Why the shutdown count is the wrong tool
A full count once a year means closing the doors, paying overtime and counting in a rush. The result is one accurate day, followed by eleven months of drift that nobody sees. When the count finally shows a large shortage, it is too late to find out when or why it happened.
Cycle counting turns the problem around. You count a small part of the warehouse every working day, so every bin is counted several times a year, and each difference is fresh enough to trace.
Decide what to count, and how often
Not every bin deserves the same attention. Count most often where errors cost most: fast movers, high-value items, items that are easy to confuse, and anything that has shown differences before. Slow, cheap items can wait. A simple way to start is to rank items into three classes and give each class its own cycle.
TipStart with a daily target your team can always meet; a small count done every day beats a large one that keeps being postponed.
Freeze the bin, not the warehouse
Warehouses close for a count because of cut-off: if goods move while you count, the book figure and the shelf figure describe different moments. You get the same accuracy by freezing only the bins you are counting, for the few minutes it takes to count them.
- Post every receipt, pick and transfer that touches the bin before you start.
- Take the book quantity for the bin at that moment.
- Hold picks and put-aways for that bin until the count is entered.
- Count, enter the result, and release the bin.
- Count early in the day or between picking waves, when the bin is quiet.
TipIf a pick is urgent while a bin is frozen, post the pick first and count the bin after it, rather than letting stock move mid-count.
Count blind, by unit and by lot
Counters who can see the book quantity tend to find it. Give them the bin and the item, not the expected figure, and compare afterwards. Ask them to count in one stated unit, pieces or cartons, so a carton of 24 is never entered as 24 cartons.
For lot-tracked goods, count each lot separately and note its expiry date. A bin that is right in total but wrong by lot will still send the wrong carton out next week.
Recount before you approve
Set a tolerance for each class, then send any line outside it for a second count by a different person. Most large differences turn out to be a miscount, a carton in the wrong bin, or a receipt nobody posted. A recount finds these before they reach the books.
Approval should sit with someone who did not count, and every difference should carry a reason. A count approved without reasons tells you the stock is wrong but not why.
TipTrack count accuracy as the share of counted lines within tolerance: if 46 of 50 lines match, accuracy is 92%.
Post the difference, then find the cause
Once a count is approved, post the difference as a stock adjustment at cost, so the shortage or surplus shows in the books in the month it was found. Do not carry it on a spreadsheet until year end.
Then look for the pattern. Repeated differences on one item usually come from a process, not from theft, and fixing the process is what makes the next count clean.
- Picks entered in cartons that left as pieces, or the reverse
- Returns put back on the shelf without being received
- Transfers or van loads moved but never posted
- Receipts put in a different bin from the one recorded
- Damaged goods thrown away without an adjustment
Vans and branches count too
Every place that holds stock needs a count cycle, not only the main warehouse. Branch warehouses can follow the same class plan on a smaller scale. In van sales, count the van’s stock at the end of every day before the cash, so a shortage is found on the day it happens and the rep can still explain it.
Checklist: Cycle count checklist
- Rank items into A, B and C classes
- Set a daily count target the team can meet
- Post every move for a bin before counting it
- Hold picks on the bin until the count is entered
- Count blind, in one stated unit
- Count lot-tracked items lot by lot
- Recount every line outside tolerance
- Approve with a reason for each difference
- Post adjustments in the month they are found