
Inventory accuracy is how closely your records match your shelves. The inventory accuracy formula is: (items counted that match the records ÷ total items counted) × 100. A score of 92% means 92 of every 100 items matched.
Our review in January 2026 covers 3 versions of the formula: by item, by location, and by dollar value. Each one has a worked example. You need a count sheet and a calculator.
Book a call
Accuracy is the share of counts that match your records, shown as a percent. A higher score is better. A score of 100% means every count matched the system.
The inventory accuracy formula is (items counted that match the records ÷ total items counted) × 100. The answer is always a percentage.
A record is a number in your system, such as a stock level. A match means the shelf count and the record agree. A miss means they differ by any amount.
The same math works for items, bin locations, and dollars. Pick one version and keep it.
Use one formula the same way every time, and your scores will mean something.
2 costs show up when counts are wrong: lost sales and trapped cash. Every sale, order, and buying choice starts from a stock number. A wrong number sends the wrong signal.
The law also cares. IRS Publication 538 says that to figure taxable income, "you must value your inventory at the beginning and end of each tax year" (IRS Publication 538). A bad count means a bad value.
A wrong count hurts sales, cash, and tax records at once.
Stockouts are times when you have none of an item that a customer wants. The system shows 10 units. The shelf holds 0. The order waits, and the customer waits too.
The FTC sets a duty to ship when you said you would for mail, internet, and phone orders (FTC order rule guide). Good counts help you keep that promise.
Accurate counts let you promise only what you can ship.
If you would rather not compare products, describe how your operation already works and we build the system around it.
No build cost. You see it running on your own process first, and the monthly subscription starts only once it is live.
Book a callOverbuying is buying stock you already own. It happens when records show too little stock. You reorder items that sit in the back room. Cash sits on shelves instead of in the bank.
GS1 builds the barcode standards behind scan counts and says they help firms "identify, capture and share" product data (GS1 barcode standards). Clean data starts with a count you trust.
A trusted count keeps you from buying what you already have.
4 steps turn counts into a score. To calculate inventory accuracy, divide the items that match by the items you counted, then multiply by 100. A fifth-grade class can do it with a pencil.
A simple count sheet has 5 columns:
The score is only as honest as the count behind it, so count each item with fresh eyes.

184 of 200 SKUs match the records in this example. A SKU is a stock keeping unit, one code for one product. A warehouse counts 200 SKUs, and the math is short.
Those 16 SKUs are your to-do list. Pull them out and look at each one. Check receipts, picks, and returns for the past 30 days.
Next, sort the 16 misses by cause. Say 9 trace to receiving, 4 to picking, and 3 to damage. Now you know where to act first.
A 92% score sounds fine until you see it as 16 wrong items out of 200. Every miss is a place where someone can ship the wrong thing.
Treat each miss as a clue, and the score becomes a repair plan.


Location accuracy is the share of bins that hold the right item in the right amount. The formula is (locations with correct items and counts ÷ total locations checked) × 100.
Say a picker checks 120 bin locations. 108 hold the correct item and count. The math is 108 ÷ 120 = 0.90, so the score is 90%.
This version helps pickers most. A picker walks to a bin and finds it empty or full of the wrong item. Each trip like that burns minutes. A warehouse with a high location score has shorter walks and fewer calls to the office.
Location accuracy tells you where stock is. Item accuracy tells you how much exists.
Track both scores, and you will know where stock sits and how much of it you hold.
Describe how the work runs today. We map it on a call and show you what it would look like built around that, before you spend anything.
Book a call
Dollar accuracy is the share of recorded value that holds true. It measures how much money sits in the gap, not how many items miss. The formula is 1 minus (absolute dollar variance ÷ total recorded inventory value), then × 100.
Absolute variance means you add up every miss and ignore the plus or minus sign. Say you are over by $3,000 on some items and short by $2,000 on others. The absolute variance is $5,000.
Now work the math. Recorded stock is $250,000. The variance is $5,000.
Inventory variance is the gap between recorded and counted stock, in units or dollars.
A dollar score shows how much money the errors put at risk.
1 missing $4,000 machine part hurts more than 400 missing $1 screws. A count score treats both as one miss. A dollar score shows the real hit.
The method changes the answer. Supply Chain News gives a case where "the traditional calculation would result in an accuracy level of 99.8%, using the unit volume based approach the accuracy level would be 99.93%" (Supply Chain News). Same stock, two answers.
Run both scores and compare them before you pick one.
Tolerance is a small gap you agree to accept before you call a count a miss. A common one is plus or minus 1 unit on low-cost items. You decide it before the count starts.
Use a tight tolerance where errors cost more:
Here is the effect. A bin of 500 screws shows 499. With a 1 unit tolerance, that bin counts as a match. A $900 part with the same gap counts as a miss.
Write the rule down and post it. Every counter must score the same way, or last month's 94% and this month's 96% mean nothing.
One written tolerance rule keeps every counter on the same scale.

95% is a commonly cited goal for the inventory accuracy rate. Many operations aim for 98% to 99%. Treat these as general benchmarks, not rules.
Open Sky Group offers a warning level: "Anything below 94% usually points to replenishment or inventory accuracy problems upstream, not shipping problems" (Open Sky Group). A score under 94% needs a look at receiving and stocking first.
Set a target for each item class. Here is one split to start with:
Readers who ask for a good inventory accuracy percentage want one number. The honest answer is a range. The trend over time matters more than one score, so a climb from 88% to 94% beats a lucky 96%.
Aim for a steady climb, not a single high score.

Off-the-shelf means fitting your process to the software. We do it the other way round, and the first look costs nothing.
Book a call
Cycle counting is the habit of counting a small group of items on a steady schedule instead of the whole building at once.
A yearly shutdown count stops work for days. A cycle count takes an hour or two. A clerk counts 30 items on Monday and a different 30 on Tuesday. The work spreads out.
Here is a plan for 600 items. At 30 items a day, the full set is counted in 20 working days. Every item gets a fresh check each month.
Each cycle count yields fresh data for the formula. Compare each stock count to the record, mark match or miss, and log the score. Over a month the scores show a pattern.
Small, frequent counts also catch errors while the cause is still fresh in someone's memory.

Frequency is set by how fast stock moves and what it costs. Most small operations should measure inventory accuracy weekly or monthly.
Use the ABC approach. ABC means you sort items by value and sales, then count the top group more frequently.
Keep the same method each time. If you count by item in March and by dollars in April, the two scores cannot be compared. Fair trends need one method.
Count the costly items most, and use one method every time.

4 causes explain most low scores. When your inventory count does not match your records, one of these is almost always the answer.
Most misses trace back to a human step, so start there.
Manual tracking is a chain of hand-offs, and each one risks a wrong number. Printed sheets, spreadsheet files, and email updates drift out of sync with your accounting software's inventory records.
One person updates the sheet. Another updates the books a day later. A third reads an old email. Now 3 versions of one number exist.
Retyping the same number in several places invites typos. The BLS tracks pay for shipping, receiving, and inventory clerks (BLS occupational data). Say 3 clerks each spend 5 hours a week fixing mismatches at $20 an hour. That is 3 × 5 × $20 × 52 = $15,600 a year.
Add the lost sales from each stockout, and the true cost climbs higher. Time spent on fixes is time not spent on counts.
Fewer hand-offs mean fewer errors.

3 checks find the cause of most mismatches, in order. Start with the cheapest one and work down.
If 8 of 10 misses come from one supplier, the problem sits at the receiving dock. If they cluster on one shift, the problem is training or timing.
Write the cause next to each miss on the count sheet. After a few weeks, the same names repeat, and the fix becomes clear.
A pattern points to a cause, and a cause is what you can fix.
No build cost. The subscription starts once it is live and doing the job, not before.
Book a call4 small habits raise inventory accuracy without a big system change. They cost little and start this week.
The NIST Manufacturing Extension Partnership offers vendor-neutral guidance on supply chain process (NIST MEP). Start with one habit this week and add another next month.
Small habits, kept daily, beat big projects that stall.

Accuracy is the base for 3 other measures. Inventory accuracy asks one question: do my records match the shelf? The others ask different questions.
Open Sky Group puts the link plainly: "Anything below 94% usually points to replenishment or inventory accuracy problems upstream" (Open Sky Group).
The Census Bureau tracks the inventories-to-sales ratio for wholesale firms (Census wholesale trade). The average days to sell inventory formula and the total cost of inventory formula both use your recorded stock. A wrong count bends every one of them.
Fix the count first, and the other measures get more reliable.
Shrinkage is the loss of stock value between what records say and what you hold. The formula is (recorded value minus actual value) ÷ recorded value × 100.
Say your books show $250,000 of stock. A count finds $245,000. The gap is $5,000. Then $5,000 ÷ $250,000 = 0.02, so shrinkage is 2%.
Shrinkage is the cost side of the same gap. Accuracy shows how often you miss. Shrinkage shows how much the miss cost you in dollars.
When all your variance is loss, shrinkage equals 100 minus your dollar accuracy. Here that is 100 minus 98, or 2%.
If some misses are overages, the two numbers will differ. Net the plus and minus for shrinkage, but ignore the signs for accuracy.
Shrinkage and accuracy tell one story from two sides.


4 mistakes ruin an accuracy score fast. Each one makes the number look better than the truth.
A score is a tool for finding problems, so do not tune it to look good.
1 afternoon is enough for your first accuracy check. Follow these steps in order.
A written score with a date is worth more than a guess in your head.
Yes, 95% is a common minimum goal. Many teams push to 98% or 99% on costly or fast-moving items. Set a separate target for each item class instead of one number for all stock.
Yes. Subtract your absolute dollar variance divided by total recorded value from 1, then multiply by 100. A $5,000 gap on $250,000 of stock scores 98%. Count scores and dollar scores answer different questions.
Start with 50 to 200 items per round. Pick across zones, value classes, and fast and slow movers. A sample that covers the whole building gives a fairer score than a large sample from one aisle.
No. Your accounting software holds your stock records. It cannot see the shelf. You must count items by hand or by scan and compare each count against that record.
Divide the locations holding the correct item and count by the total locations checked, then multiply by 100. If 108 of 120 bins are right, your location score is 90%.
Absolute dollar variance is the total of every count miss with the plus and minus signs removed. If you are over by $3,000 and short by $2,000, the absolute variance is $5,000.
Run one weekly or monthly. Count your highest-value items most often and cheap items once or twice a year. Keep the same method each round so scores compare fairly.
A 30 minute call, your operation mapped, and a clear picture of what we would build. No obligation and nothing to install.
Book a callMore on the same work.