Retail & hospitality operations

Retail Cycle Count Process: Seven Steps from Scope to Sign-Off

Marko Čačić Co-founder
Store employee counting boxes on a drugstore shelf with a blank count sheet on a clipboard

The system says six, the shelf holds four, and nobody can say when the difference started. A yearly count finds the gap months late, when deliveries, returns and transfers have blurred together. A retail cycle count process counts one planned part of the store at a time instead, so the cause is still traceable when you find the gap.

This guide gives store teams a routine for one count: choose what to count, set a cut-off, count blind, recount, trace the cause, adjust the record and close the count. It also covers who does what and how to handle stock that moves while you count.

The store team counts and follows up. Stock quantities stay in your inventory system, and any year-end count your accountant requires is a separate matter. It is one routine inside our wider guide to retail operations.

Cycle count, full count or spot check?

Stores use three kinds of count, and they answer different questions. A full count covers everything at once, usually a few times a year. A cycle count covers one planned slice at a time, so the rest of the store keeps trading. A spot check is a one-off count of a single product or bay, run because something looks wrong.

Shopify’s cycle counting guide describes a cycle count as counting a small group of products at scheduled times, with less disruption than a full physical count. Because each count covers only a slice, the schedule decides whether everything is eventually checked. Check what your system does with the items you did not count: Lightspeed Retail (R-Series) lists them as missed and makes no change to them unless you zero them, so in a partial count leave them alone.

A cycle count does not always replace the yearly count. Your accountant or local rules may still require a full count, so ask before you rely on cycle counts alone.

The retail cycle count process at a glance

One count, seven steps, each with a result you can check:

  1. Choose the scope: which part of the store, which products and which day.

  2. Set the cut-off: stop movement for that scope and note the system quantity.

  3. Count blind: the counter writes down what is there, without seeing the system quantity.

  4. Recount: a second person counts every difference beyond the tolerance again.

  5. Trace the cause: check deliveries, transfers, returns and holds before changing anything.

  6. Adjust: one named approver changes the record and gives a reason.

  7. Close and review: sign and file the sheet, then check the results every week.

Three roles run the steps: a counter, a checker who recounts and traces, and an approver who signs off. In a small store one person may hold two of these roles, but whoever approves an adjustment should not be the person who counted. Poskio’s guide recommends the same separation.

Choose what to count this week

A cycle count starts with a decision made before anyone picks up a clipboard: what is in this count. Pick one of four ways to choose, and write down which one you used.

Way 1
By value
Count the products that matter most more often, and the slow, low-value ones less often. Effort follows the money.
Way 2
By place
Count one aisle, bay or room at a time until the whole store is covered. It is the easiest plan to run and the easiest to check for gaps.
Way 3
By trigger
Count an item after something happens to it: a delivery, a transfer, a return, an adjustment, or a scan that does not match the stock on hand.
Way 4
By exception
Count what the system flags: negative stock, unusual adjustments, or items that sold while the record said none were in stock.

A common mix is a fixed schedule for coverage plus a short list of exceptions. For the value method, A items are the few that drive most of your sales or margin. A common starting point is to count A items monthly, B items quarterly and C items once or twice a year. The guides from Shopify and Poskio both use this split.

Whichever way you choose, decide it in advance, as Shopify’s guide advises, and keep the list so the next count starts where the last one stopped. Include the items your system shows at zero. If one of them is on the shelf, the record is wrong and replenishment cannot see the stock. Shopify’s guide calls these zero counts.

Then size the slice so it fits into normal days. Shopify’s guide, for example, spreads 2,400 products over eight weeks, which comes to about 300 a week. The same sum works for any scope:

Items to count per week = items in scope ÷ weeks to cover the scope

Pick a slice your team can finish in one short session, ideally before opening. If counts keep running over, make the slice smaller instead of rushing the count.

Set the cut-off before anyone counts

The cut-off is the point in time the count is compared with. Anything that changes the stock of the items in scope has to be finished, or written down, before it.

Stop the movement

Finish receiving, put-away, transfers and returns for the products in scope before the count starts. Shopify’s guide lists the usual causes of false differences: active receiving, shipping rushes, unresolved transfers, shift changes, and receipts, returns or adjustments that are still being recorded. Our retail goods receiving checklist and retail stock transfer process cover the two doors stock comes through.

The best time is before opening or in a quiet hour, when little is being sold, received or moved. Mark the area: put a tag on the shelf edge and park waiting deliveries away from it, so nobody adds stock to the bay or takes stock out of it during the count.

Store employee fixing a blank tag to a shelf bay before a cycle count while a delivery cage waits in the aisle

Write down the system quantity

The shift lead notes the system quantity for each item and the time, for example 8:30 AM, and keeps that copy. The counter’s sheet stays blank, because the count is blind.

Count only what is in stock and ready to sell. Lightspeed’s help tells you to leave out stock reserved for special orders, layaways and work orders. Goods in a hold area are tracked on their own case record, as in our retail damaged goods procedure, so keep them out of the shelf count.

Account for what moves anyway

Sales rarely stop for a count, and systems handle this differently. Lightspeed Retail (R-Series) records the system quantity when the first unit of an item is counted and marks items that changed afterwards with a QOH Changed button. Its help also says not to sell uncounted units of a partly counted item, and not to count units that arrive through purchase orders, transfers or refunds once the item has been counted.

If the store trades during the count, write the time on every line and compare each count with the system quantity at that time. Do not compare it with the quantity on screen when you finally enter the result.

Run the count: a counter, a checker and a blind sheet

Name the counter, the checker and the approver before the count starts. The counter counts. The checker recounts any difference beyond the tolerance. The approver reviews the cause and signs off the adjustment.

The counter works from a blind sheet: product, location and unit, with an empty box for the quantity and no system figure anywhere. Showing the expected number invites the counter to confirm it, a mistake that Poskio’s guide lists among the common ones.

Tidy the bay first: return misplaced products to their home and group identical products together. Then count every place the product can be, including the shelf, backstock, displays and the returns area, and write the unit on each line, because a single item, a pack and a case are different counts. Our retail stockroom organization guide shows how to give every storage position an address, which tells the counter where to look.

The counter does not fix a number on the spot or ask what the system says. A count that looks odd is written down as counted, and the checker deals with it later.

Recount and trace before you adjust

When the counts are in, the lead compares each line with the system quantity noted at the cut-off. Work out the difference the same way every time:

Variance = counted quantity − system quantity at the time of the count

A negative variance means the shelf holds fewer units than the record shows, and a positive one means more. Keep gains and losses apart: +3 on one item and −3 on another are two problems, not a clean zero.

Set the tolerance before the count, not after. Poskio’s guide notes that tighter tolerances for higher-value items are common, often an exact match for the items that matter most and a small unit margin for the cheap ones. Any line outside the tolerance goes to the checker, who counts it again without seeing the first count. Shopify’s guide gives the same order: recount discrepancies before adjusting records.

Two store employees recounting a shelf bay in a drugstore stockroom with a clipboard count sheet

If the recount confirms the difference, do not change the record yet. The cause often sits in another record, so check these in order:

  • Deliveries: received late, received twice, or posted against a similar product.

  • Transfers: sent or received, but not posted at one of the two stores.

  • Returns: customer returns put back on the shelf but not recorded, or recorded but not put back.

  • Holds and damage: goods in a hold area, or thrown away without a record.

  • Other places: backstock, displays, the returns area or another room.

  • Similar products: the wrong size, flavor or color counted, or sold under the wrong code.

  • Manual adjustments: recent changes by a user, visible in the item history.

For each confirmed difference, write down the item, the amount, the location, the date and the recent activity, such as receiving, transfers, returns or sales. For a large difference, add a photo of the bay to the sheet. If the cause is a missing delivery or transfer entry, fix that entry. If you find no cause, the difference goes to the approver as it is. For a difference that stays unexplained, our guide to running a retail inventory discrepancy investigation shows how to rebuild the timeline from the records and correct what caused it.

Adjust the record and close the count

One named approver reviews the sheet, the recount and the cause, then changes the quantity in the inventory system with a reason. Use the count reason, not a general correction. In Shopify, the Count reason is for the discrepancy found after counting physical stock, while Correction is the default general reason. Square lists Inventory recount among its predefined reasons.

Store manager signing a count sheet at a stockroom desk beside a counted shelf bay

Adjust only what the count proves. If the cause was a missing receiving or transfer entry, post that entry instead, because a count adjustment on top of it would hide the real fault. For how stock records are kept in the first place, see our retail inventory system guide.

Check before you reconcile. Lightspeed’s help warns that reconciling an inventory count cannot be undone, and other systems may work the same way. Look at the recounts and the approver’s name first, and check the adjustment history afterwards: every Shopify update is recorded there, so the entry can be traced later.

Close the count when three things are true: the system matches the approved result, the sheet is signed by the counter, the checker and the approver with the date, and the next count date for this area is set. File the sheet, or a photo of it, where the next shift lead can find it.

Example: one hair-care bay on a Monday morning

This is an invented example to show the sequence, not typical results. A drugstore counts one bay of hair-care products on Monday, Nov 2, 2026. The shift lead notes the system quantities at the 8:30 AM cut-off, and the counter finishes the bay by 9:20 AM. The tolerance is an exact match for items over $10 and one unit either way for cheaper ones.

System 14 · Counted 14
Argan shampoo, 250 ml ($7)
Match. No action.
System 9 · Counted 8
Daily conditioner, 250 ml ($6)
One unit short on a cheap item, inside the tolerance. The approver adjusts by −1 with the Count reason.
System 6 · Counted 3
Hair serum, 50 ml ($18)
Over $10, so an exact match was needed. The recount is also 3. The trace finds three units sent to another store on Saturday and never posted. The lead posts the transfer, and no count adjustment is made.
System 0 · Counted 4
Dry shampoo spray ($9)
A zero count: the system says none and the recount finds four on the shelf. The trace finds Friday’s delivery posted against a similar scent. The receiving entry is corrected and the other scent goes on the next count.
System 12 · Counted 9
Hand cream, 100 ml ($8)
Three short, outside the tolerance. The recount is also 9 and the trace finds nothing. The approver adjusts by −3 with the Count reason, and the item goes on next week’s list.

Of the five lines, one matched, two ended in a count adjustment and two were fixed in another record. Without the trace, all four differences would have been written off as count errors, and the transfer and receiving mistakes would have gone unnoticed.

Review the numbers every week

Once a week, spend ten minutes on the counts of the past seven days. Look at three things: how many lines matched, whether the same products or bays keep coming back, and whether the scheduled counts happened at all.

Count accuracy shows how often the record was right. Count a line as a match when it falls inside the tolerance you set before the count:

Count accuracy = (lines within tolerance ÷ lines counted) × 100

If 46 of 50 counted lines are within tolerance, count accuracy is 92%. Poskio’s guide measures inventory record accuracy the same way. Coverage is the share of scheduled counts that actually happened. It drops when counts are skipped on busy days, and then the accuracy figure rests on the easier counts. Our guide to retail inventory accuracy adds a second measure, how far off the misses are, and shows where records drift.

Repeats matter more than single misses. The same product, bay or delivery day showing up in consecutive counts points to a process problem, so change the cause: give the product one home, fix the label, or correct how deliveries are posted, as in our retail goods receiving checklist. Use the numbers to find stores and areas that need help, not to rank people.

Run the routine across many stores

The routine is simple to explain and hard to keep going in many stores at once. Counts get skipped on busy days, signed sheets stay in a drawer, and nobody above store level can see which differences are still open.

Bitreport task management can put each week’s slice on the shift lead’s list as a recurring task with a due date, ask for a photo of the signed sheet as proof, and give every recount or follow-up an owner and a deadline. A reviewer can mark up a photo and send it back with a note until it shows what is needed. On store visits, an audit item can check that last week’s sheet is filed, and a failed item creates a corrective task.

Bitreport is not an inventory system or a scanner-based count tool, and it does not post stock adjustments, so the count and the adjustment stay in your inventory system. To see how recurring tasks and photo proof look across locations, book a Bitreport demo.

Frequently asked questions

What is the difference between a cycle count and a full count?

A full count covers everything at once, usually a few times a year. A cycle count covers one planned slice at a time, so the rest of the store keeps trading.

How often should a store run a cycle count?

A common starting point is to count A items monthly, B items quarterly and C items once or twice a year. Pick a slice your team can finish in one short session, ideally before opening.

Can you run a cycle count while the store is open?

The best time is before opening or in a quiet hour, when little is being sold, received or moved. If the store trades during the count, write the time on every line and compare each count with the system quantity at that time.

Does a cycle count replace the yearly count?

A cycle count does not always replace the yearly count. Your accountant or local rules may still require a full count, so ask before you rely on cycle counts alone.

Take control of your operations, before chaos takes control of you.

This field is for validation purposes and should be left unchanged.