Bclark:
Unless I have missed something in my travels, there is no system that I know of that calculates total demand by subtracting emergency purchases. By definition an Emergency Purchase is a receipt and not a sale. I believe what you are referring to is the level of service calculation available on most systems. (Please correct me if I am wrong!)
In regard to the scenario you presented...
First, in most markets I would be surprised if a customer was very likely to return in 1 or 2 days. I would guess that if you are the only OEM supplier in town your odds increase considerably.
Regardless of this, look at this situation this way. If we dont post a Lost Sale during the initial request, how, and do we want to put forth the added effort, of tracking every request made and then trying to determine who actually returned for purchase. On the other hand if we post a Lost Sale at the initial request and then sell the part to the same customer with in a few days, in the vast majority of cases it will make no difference to us in the long run. Thats right, it wont matter if we have two hits, because for purposes of Phase In, all sales and Lost Sales in 1 month are considered 1 demand.
In Reynolds look in 2323 or in ADP look in MNT1. You will see that the way the Phase In criteria is worded is similar to PHASE IN IF THERE IS DEMAND IN 3 MONTHS OUT OF 12 MONTHS. In actuality you could post 20 Lost Sales in a single month and a part still would not phase in if there were no demand in previous 11 months. Caution! If you use TDN or UCS, you need to keep in mind that these systems track pieces not total demand by month.
To answer you final question, at what point does a request become a lost sale? Answer: At the point an inquiry is made on a part that is not in stock and the customer does not purchase, special order, or ask you to pick up the part.
David S.