Epicor Forecast Netting

In Epicor currently we have our demands loaded on a weekly basis. Our netting is happening looking 7 days forward and 3 days back. The issue we are seeing is that our orders do not come in as smoothly as our demand is loaded. So what happens for example is if we have 10K demand each week on the first week of the month we are receiving an order for 30K. We would want the system to know that 20K of that is for future weeks demands. Instead the system is decrementing just the first week and still telling MRP to order more for weeks 2 and 3. There are three solutions we thought of but each with other drawbacks.

  1. Extend our horizon – Issue here though is that they the system will potentially take demand out of future months or maybe even use orders from prior months to take demand out of the current month
  2. Have planners every week adjust the forecast - Just a lot of manual work and potential error prone
  3. Load all demand at the first of the month and have the horizons the total months – This would then pull up future orders and in essence add 2 weeks of static inventory.

We are wondering if there is some way others are dealing with this in the system without too much manual intervention.


Rory St. Lawrence
Epicor Project Manager
Maesa LLC


The key phrase you mention is

How would Epicor know that?

1 Like

I think that’s what Rory is trying to get an answer to… how do we tell the system that 20K is for the next two weeks’ demand, right? And is there any other way than the three Rory listed?

That’s exactly what my Supply Chain Manager is trying to get to. Is there any option other than the three I mentioned.

1 Like

If you do some historical analysis, would changing your horizons to 14/3 have given you better performance over the past 6-9 months? And these values should probably be looked at at least quarterly, and will almost surely need seasonal adjustment.

1 Like

i have seen some companies do a plus & minus 30 days on the forecast window… I believe that the system consumes the oldest active forecast first. Remember that MRP looks at the largest value… either the forecast, or the orders… so if you have a forecast for 100, but sales of 110, then 100 of the forecast was already being planned, and jobs/pos are created, only the extra 10 will be missing, and MRP will take that into account and plan more.
I would strongly hesitate only one forecast per month. Reason is that puts the bulk of your planned orders on that date as well, and your shop will not be very smooth. You will end up purchasing material too early, and making things too early. We had a customer who did monthly forecast buckets, and then when the orders started coming in, they always wanted to correct the timing. They ended up with many change orders on their POs to push things out. When we went to weekly forecast buckets, everything smoothed out, and many of the change orders went away.