Generic selectors
Exact matches only
Search in title
Search in content
Search in posts
Search in pages
Filter by Categories
Glossary
News

Probability applied to Capacity and Material Planning

We discussed in an earlier article the use of “Probability instead of Accuracy as a Measure in Forecasting” and now we want to take this one step further and explore the use of probability information derived from the forecasting process for Capacity and Material Planning.

The Problem Statement

Demand for items is subject to changes from one period to the next; some items have higher and some items lower volatility, but the demand changes! If responsiveness of supply cannot be achieved due to long-lead times, short product lifecycles and the need to manage production commitments in the mid-term (slush period) for external suppliers, then the probability distribution of demand (see Article: “Probability instead if Accuracy as a Measure in Forecasting”) must be better understood and decomposed.

Planning capacity or material in volatile and changing demand situations will result into a high capacity and material plan instability for the suppliers and factories. Most measures such as plan adherence only measure against the last plan committed to at the release point. In addition, the history of how a plan and its underlying demand has evolved across multiple planning cycles is usually not available to judge the quality of a plan. Hence any measure of the quality or inherent risk profile of a plan is impossible.

Given that the biggest challenge of any planner is what level of capacity he or she can commit to, traditional methods are of little use; hence the probability method of assessing the underlying risk to a capacity and material plan come into play.

Method of how to apply Probability to Planning Decision

The method is straightforward. A capacity or material plan for any planning period (a week or a month) is composed of a mix of product consuming this capacity or material. This mix changes from one planning period to the next across the planning horizon. Each product consumes the capacity or material in discrete multiples of unit, for example a piece or a pair of shoes.

Two factors come into play:

  • what is the volatility of the mix of products, and
  • what is the volatility of the quantity of a single product within the mix

Answering these questions, when looking just at a single plan is impossible. As a consequence, we must derive the answer by:

  • Understanding the mix stability across the last n planning cycles (probability of a mix change)
  • Understanding the probability distribution of a single product both for the last plan AND across the last n planning cycles and may be extended to like products historically during the product launch stages

Different Method of deriving Probability

To understand the mix stability the method applied is the analysis over n planning cycles of:

Calculate the CoV (Coefficient of Variance) for each product in the mix and derive the capacity consumption impact per product and the associated CoV. This method will provide a good understanding of the percentage of capacity driven by “stable” products (those with low CoV).

Adding the second layer of probability, the method for understanding probability distribution of the quantity of a single product with the mix was described in the article mentioned above. This method will provide a good understanding of the percentage of quantity driven by “stable” demand. Other method for analyzing the demand can also be applied.

The Result

Both methods together will provide the planner with a powerful way to determine the baseline load for a supplier or factory. In some examples where we applied this method we found that between 15% to 20% of the products planned in a period, they drove 55% to 65% of the capacity. Close to a Pareto!

Allowing the loading of a factory to this level, and at the same time committing to this baseline load, also means that the planner can spend more time dealing with the volatile products and quantities. Working with sales and merchandising, this information provides a valuable base for analyzing the sales and marketing risk and leads to better decision on both sides on how to manage this risk.