What This Article Covers
- How does best fit work?
- How does it work in SAP DP?
- What is he problem with best fit in SAP DP?
Best Fit Forecasting?
Best fit forecasting is provided in most supply chain forecasting applications, which fits a forecast model to a demand history. A full definition of can be found here.
How It Actually Works
This essentially sets up different forecasting methods in competition with one another per demand history trend line. The software goes back in history and uses previous period to forecast more recent period for which the actual demand is known. By comparing multiple forecast methods and compare the error between the forecast and the actual, the software picks a forecast methodology which “best fits” the historical trend.
In SAP SCM DP, this can done by selecting the Auto Model Sel 1 or 2 on the Model tab of the planning book.
The only decision the planner has to make is the time horizon which the best fit calculation will be made. This is controlled under the horizons tab in the planning book. Obviously, different time selections can yield different results. However, at account after account I am finding that this functionality does not work. That is returns the constant model, and so clients are not able to use it. Here are two samples.
As you can see, the blue line, being the forecast does not match the yellow line which is the historical pattern. My client noted
Auto Selection Model seems to always choose a constant model for the statistical forecast, even when there is a clear seasonal and trend pattern. This does not meet the business’s need to account for seasonality and trend.
I have told another forecasting vendor about the problems with best fit in SAP DP, and they find it almost impossible to believe that a commercial forecasting system would have such a problem performing a best fit analysis. I get that response a lot with SAP DP.
On SAPExpert, an article recommends using a macro to perform best fit functionality, however, this is completely unacceptable. Best fit functionality is “core functionality” for any enterprise forecasting software, and the fact is it should simply work.
However, I have found that best fit can be made to work in SAP DP, but it must be run in batch mode. If best fit can be made to work, SAP has good mathematics, but the issue is that it is so difficult to get to work. I would prefer to run best fit outside of SAP, in software that is much better at doing it, and then coding the forecasting method to be used in SAP DP after the fact. This can allow best fit to be run much more often, and much more efficiently.
References
Wikipedia
Supply Chain Strategy, McGraw Hill, Edward Frazelle
Sales and Inventory Planning with SAP APO by SAP Press
Supply Chain Management for Advanced Planning, Springer Press
What This Article Covers
Do you have experience using best fit in SAP DP? If so, please comment below about how well you think it works.
Pingback: DP « SAP SCM Planning
Hi,, can I got a further explanation on this?
I’m highlighting your quotes saying that “Research over time indicated that in many cases sophisticated models could not defeat simple time series techniques.”
So were you saying that the Causal Model is always better than Time Series Model?
Is causal model available in SAP APO?
Looking forward to have your reply at your earliest convenient time. Or you may email me personally.
Thanks, benhard
Hi Ben,
Actually, I am somewhat proposing the opposite. This research is SAP independent. For decades software companies and consulting companies have been promoting “advanced forecasting” when in fact forecasts have not improved much. Wall Street in a way proposes the same thing…advanced models and math can provide for better forecasts and in therefore allows them to take on more risks — and we can see how that worked out.
SAP DP has multiple regression capability that can be set in the Forecast Profile, and on the MLR tab. You can enter any number of causals. However, it must be lacking in some way because SAP invested heavily in SPP causal forecasting for service parts and gave it an entirely different name. SAP has done a a poor job of explaining how this functionality is different and right now I don’t have the time to research it.
But all of this skirts an important issue. The vast majority of companies do not have even one causal (that is accurate) that they can put into these models. And this is one of major reasons that causal and MLR forecasting is much more talked about, than actually implemented.
Good point, Shaun
Except for GUI (which is very important for Demand Planning – SAP would deny this of course:)) and some key functionalities, having “complex” algorithm (or even AI) will not help in getting a good forecast
They key to have a good forecast is:
1. Process (you are right about that) but not just SO&P (where most consultants are making a living) but also end-user process (how do you help an end-user with a maximum of a college degree recognize when forecast is good?)
2. Data quality but not only initial clean-up but also consistent data management (no company does this IMHO)
3. GUI, GUI, GUI – it is somewhat tied to point #1 above
Keep the fire going:)
I find it strange that S&OP would be such a big area right now, since there is absolutely no equality in power between sales, finance and operations. If sales and finance are so much more powerful than operations, and in essence dictate to operations, how can software bring them all together to make shared decisions. Maybe someone can answer this for me.
Well, it is a myth:) However, depending on WHO implements S&OP process such a marriage (Sales, Operations, Finance) is doable.
Of course, it also heavily dependent if there is an influential insider champion (C-level) who would ensure that the process is followed after consultants are gone.
Software is only a facilitator with a role providing a convenient easy-to-use platform to facilitate and encourage the process
This is why software vendors are never successful in this:)
Pingback: DP – Demand Planning « SAP SCM Planning
Pingback: SPP DRP Consolidation and Forecast Profile Maintenance « SAP SCM Planning
Pingback: DP and Best Fit Forecasting Definition « SAP SCM Planning