A cycle count sheet should capture four things: what the system said, what was actually there, the difference, and the likely cause. Recording only the correction throws away the one piece of information the count produced. A line that comes up short three times is telling you about receiving, picking or shrinkage, and you cannot see that without the cause column.
When to use it
Use it for weekly rolling counts rather than an annual full count. Counting a few lines often finds errors while the delivery, the picker and the paperwork are still identifiable.
Prioritise by value and movement. Fast-moving, high-value lines are where drift costs most and compounds quickest; a slow, cheap line can wait months without anyone being worse off.
The template
One row per line counted. The cause column is the reason the sheet exists.
CYCLE COUNT
Date: [date] Counted by: [name] Location: [where]
Counted before opening / after closing: [which]
SKU | Description | System | Counted | Diff | Likely cause
---------- | ------------- | ------ | ------- | ---- | ------------------
| | | | |
| | | | |
| | | | |
CAUSE CODES
R receiving - not counted against a PO, or counted wrong
P picking or dispatch error
D damage not written off
A previous adjustment with no recorded reason
U unknown - investigate if it repeats
REPEATS
Any SKU appearing here for the second time in 90 days:
[SKU] - [how many times] - [what we are doing about it]
ADJUSTMENTS MADE
[ ] System updated
[ ] Causes recorded above BEFORE adjusting
[ ] Repeats flagged to [name]How to adapt it
- Keep "causes recorded before adjusting". Once the number is corrected, nobody goes back and writes down why.
- Keep the U code. Forcing a guess produces bad data; an honest unknown that repeats is itself a finding.
- Count before opening or after closing. Counting a location while picking is happening in it produces discrepancies that are artefacts of the count.
- Add a batch or expiry column if you handle dated stock.
Why it is this short
A long count sheet does not get completed on a busy afternoon, and a count that is abandoned halfway is worse than none because the partial data looks complete.
The repeats block is small and does most of the work. One discrepancy is an incident; the same SKU three times is a process fault with an address.
Common questions
What should a stock count sheet record?
System quantity, counted quantity, the difference, and the likely cause. The cause is the part that turns a count into information rather than a correction.
How often should I cycle count?
Weekly, a handful of lines, prioritised by value and movement. Fast-moving high-value lines might be monthly; slow cheap lines once or twice a year.
Why record the cause before adjusting?
Because once the number is corrected nobody goes back to write down why, and the information the count produced is lost. The same discrepancy then returns indefinitely.