How "LAG based Forecast error calculations in Demand Planning" scenario works ?

[former_member666725]

Hi IBP Experts,

I have to configure a scenario using "LAG based Forecast error calculations in Demand Planning".

I would like to know how to execute the LAG based Forecast error calculations scenario and how to interpret the data and understand the scenario how it is working ?

Thanks in Advance for your help.

Parthi

[lingaiahvanam · Accepted solution]

Hi,

From a business point of view, there are several possibilities in which forecast errors are calculated on which level and for which lags.

Within SAPIBP1 the following assumptions are made:

For Demand and S&OP purposes:

• Forecast Accuracy is measured for Statistical Forecast Qty, Consensus Demand Plan Qty, Consensus Demand Final and Combined Final Demand
• Forecast Error is calculated and stored in monthly buckets for 1 Month and 3 Months Lag
• Beside of the forecast error, the corresponding forecast bias is calculated and stored as well
• All these calculations are done on Product/Location/Customer-Level
• Exactly one forecast error measure is used (e.g. MAPE)

Statistical Forecast Qty:

• STATISTICALFORECASTQTYLAG: Lag-specific key figure, means “LAG” is used as root. Used to store the snapshots of key figure STATISTICALFORECASTQTY for all relevant monthly lags (in our example 1M lag + 3M lag).
• STATISTICALFORECASTQTYERROR: Contains the calculated Statistical Forecast Qty Error per month and by lag (in our example 1M lag + 3M lag)
• STATISTICALFORECASTQTYBIAS: Contains the calculated Statistical Forecast Qty Bias per month and by lag (in our example 1M lag + 3M lag)

An illustration of how the STATISTICALFCSTQTYLAG data is retrieved from STATISTICALFCSTQTY by using the Lag Based Snapshot operator. LAG = 1:

With the above statistical lag data, create new ley figures and wire calculation actual demand vs statistical forecast lag key figures data to calculate the "Forecast Error Calculation".

An example represented same applicable to customer requirements but overall process remains same.

Best Regards,

Lingaiah