Cycle counting instead of the Sunday audit
The full stock count is a ritual: shop closed, everyone in on a Sunday, a number at the end that nobody quite believes. There is a better way to spend the same hours, and there is published support for it.
What the research actually supports
Be careful here, because this topic is thick with invented statistics. What is genuinely supported is this: in their study of inventory record inaccuracy, DeHoratius and Raman identify inventory auditing practices as a factor that mitigates record inaccuracy.
That is a real, peer-reviewed basis for "count more often." It is not a promise of a specific improvement, and anyone offering you one should be asked for the publication.
Frequency by value, not by convenience
The standard approach is to count different things at different rates according to what they are worth — an ABC classification. This is ordinary practice rather than a research finding, and it should be presented that way.
| Class | Roughly | Count |
|---|---|---|
| A — high value or fast moving | top ~20% of value | Monthly or quarterly |
| B — moderate | next ~30% | Two or three times a year |
| C — low value, slow | the long tail | Annually |
Two refinements are worth more than the classification itself. Count a line more often if it keeps being wrong, regardless of its class — persistent drift is information. And count fast movers when they are low, because that is when the count is quick and when the number matters most.
Post the count line by line
This is the part most businesses get wrong even after adopting cycle counting, and it wastes most of the benefit.
If a count ends in a single reconciling adjustment — one entry that brings the total back into line — you have learned that you were wrong. If it is posted line by line, with the counted figure against the expected figure for each item, you learn which items drift, in which direction, and by how much. That is a diagnosis rather than a correction.
Drift with a pattern almost always has a cause you can find: a line sold in pieces and received in cartons with the wrong conversion, two products that look identical on the shelf, a variant that staff ring up as its sibling, breakage that nobody records. None of those are visible in a total.
Corrections should be movements
A related point about how the adjustment is recorded. If a count edits the stock figure in place, the history of that item now says it always held what you just counted. The evidence of the drift is gone, and with it any chance of finding the cause.
If instead the correction is written as a new movement — a dated adjustment, attributed to a person, with the before and after — then the record of what happened survives, and the third time the same line drifts you can see the first two.
You probably still need the annual count
Cycle counting complements a full physical count; it does not automatically replace it. Whether you can stop doing the annual count depends on your auditor, your reporting framework, and how much confidence your cycle counting has actually earned. Many businesses run both — the annual count for the accounts, cycle counts for the operation.
The realistic promise is not that you will never close the shop again. It is that when you do, the number at the end will not be a surprise.
Sources
- DeHoratius, N. & Raman, A. (2008). Inventory Record Inaccuracy: An Empirical AnalysisManagement Science 54(4), 627–641
Figures on this site are attributed to the organisation that published them, with a link. Where sources disagree, we say so rather than choosing the flattering number.