What is Lag MDT in IBP Demand?

[pravin_tikar]

Hello All,

I know little about Lag MDT which used for lag information within Demand sensing.

Could you please explain with example?

Thanks,

Pravin Tikar

[Former Member · Accepted solution]

Hi Pravin,

It is the number of weeks between the period when a forecast is calculated and the period that the forecast is calculated for. For example, if the forecast is calculated on CW10 and Lag value is 2 weeks, then the forecast will be calculated for CW12.

Hope it helps.

Best regards

Rinju

[pravin_tikar · Accepted solution]

Thanks Rinji,

I am still trying to understand. See below example.

MONTH      March.      April.     May.     June.     July

March.        125.          130.      175.       210.       225

      Lag.         0.              1.           2.           3.           4

April.                             135.      185.      220.       235

       Lag.                         0.           1.           2.           3

May.                                            170.      225.      225

       Lag.                                        0.          1.           3

Actuals.      128.          135.       172.       225

Lag 0= forecast calculated in period(x) for period (x)

Lag 1= forecast calculated in period (x) for period (x+1)

.......

Do you understand above table?

thanks,

Pravin Tikar

[anna_linden · Accepted solution]

Hi Pravin,

LAGs represent the number of periods that you would like to forecast. In the example above it is e.g. 4 months. Lag0 is always the current period. Lag 1 is the period after, etc.

But what you see on the above table is the evolution of the forecast.

Example:

For the past year, every month, you have created a forecast. And you have tracked the results  e.g, in a snapshot key figure. And now you want to compare them to the actuals to find out how reliable your forecast is in which forecasting horizon. To be able to have comparable results, you want to track the reliability of the forecast e.g. in Lag2 or Lag4. Because of the nature of the forecast, you will normally always have better results in the near term horizon.

You would e.g. try to compare always the lag2 forecast results versus the actuals. That is done during measurement of the forecast accuracy.

Best Regards,
Anna

[pravin_tikar · Accepted solution]

Thanks Ana for nice Note.