levels at SDCs to handle storage limitation

As a precursor to this project, the inventory targets were recalculated using Multi-Echelon Inventory Optimisation which provided a solid base to the distribution optimisation and complemented the whole process. This involved calculation safety stock levels at DCs and SDCs as well as maximum stock boundaries at the SDCs. The maximum stock boundaries were essential to the distribution strategy as SDCs have a limited storage capacity and the stock must always float between min and max boundary at the SDCs.

Distribution Logic: The distribution strategy involved pushing the stock from Buffer Warehouse to the SDCs up to the maximum stock level and pushing the remaining cycle stock to the DCs. SDCs would then pull the stock from DCs according to the demand. The max stock was a hard constraint meaning it cannot be violated unless there is an out of stock risk. The DCs had higher max boundaries and a soft max constraint which triggered shipments to SDCs when max at DCs is reached leading to stock rotation. This was supplemented by fair-share which ensures stock is evenly distributed in case of over and under supply.

This was achieved with the help of designing the cost model which drives the optimiser behaviour and making sure that the outcome of optimiser follows business rules thus making it predictable. 

Optimiser Performance: One key element of the project was to achieve a good performance from the optimiser runs so that results can be trusted and executed with confidence. When dealing with large optimisation problems, it is important to strike the right balance between problem size, number of constraints and runtime. It involved removing unnecessary constraints, ensuring master data cleanliness, providing sufficient runtime and breaking problems into smaller sub-problems where necessary.

Conclusion 

Optimiser is often considered as a black box which is far from true and many companies are wary of investing in Optimiser. If modelled in the right manner, it can be highly predictable and achieve great results for a business. It is also the most powerful algorithm when it comes to available supply chain planning algorithms in the SAP SCM packages like APO and IBP.

Olivehorse has a vast experience in dealing with complex optimiser implementations and has several experts in the area of supply chain optimisation using SAP APO or IBP.

If you are struggling to achieve the true potential of your current investment in Optimiser or are in the process of designing and implementing optimiser, we would love to help you with your Optimiser journey.

Should you be interested in seeing how Olivehorse can help your organisation, please get in touch now on: contactus@olivehorse.com 

Chandra Shaktawat is a Principal Consultant here at Olivehorse and has been with the team since 2015.