Raw Material Order Optimisation

Raw Material Supply planning is a job which goes unnoticed until it doesn’t. When raw materials are not available to meet production plans, it can cause significant disruptions to both manufacturing and customer orders. While resilience has been a key trend in recent years, there is now a realisation that resilience needs to be balanced with supply chain costs. So what is the best practice for this?

  There are many bespoke and SaaS solutions on the market that promise inventory optimisation. In the majority of cases, these set minimum safety stock targets and re-order set quantities whenever inventory levels hit the safety target. A very small number of solution actually optimise inventory. The other challenge with many SaaS solutions is that they are black box, and are under-utilised once they are implemented because planners don’t understand or trust the recommendations.

At the end of the day, the decision that raw material planners need to make is when and how much to order for each material and factory. So how can organisations best equip their planners to optimise orders rather than inventory? Lets start with what is driving these decisions.

Demand for raw materials is driven by production plans at manufacturing sites. The raw material required is based on the bill of materials (BOM) for the product being manufactured, the run size and any shrink/waste assumption for raw materials lost at the start or end of run. However, because raw material lead times can be many weeks, planners can’t order just enough to meet the current production plans. Production plans change, and raw material supply needs to be able to cope with some amount of change. But as always with supply resilience, there is a cost-benefit trade-off.

To manage this, Purpose AI use probability distributions. We track changes in production plans over time to see how often and by how much production plans change. Then we use this to understand how often and by how much production plans are likely to change. These distributions are used to optimise the trade-off between the cost of holding more inventory and the cost of shortfall if production plans increase. These probability distributions can be done at a product or category level.

One thing to understand is that probability distributions should not just be added together. There is a 15% chance that raw material demand for the product in the chart above will increase by more than 10% (you have to sum all the probabilities for changes >10%). But there will be a smaller probability that the total demand for this raw material will increase by >10%. This is because the demand for all products is unlikely to go up by >10% at the same time, so there will be some amount of netting, or averaging out, to a lower overall probability. There is clever maths behind this that we won’t discuss here, but with these insights on the probability of demand changes, it is possible to make much more informed decisions on raw material ordering.

Before optimising raw material orders, supply factors also need to be considered. There will be a lead time from placing a raw material order to that material being available at the factory. While most raw material planning processes use a fixed lead time, it is much more accurate to also consider lead time as a probability distribution. There are frequently supply disruptions (logistics delays, unavailability of materials from primary supplier). Depending on the specifics of the raw materials, it can also be beneficial to consider the probability of quality issues upon arrival. Quality distributions are best calculated separately to the lead time distribution.

The probability distributions used for demand and supply should be tailored to the most significant factors in order to get the most optimised raw material orders. These distributions are then used to calculate the probability and impact of having a shortfall of raw materials based on different scenarios for order date and order quantity. Other cost factors such as the probability of waste raw materials if demand comes in lower, the costs of holding raw material inventory and fixed and variable logistics costs are then added to get a total order cost for each scenario. The optimum order is the order which has the lowest cost per unit delivered. These scenarios can be done for a single material/location, or across multiple locations simultaneously.

With a prescriptive, optimised order plan created for each location and raw material, it doesn’t make sense for planner’s to manually review each order. Instead, planner’s should focus on the ~10% of orders where they can add value. There are multiple criteria to determine which orders fall under this banner, and some testing is required to customise this to an organisation. But fundamentally, orders where the risk of stock-out or the cost of the order is above a defined threshold should be flagged for review. Planners can then go through this list and review if any change should be made to the date or quantity of the order.

When we implemented this solution with a Top 10 global food company, raw material stock-outs were reduced by 18%, raw material waste was reduced by 15% and inventory levels dropped by >5%. Planners trusted the system to plan >70% of orders with no human review which unlocked time for strategic projects to improve cost and resilience. This solution has proven so effective, we now offer a cost price, 4-6 week, proof of concept to show organisations how much value it could deliver for them.

Purpose AI is an AI and Analytics Consultancy delivering data-driven solutions for complex Supply Chain and Commercial decision making. We work across the Consumer Products and Retail Industries and have worked with many Top 20 CPGs and Retailers in Europe and North America. Our USP is that we have developed industrialised solutions to common industry problems, reducing the time and costs to deliver value. But unlike SaaS providers, there is no licensing costs, we customise our solutions to your requirements and embed a strong business process to ensure they deliver value.