sgmentationForecasting is rarely anyone’s favorite activity.  In fact most consider it a pain, or possibly a black art.  It is even common for companies to say they do not forecast because it is impossible in their industry (absolutely not true).  When talking about MRO stock these types of attitudes are even more evident than with regular stock.  The problem, of course, is that forecasting is necessary for people within the company to do their job and if there is no formal forecast, people will just make it up as they go, which means no consistency between departments or processes.  Not good for efficiencies, effectiveness, or optimization of your processes.  Since MRO stock tends to have much higher SKU counts and much more erratic demand patterns it is even more difficult to forecast.  One of the more effective methods for forecasting materials with erratic demand patterns is a method called segmentation forecasting.  The basic assumption for this method is that the reason for the erratic demand is that there are several different types of demand influencing the total demand quantity.  The obvious answer to that is to determine what types of demand there is for the material, forecasting each of the demands separately and then adding all the forecasts back together.  Since MRO stock often has different types of demand patterns this is an excellent methodology to use for it.

The first step, obviously, is to first determine what types of demands the MRO stock might have.  While the answer to that would be specific to any given industry or organization I would expect to see some or all of these types of demand.

  • Preventative Maintenance – Demand that is created by participation in a preventative maintenance program
  • Breakdown Repairs – Demand that is created by unscheduled breakdowns of the machinery or process
  • Project part – Demand that is created by specific projects within the organization (one time or short term)
  • Direct Order part – Demand that is created outside normal requirements and process for specific but non-normal usage.

The second step would be to define a forecasting process for each of these types of demand. For example, Preventative maintenance demand can be directly calculated by the computer if a firm Preventative maintenance schedule and individual Maintenance BOM’s exist. (There actually may be multiple BOMS for a particular process if you do major / minor maintenance. So long as the schedule notes which type and BOM to use there is no problem. For instance, if you do oil changes twice a year but every forth maintenance cycle you also change the transmission fluid then the type one BOM would only reference a quantity of oil but the type two BOM would show both oil and transmission fluid.) Breakdown repairs would probably be a straight statistical process with a sliding scale based on age of the equipment. Project and Direct Order demand would probably require a manual forecast from the people creating the demand. In order to control the processes it would be a good idea to set up a RACI chart detailing forecast responsibilities.Segmented Forecast – Part # 123

The third step would be for whoever is responsible (refer to RACI) to actually run each of the separate forecasts as per the processes defined in the second step. These forecasts would then be gathered up and added together for each separate SKU. This would give you a forecast chart for each SKU that looks something like this:

This four are relatively common in any company. There may well be other types of demand that are specific to your industry.


Total Forecast Preventative Maintenance Breakdown Repairs Projects Direct Orders











Mar 23 20 3  



100 20   80  
May 48 20 3  



20 20      


23 20 3  





Sep 28 20 3  



130 20   110



23 20




40 20    


Total 503




The total forecast is treated the same as any other forecast and can then be shared with other interested departments such as purchasing and finance to be used for their purposes as well as serving as a basis for determining inventory stocking levels, replenishment triggers, etc.

If you would like to discuss more about Forecasting either MRO or regular stock and how to improve it in your organization, I would be more than happy to meet with you and to explore this in more detail.  Please feel free to contact me at

If you would like to read more about this and other topics check out my other postings on my website –

%d bloggers like this: