Some might label an ‘accurate forecast’ as an oxymoron. Regardless of the validity of the joke, demand planning revolves around a company’s ability to (as accurately as possible) predict demand and then plan around it.
Thankfully, demand is not entirely random. While there is noise in most demand data, oftentimes there are also underlying trends or patterns in historical data that computers can pick up on, and then use that insight to forecast future demand. Using historical data to forecast future demand is a necessity for any company aiming to satisfy its customers, but a company can’t just choose a forecasting method at random and use that as the basis for its supply chain decisions. It is monumentally important to choose the right forecasting method, and within SAP IBP, the simplest solution to choosing the best forecast for your data is to use Best Fit Forecasting.
Best Fit Forecasting is a method that tests different models against the data in your system and ranks the various models based on the forecast errors of the outputs. The forecasting model with the lowest overall error is selected as the best fit, which will then be used to forecast future demand based on trends in historical data that may not be obvious to the user.
See Best Fit Forecasting in action with this Five Minute Feature
The importance of letting IBP choose the forecast for you can be seen in the complexities of all the different forecasting options (over 15 to choose from!) and which scenarios they are best in.
Even simple, double, and triple exponential smoothing forecasts have significant differences between them that make one better for different types of data than another; between the three, double exponential smoothing is best used when data presents a trend with no seasonality, and triple exponential smoothing is best used when data presents a trend and seasonality. Without letting a software analyze the data and make the best decision, a user may not be able to tell with the naked eye when there is a trend or seasonality or both.