Talk to an expert

Tell us what you need and we will call you back.

🇮🇳 +91

    Stock audit glossary · Counting methods

    Cycle Count

    noun · also "cycle counting"

    Definition

    A cycle count checks part of the stock at a time on a rolling schedule, so that every item is counted over a period while the site keeps working. Groups of locations or items are counted in turn, usually with high-value and fast-moving lines counted more often. It gives continuous checking through the year, but no single date on which the whole stock was verified.

    • Reviewed by CA Sundram Gupta · Chartered Accountant
    • Last reviewed 23 October 2026
    • 6 min read

    See related terms

    At a glance

    Counts
    Part of the stock at a time On a rolling schedule
    The site
    Keeps working Only the group being counted is frozen
    Frequency
    Set by value and movement High-value lines most often
    Gives
    Early warning of errors While the cause can still be traced
    Does not give
    One dated position A full count does that

    Meaning

    01 What a cycle count is

    A business that cannot stop trading has no day on which everything stands still, so a single full count is costly or impossible. A rolling programme solves this: the stock is split into groups, each group is counted in turn, and every item is reached over a period rather than at one moment. The schedule is weighted so that high-value and fast-moving lines come round far more often than slow ones.

    The result is continuous checking of how well stock is recorded. Differences show up soon after they arise, while the documents that caused them are still at hand, instead of a year of errors appearing at once with no way to trace them. The limit follows from the same feature: different areas are counted on different days, so no single date exists on which the whole site was verified. Where a dated position is needed, a full count supplies it.

    At the audit

    02 What an auditor checks in a cycle count programme

    Where a business counts in rotation, the programme itself is checked.

    What the programme should show

    • Coverage, two ways. How much of the site has been reached, by number of locations and by value.
    • When each group was last counted. So a group that has not come round for months is visible.
    • What happened to each difference. Investigated, corrected and posted with its cause.

    Warning signs

    • Groups that dropped out. The programme reports plenty of counts while part of the range is never reached.
    • Zones counted while picking went on. The result cannot be reconciled with anything.
    • Differences adjusted away. The count becomes ledger tidying rather than a check.
    The auditor tests the programme, not just the stock. A sample of recently counted locations is recounted independently to see whether the programme’s results hold.

    In practice

    03 How a cycle count programme is built

    ItemsCounts a yearCounts needed
    Group Items × counts a year Sample counts
    A: high value 180 × 12 (monthly) 2,160
    B: middle 540 × 4 (quarterly) 2,160
    C: low value 2,880 × 1 (once a year) 2,880
    All 3,600 items Every item reached within the year 7,200

    2,160 + 2,160 + 2,880 = 7,200 counts a year, about 600 a month. Here group A holds 72% of the value, B 22% and C 6%. Sample figures for illustration only.

    Five steps, in order

    1. Split the stock into groups

      By location, category or value band, and give each group a counting frequency, tighter for high-value and fast-moving lines.

    2. Freeze the group being counted

      No picking or replenishment in that group while the count is on.

    3. Count without the system figure

      The counter is not shown the expected quantity, so the count is independent.

    4. Recount and investigate differences

      At once, while the goods are still there, then trace each one through the movement history.

    5. Post the correction and track coverage

      Record the cause, and keep a running record of which groups were counted and when.

    Worked example

    04 One month of cycle counts

    Sample results from the schedule above.

    A distribution centre’s 600 counts in one month, sorted by result
    Result Counts What it means
    Agreed with the system 576 No action needed
    Receipt or issue not yet posted 15 Timing: corrected once posted
    Goods in the wrong location 6 Found nearby, moved and re-recorded
    Real shortage 3 Traced through the movement history

    The differences matter more than the accuracy figure

    576 of the 600 counts agreed, an accuracy of 96%. The value is in the 24 that did not: most were recording delays, which point to a process to fix, and three were real losses that need explaining.

    Because the counts are spread across the month, these causes surface while the documents are still at hand. An annual count would show the same differences all at once, with the trail gone cold.

    Watch out

    05 Common mistakes with cycle counts

    A rolling programme fails quietly, which is why these are costly.

    Counting every line equally often

    Most of the effort lands on slow items where nothing has changed.

    Weight the schedule by value and movement.

    Adjusting differences without a cause

    The same difference comes back next cycle.

    Trace each difference before posting it.

    Counting while picking goes on

    The result cannot be reconciled.

    Freeze the group for the length of the count.

    Quoting a dated position

    No single day covered the whole site.

    Use a full count where a dated figure is needed.

    Groups falling out of the rotation

    Activity looks high while part of the range is never reached.

    Report when each group was last counted.

    Counting the easy locations

    Tidy, reachable racks are counted again and again; awkward ones wait.

    Fix the calendar in advance and hold to it.

    Questions

    06 Cycle count FAQ

    How to do cycle counts?

    Group the stock by value and movement, give each group a counting frequency, and draw up a calendar that reaches every item within the period. On the day, freeze the group being counted, count without being shown the system quantity, recount any difference at once, and trace each difference through the movement history before posting the correction.

    How is cycle count calculated?

    Two figures are usually worked out. The counts needed in a year are the items in each group multiplied by how often that group is counted, added across the groups. Count accuracy is the number of counts that agreed with the system divided by the number of counts made; 576 agreeing out of 600, for example, is 96%.

    How much cycle count is normal?

    There is no fixed figure. Each business sets its own schedule, usually counting high-value and fast-moving items most often and slow, low-value items less often but still within the period. What matters is that every item is reached and that the coverage can be shown.

    CA Sundram Gupta

    Reviewed by CA Sundram Gupta · Chartered Accountant

    Patron Accounting LLP. This page makes no regulatory statement.

    Last reviewed: 23 October 2026 Applies to: businesses that count stock on a rolling schedule, and the auditors who rely on it

    Sources: general stock-count and audit practice; no statutory statement is made on this page. General information, not professional advice.

    Need help with cycle counts? See our stock audit service. A stock audit tests a rolling programme as well as the stock.

    Talk to an expert

    Tell us what you need and the right person calls you back.

    🇮🇳 +91

      By sending this form you agree to our privacy policy.