When we did our Snapshot on Physical Inventory day in December, it did not pick up transactions that had happened that day.
We did all job closings, received jobs to inventory (for the SO that were sitting in the warehouse ready to ship), PO Receipts and SO invoices then ran Capture COS/WIP before we ran Start Count Sequence so everything should have been posted.
However, when we got to January and started doing our normal finished goods cycle counts, we found that they were off by the qty we reported to inventory the day of physical inventory.
We have come up with some possible causes to this but they don’t all make sense:
1 - Employees weren’t all logged out of jobs
2 - At Physical Inventory, the systems goes back a designated # of hours prior to grab the snapshot.
3 - we are missing something that we need to be doing.
This would be the first thing I look for and has bitten us in the butt before. One year we tried to do more steps ahead of time so we could start counting earlier on the day of. Any part/location combo that was empty the day before the count but then had parts moved into it were all considered write-ups even though a proper snapshot would have captured the starting quantity. We had to pretty much go back and find all the affected tags and figure out the actual snapshot quantity so we could back into the correct adjustments.
I will say we had a precount exercise where we counted sealed boxes and put a label on the tape. If the label was unbroken, we’d take the count off the box. It helped somewhat and helped to identify slower moving items if they have multiple labels over several physicals.
I have some skepticism. Not in that process precisely, but in the idea that the counters (people) will take the notion of “I don’t have to open boxes anymore!” and run with it.
In the end, it’s all about management. People that (professionally) care have to oversee the counting.
Last year they had the QA team be the PI supervisors on the floor. I thought that was a clever choice. I used to work in quality, and if anyone loves to point out other people’s flaws and mistakes, it’s QA!
In a previous life, @Rick_Bird is supporting them too. Every morning they shutdown the warehouse, which is already locked down to warehouse workers, and they count every part that was active the previous business day along with random parts due to be counted. If an error occurred, it was usually on the previous day.
The tags were downloaded and sorted by warehouse location. Each counter would get the next tag (tags for serialized parts), and then those counts were pasted back into the Cycle Count system. It was pretty slick.
Preparation does not help!
