In my 10+ years being a Demand Planning consultant, I am often told the story of how the lack of forecast accuracy is the core issue causing the company all of its inventory woes.
While it is a fact that low forecast accuracy can indeed cause a host of issues in a company’s supply chain, I find that many companies forget about the fact that the forecast must travel to the planning system in order to be useful. You see, it is HOW the forecast gets integrated into the supply plan that really makes a difference to the inventory situation, and this is influenced heavily by the strategy that is executed around the consumption of this forecast by the incoming sales orders. It is also imperative that this “consumption strategy” aligns with the information that is being fed to your forecasters by your sales staff. What I mean by this is that if your forecasting system uses monthly buckets, for instance, then your forecast consumption strategy must take this into account.
It is here that conflicts typically occur as the demand side of the balance equation typically wants to align their data with the sources of that information (the sales team, which usually plans in months or weeks) while the supply side wants the forecast data to be congruent with the supply/capacity planning cycles (which are usually shorter periods like weeks or days).
If the sales input is collected at the monthly level, but then subsequently broken up into weekly/daily buckets to make the information easier to consume for the supply planning process, additional care must be taken to ensure that the original intent of the forecast was maintained.
Top 5 Tips on forecast consumption:
Make sure that your consumption strategy aligns with your forecast inputs:
If you sales people are giving you forecast information in monthly buckets, be very careful about breaking that quantity into weekly buckets or smaller as this may negatively impact how that forecast gets consumed
Pay careful attention to your consumption periods:
As we all know, a forecast, by definition, is wrong. Knowing this fact, the consumption periods are designed to aid you in mitigating this error by allowing the incoming sales orders to “drift” backward and forward to ensure that they find a forecast requirement to consume. Imagine this: if you release a monthly forecast number with no forward/backward period, then ALL sales orders that were anticipated by this forecast must come in during that month or they will be considered to be incremental demand (even if they come in 1 day after the end of the month). We, as forecasters, cannot account for things like clients taking vacations, natural disasters, etc. that keep orders from being entered, but we can protect ourselves from those effects.
Make sure that your retention strategy aligns with your business:
If you have set the backward consumption period to allow consumption to go back 30 days but you are deleting all unconsumed requirements that are 7 days or older, there will never be an instance where consumption will happen beyond that 7 days.
I’ll be honest here…I’ve never understood why SAP forecasting and APO Demand Planning are bucketed based tools that then feed a moving window planning system. This tends to really confuse the sales people who have been asked to provide the input to the forecast. The bucketed consumption solution is a new solution delivered by SCM enhancement pack (1) that can address this issue. When you release a forecast for 1000 units to the month of March, only sales orders that were entered in March can consume that forecast. This obviously has issues with orders that are entered before/after the month, but it really helps align the sales/forecast/production teams (which is the point, if I’m not mistaken).
These are not “set and forget” settings:
I have seen many clients bring in consultants to subdivide their product base and then assign specific consumption strategies to the various buckets. This approach is a good one as long as the client understands that it is the process used to determine those buckets and the products assigned that are the real benefit. The client MUST regularly redeploy those strategies and redefine those buckets to ensure that products whose profiles have significantly changed recently are handled appropriately. It would be great if APO would just do the analysis for us and assign the consumption parameters for us… but this is just simply not the tool we currently have (without significant customization).
Click to read my next blog where I will outline a real life scenario where misunderstanding of the concepts above led to poor forecast consumption and increased inventory and supply chain costs.
A Reference of Forecast Consumption Settings:
- Consumption Mode
o Backward: Sales requirement only searches backward in time to consume an appropriate forecast requirement
o Forward: Sales requirement only searches forward in time to consume an appropriate forecast requirement
o Backward/Forward: Sales requirement first searches backward in time to consume an appropriate forecast requirement; if none are found, the search continues in the forward direction
o Forward/Backward (exists in ECC only): Sales requirement first searches forward in time to consume an appropriate forecast requirement; if none are found, the search continues in the backward direction
o Bucketed (exists in SCM only and requires EHP1): Sales requirement searches for appropriate forecast requirements to consume inside a specified planning bucket (e.g. monthly buckets)
- Backward Consumption Period
o Defines the number of days backward that the system will search for a suitable forecast requirement before ending the search or continuing in the forward direction (depending on consumption strategy)
- Forward Consumption Period
o Defines the number of days forward that the system will search for a suitable forecast requirement before ending the search or continuing in the forward direction (depending on consumption strategy)
- Unconsumed Forecast Deletion Period
o This is the number of days in the past after which unconsumed forecast quantities are either deleted or rolled forward into a future period. Typically, this period is at least as long as the backward consumption period in order to allow for “late” sales orders (i.e. later than the forecaster anticipated)
Have questions? Reach out to us … we’re happy to talk!