A full physical inventory finds every error eventually, but it usually happens once or twice a year, and it stops normal work while it runs. By the time it finds a problem, that problem may have been feeding wrong decisions for months. A cycle count fixes that timing gap. It counts one small, defined group of items on a regular rotating schedule, so every item in the store gets checked several times a year without ever closing the whole operation.
This lesson gives you the method a factory storekeeper uses to choose what to count, how often, and what to do when a count falls outside the agreed tolerance. You will apply it to Delta Garments' own fabric and trims store, using Order B's trims as the worked example, because a denim order carries several high-value components, such as rivets, zips and back-patch labels, that are exactly the kind of stock a cycle count is built to protect.
You need: the difference between book stock and physical count, and the inventory accuracy rate calculation, both from Lesson 3.1.
By the end of this lesson you can:
Not covered here: sorting a variance into theft, damage or administrative error (Lesson 3.3), or costing the impact of an inaccuracy on an order's own margin (Lesson 3.4). This lesson only teaches how to run the count itself.
| Term | Meaning |
|---|---|
| Cycle count | Counting one small group of items on a regular rotating schedule, instead of counting the whole stockroom at once. |
| Full physical inventory | Counting every item in a location at the same time, usually once or twice a year, closing normal activity while it runs. |
| Count tolerance | The size of discrepancy a store agrees to accept as normal variance before it triggers a closer check. |
| Blind count | A count taken without the counter seeing the book stock figure first, so the result is not influenced by what it is expected to show. |
| High-usage item | A stock item consumed often or in large quantity, usually counted more frequently because an error compounds faster. |
| Count escalation | Moving from a small cycle count to a full count of an item group once a result repeats or is too large to trust the cycle count alone. |
A full physical inventory is thorough but slow: an error that appears the week after last year's count will not be caught until next year's count, by which time it may have driven a dozen wrong reorder or production decisions. A cycle count trades completeness in one sitting for frequency over the year. By counting a defined group, such as Order B (the jeans order)'s high-value trims, every week instead of once a year, the same item is checked around 52 times a year rather than once, so an error is caught within days of appearing rather than within months.
| What can differ | The shapes you meet | Why it matters to you |
|---|---|---|
| Item value | High-value trims, such as Order B (the jeans order)'s rivets, zips and back-patch labels, or a fine yarn count, justify a tighter tolerance and a more frequent count than low-cost, bulky items such as poly bags. | A small error on a high-value item costs more per unit, so it deserves checking more often. |
| Usage speed | A fast-moving basic like Order A's fabric roll turns over daily; a slower item, such as a specific trim colour held for a smaller repeat order, moves once a month or less. | Fast-moving stock has more chances to drift between counts, so it usually needs a shorter count cycle than slow-moving stock. |
| Business model | A Make to Order factory like Delta Garments counts against a specific confirmed order's own requirement, so a shortfall threatens a fixed shipment date. A Make to Stock or hybrid operation counts against a rolling forecast, so a shortfall threatens the next production run rather than one dated shipment. | The consequence of an uncaught error differs, but the cycle-count method is the same in both cases. |
| Market | An export order under a fixed shipment date, like Order B's jeans, has less slack to absorb a late-discovered shortage than a domestic Make to Stock chain replenishing its own shops on a rolling basis. | A tighter deadline argues for counting the critical items more often, not for skipping the count. |
| Product family | Socks and hosiery count yarn cones and elastic by weight or reel; terry towels count yarn and trim by weight; home-textiles count fabric rolls by length; a garment factory like Delta Garments counts fabric by weight or length and trims by piece or by the box. | The unit counted changes by product family, but every family still needs a defined group, a frequency and a tolerance. |
Delta Garments' fabric and trims store holds materials for both Order A and Order B (the jeans order). Karim, the production manager, and the storekeeper agree that Order B's high-value trims, rivets, zips and back-patch labels, are counted every week, while lower-value items such as poly bags and cartons are counted only once a quarter. The agreed count tolerance for the high-value group is 5 units.
This week's cycle count of Order B (the jeans order)'s zip stock is run as a blind count: the storekeeper counts the shelf first, without looking at the book stock figure, then compares the two numbers afterwards. That order protects the count from being unconsciously adjusted to match what the system already claims.
The worked numberThe 4-unit variance is inside the 5-unit tolerance, so this result does not trigger an escalation. It is logged in the cycle count schedule as within tolerance, and the book stock is adjusted to match the confirmed physical count of 516. The value of the exercise is not this one clean result; it is that Delta Garments now has a fresh, confirmed number for a component that Order B (the jeans order)'s shipment date depends on, instead of trusting a book stock figure that had not been checked since the last cycle.
Once a counter has seen the expected number, it becomes very easy to stop counting the moment the shelf seems to match, especially under time pressure, even if a few units were miscounted or missed. That is not a cycle count; it is a confirmation exercise that will report false accuracy. The cost of this mistake is not the count itself, which still takes the same time either way; it is that a real shortage can sit undetected behind an apparently clean cycle count for months, defeating the entire purpose of counting more often in the first place.
Instead of counting everything once a year and hoping nothing goes wrong in between, a cycle count checks a small group of items often, so an error is caught within days rather than months. The most valuable or fastest-moving items are counted most often. Every count is done without looking at the system's number first, so the result is honest. If the count matches within an agreed margin, the record is refreshed and the store moves on. If it does not, the group gets a full, closer count before the small check is trusted again.
The cycle count schedule shows Order B (the jeans order)'s rivets counted blind at 1,180 units against a book stock of 1,220 units, a variance of 40 units against an agreed tolerance of 15 units. Last week's count on the same item also came in outside tolerance, short by 22 units.
Should the storekeeper simply correct the book stock again and continue the normal weekly cycle count schedule? Decide before opening the answer.
No. Two consecutive out-of-tolerance results on the same item are the escalation trigger; the storekeeper should run a full physical count of the rivet stock now, not wait for next week's cycle count.
1. A single out-of-tolerance result can be normal variance, but a repeat on the same item is a pattern, not a one-off, and a cycle count's small sample is no longer a reliable check on its own. 2. Order B (the jeans order)'s sewing line depends on this stock within two weeks, so a shortage found late, after a third weekly count, would leave far less time to reorder or substitute the missing rivets. 3. A full count of the rivets, rather than another partial cycle count, gives a confirmed total the production plan can be built on before the sewing line starts.
The method's own rule, escalate on a repeated or large variance, exists exactly for this situation. Correcting the book stock a second time without escalating would treat a real, worsening shortage as routine noise.
Order B's jeans carry a buyer target FOB of USD 9.80 per piece on 3,000 pieces, giving a total order value of:
If a rivet shortage stops the sewing line even for part of a day within Order B (the jeans order)'s fixed 22-week timeline (docs/02 section 7: 3 weeks approvals, 7 fabric, 3 sewing, 3 washing, 1 inspection and documents, 4 shipping, 1 warehouse), a delay inside the 3-week sewing stage has no slack built in before the fixed port cut-off further down the timeline. The cost of a missed trim is therefore rarely the trim's own small unit price; it is the risk to a shipment date on a USD 29,400 order that a cycle count, run and escalated on time, is built to prevent.
Use the T21-C03 Inventory accuracy and shrinkage tracker workbook, "Cycle count schedule" sheet. For each item group, record its counting frequency, its agreed tolerance, the last count date, whether the result was within tolerance, and the next action. Start with Delta Garments' Order B (the jeans order) trims example above, then add your own item groups.
[Download the Inventory accuracy and shrinkage tracker workbook](../../templates/T21-C03-inventory-accuracy-shrinkage-tracker.xlsx)
Download the Inventory accuracy and shrinkage tracker (Excel)
1. Why does a weekly cycle count catch an error faster than a yearly full inventory, checking far fewer items at once?
2. Delta Garments counts Order B (the jeans order)'s zip stock blind at 516 units against a book stock of 520 units, with an agreed tolerance of 5 units. What should happen next?
3. Another item group shows a 30-unit variance against a 10-unit tolerance, on this one count only so far. What is the correct next step?
| EN | AR | FR | TR | Meaning |
|---|---|---|---|---|
| Cycle count | الجرد الدوري | Inventaire tournant | Döngüsel sayım | Counting one small group of items on a regular rotating schedule, instead of counting the whole stockroom at once. |
| Full physical inventory | الجرد الفعلي الشامل | Inventaire physique complet | Tam fiziksel envanter | Counting every item in a location at the same time, usually once or twice a year, closing normal activity while it runs. |
| Count tolerance | هامش تسامح الجرد | Tolérance de comptage | Sayım toleransı | The size of discrepancy a store agrees to accept as normal variance before it triggers a closer check. |
| Blind count | الجرد الأعمى | Comptage à l'aveugle | Kör sayım | A count taken without the counter seeing the book stock figure first, so the result is not influenced by what it is expected to show. |
| High-usage item | صنف عالي الاستهلاك | Article à forte rotation | Yüksek kullanımlı ürün | A stock item consumed often or in large quantity, usually counted more frequently because an error compounds faster. |
| Count escalation | التصعيد | Escalade | Yükseltme (eskalasyon) | Moving from a small cycle count to a full count of an item group once a result repeats or is too large to trust the cycle count alone. |
Next you will learn how to sort a confirmed stock variance into its real cause, so the corrective action fits what actually happened. Back to the course map