Suppose that we’ve come up with a really sophisticated and accurate forecast of our call volume (for example), but that we know we need to amend the forecast to reflect some planned change in our business.  For example, a given call type might become automated so that 10% of your volume is expected to be dealt with by an automated system; or perhaps you have a change of risk management policy that is expected to change the volume of calls entering a credit and risk management centre; or it may be that you are running a customer satisfaction and problem resolution initiative that expects to reduce call centre volumes by 10-15%.  Having gone to the trouble of producing a good forecast, you now have to change it – but how?

Most changes of this type happen on a specific date or week, and the progress or otherwise is often more noticeable week-to-week than day-to-day.  The forecasts, then, should usually be first expressed at a weekly level.

Suppose that we make the required adjustment to the forecast and proceed to plan headcount on the basis of that revised forecast.  If the project changes, causing the forecast adjustment to be higher, lower, earlier, later, or in any way different that would affect the forecast, you need to go back to your forecast and make the adjustment.  If you’ve lost traceability of your original forecast, you’ll need to undo the adjustment already made, and then re-enter the updated adjustment.  That’s not too bad, unless there have been three other changes that have also been applied in the meantime, in which case you’d have to undo those three, then undo the original adjustment, apply the updated adjustment, and then re-apply the later three (assuming that those three don’t interact with the first).