Cycle counting for a small shop: a practical schedule and process (2026)
How to set up cycle counts in a small shop: which items to count how often, a worked weekly schedule, how to run a blind count, and how to measure inventory accuracy.
Key Takeaways
- Count by value, not alphabetically. Rank items by the cost of stock that moves through them each year. Count the top group most often; that is where errors cost the most money.
- A 400-item shop needs about 45 item counts a week. With every item counted at least quarterly and the top items monthly, the worked schedule below needs roughly 9 item counts a day, five days a week.
- Count blind. The person counting should not see the expected quantity until they have entered what they found. Otherwise counts drift towards confirming the system.
- Investigate before you adjust. A difference is often an unrecorded delivery, a sale rung up under the wrong variant, or stock in a second location. Fix the cause, then post the adjustment with a reason.
- Track two accuracy numbers. The share of items that matched, and the absolute value of differences as a share of the value counted. Both should improve month on month.
Quick-Start Prioritization Framework
| Shop size | Suggested approach | Rough weekly effort |
|---|---|---|
| Under 100 items | Count one shelf or category a day; every item monthly | 15 to 30 minutes |
| 100 to 500 items | ABC schedule: A monthly, B and C quarterly | 30 to 60 minutes |
| 500 to 2,000 items | ABC schedule plus counts triggered by zero stock or negative stock | 1 to 3 hours |
| Any size, before year end | Full count only if your accountant needs one, or cycle counts that cover every item | Depends on coverage |
Cycle counting means checking a small slice of your stock on a regular schedule instead of closing for a full stocktake once a year. Done well, it takes minutes a day, catches errors while they are still fresh enough to explain, and keeps the numbers your reorder points depend on trustworthy. This guide shows how to decide what to count and how often, builds a real schedule for a 400-item shop, and covers the counting process, the accuracy maths and what to do with the differences you find.
Cycle counts vs a full stocktake
A full count, or stocktake, checks every item in a location at once. A cycle count checks a subset at regular intervals. Square’s help centre describes the two this way, and adds a detail worth knowing: in a Square full count, any stock you don’t count is set to zero when the count is approved. That is the right behaviour for a full count and a nasty surprise if you meant to count one shelf.
Why small shops prefer cycle counts
You don’t close the shop. Errors are found weeks after they happen rather than months, so someone still remembers the delivery that came in short. And counting a few items daily becomes a habit rather than a dreaded event.
When a full count still makes sense
When you are switching systems and need a clean starting number for everything. When accuracy has slipped so far that cycle counts can’t keep up. And when your accountant needs a year-end count; many accept a well-documented cycle count programme instead, so ask them first.
How often is enough
Square’s help centre recommends counting each item variation at least once every 90 days. That is a sensible floor for a small shop. Your best-sellers and most expensive items deserve more.
Decide what to count with ABC analysis
ABC analysis sorts items by how much money flows through them, so your counting time goes where mistakes are expensive.
Step 1: calculate annual cost value per item
For each item, multiply units sold in the last 12 months by unit cost. A candle that sold 2,190 units at $4.50 cost has an annual cost value of $9,855. A greeting card that sold 600 units at $1.20 has $720.
Step 2: sort and split
Sort from highest to lowest. A common split: the top 20% of items are A, the next 30% are B, the bottom 50% are C. In most shops the A items account for the large majority of annual cost value, but check your own numbers rather than assuming.
Step 3: move items up a class when they deserve it
Some cheap items deserve A treatment anyway: small, easily pocketed products; items customers come in for specifically; anything that keeps going negative in your system. Promote them by hand.
A worked schedule for a 400-item shop
Here is a full schedule for a shop with 400 stocked variants, using the 20/30/50 split and Square’s 90-day floor.
| Class | Items | Count frequency | Counts per year |
|---|---|---|---|
| A | 80 | Monthly (12 a year) | 960 |
| B | 120 | Quarterly (4 a year) | 480 |
| C | 200 | Quarterly (4 a year) | 800 |
| Total | 400 | 2,240 |
Turning it into a daily list
2,240 counts ÷ 52 weeks ≈ 43 counts a week, or about 9 a day over five days. If counting one item takes around a minute including finding it, that is roughly 10 minutes a day. Your timing will differ; time the first week and adjust.
Group counts by location
Counting nine items scattered across the shop is slower than counting nine on one shelf. Build each day’s list so the A items due that day are joined by B and C items stored nearby.
Add triggered counts
On top of the schedule, count an item whenever the system shows zero or a negative number, when a customer can’t find something the system says you have, and when a delivery is received with a discrepancy. These counts catch errors at the moment they become visible.
How to run a cycle count properly
The counting itself is where accuracy is won or lost. Agree a short routine and stick to it.
Pick a quiet time and freeze movements
Count before opening or in a lull. While you count an item, don’t sell or receive it. If a customer buys one mid-count, note it and subtract it before entering the total.
Count every place the item lives
Shop floor, back room, display, window, the box under the counter. Missing a second location is the most common cause of a “shortage” that isn’t real.
Count blind, then recount differences
The counter enters what they found before seeing the expected quantity. If the result differs from the system by more than your tolerance, a second person recounts. Only then is the number accepted.
Check cutoff
Look for deliveries that arrived but haven’t been entered, sales that are parked or on hold, and returns waiting to be processed. Each one makes the count look wrong when it isn’t.
Measuring inventory accuracy
Two numbers tell you whether your counts are getting better, and they are easy to calculate from each week’s results.
Item accuracy
Item accuracy = items whose count matched the system (within tolerance) ÷ items counted. If you counted 43 items this week and 36 matched, item accuracy is 36 ÷ 43 ≈ 84%.
Value accuracy
Add up the absolute value of every difference at cost, ignoring whether it was over or short, and divide by the total cost value of what you counted. If the week’s counted stock was worth $8,400 and differences added up to $312, absolute value variance is 312 ÷ 8,400 ≈ 3.7%.
Track the net figure (overs minus shorts) separately. A small net with a large absolute variance means errors are cancelling out, often a sign of sales rung up under the wrong variant.
Set tolerances by class
Exact match for A items. For cheap C items sold in bulk, such as loose stickers, a tolerance of a couple of units or 2% is reasonable. Write the tolerances down so everyone counts against the same rule.
What to do with the differences
A count difference is information. Adjust the number, but first find out why it was wrong, or you will be correcting the same item next month.
| What you see | Likely cause | What to check |
|---|---|---|
| One variant over, its sibling short by the same amount | Sold under the wrong size or colour | Recent sales of both variants, staff training at the till |
| Short by a full case or pack | Delivery entered but not received, or received short | Packing slip vs purchase order vs what was entered |
| Short by a few units repeatedly | Theft, unrecorded damage or samples | Damage log, item placement, who has access |
| Over | Delivery not entered, return not processed, a second location missed | Unprocessed paperwork, back room |
| System shows negative stock | Sales recorded before the delivery was entered | Receiving routine and timing |
Post adjustments with a reason
Use a short, fixed list of reasons: count correction, damaged, theft suspected, supplier short, wrong variant sold. A fixed list lets you total adjustments by reason at the end of the month.
Look for patterns monthly
Once a month, total adjustments by category, reason and the person who received or sold the stock. One category drifting every month tells you more than any single count.
Tools: paper, spreadsheet or software
Cycle counting works with a clipboard. Software makes it faster and keeps the history.
Spreadsheet
Keep a tab with every item, its class, last count date and next due date. Filter by due date each morning to get the list. Record counts in a separate tab so you keep the history rather than overwriting the stock number.
Inventory and POS software
Square documents an inventory counting tool for full or cycle counts on its Plus and Premium plans. Sortly lists stock counts that you can create, assign and track. Many tools let you count by scanning, which reduces miscounts on similar-looking items.
Your own app
If you’d rather not do this yourself in a spreadsheet and no tool matches your process, an AI app builder such as Floot can build a small counting app with your ABC classes, a daily list, blind entry and an adjustments report, which you then test with a week of real counts.
Common mistakes
Most cycle counting programmes that fail do so for the same handful of reasons.
- Showing counters the expected quantity first, so they confirm the number instead of counting.
- Adjusting the number without looking for the cause, so the same error repeats.
- Counting only the shop floor and forgetting the back room or display stock.
- Running a “full count” for one category in software that zeroes every item not counted.
- Counting while deliveries are half-entered or sales are parked.
- Letting the schedule slip for a busy month and never catching up. A shorter list done daily beats a long list done occasionally.
- Measuring only item accuracy, which hides large value errors on expensive items.
Frequently Asked Questions
What is cycle counting in inventory?
Cycle counting is counting a small portion of your stock on a regular schedule, so every item is checked several times a year without closing for a full stocktake.
How often should a small shop do cycle counts?
Count something most days. Each item should be counted at least every 90 days, which is Square’s recommendation, with best-sellers and expensive items counted monthly or more often.
What is ABC cycle counting?
Ranking items by annual cost value (units sold × unit cost), splitting them into A, B and C groups, and counting the A group most often. It puts counting time where errors cost the most.
What is a blind count?
A count where the person counting does not see the expected quantity until after they enter what they found. It prevents people from simply confirming the system’s number.
Can cycle counts replace a year-end physical inventory?
Often, if every item is counted during the year and you keep records of each count and adjustment. Ask your accountant what they need before dropping the year-end count.
What is a good inventory accuracy rate?
There is no universal benchmark for small shops. Set your own baseline in the first month, aim for exact matches on A items, and watch the trend. Rising accuracy matters more than any single figure.
Sources
- Square Support, Conduct, review, and approve inventory counts, 2026 https://squareup.com/help/us/en/article/8249-conduct-full-inventory-counts-with-square-for-retail
- Square, Square for Retail pricing and features, 2026 https://squareup.com/us/en/point-of-sale/retail/pricing
- Sortly, Pricing and plan comparison, 2026 https://www.sortly.com/pricing/