Retail & hospitality operations

Retail Inventory Accuracy: How to Measure It and What to Fix First

Marko Čačić Co-founder
Hardware store employee in a blue apron scanning a shelf label with a handheld barcode scanner next to bins of screws and bolts

In a 2008 study by Nicole DeHoratius and Ananth Raman of nearly 370,000 inventory records from 37 stores of one retail chain, 65% of the records did not match what was on the shelf. That is one retailer’s result, not an industry average, but it shows how easily records drift. Retail inventory accuracy is the share of your stock records that agree with what the store holds, and every reorder, refill and “in stock” answer starts from those records.

This guide shows how to measure it with two numbers, where records drift and what to change first. It is the overview for our guides to the routines behind it, starting with the retail cycle count process and the retail inventory discrepancy investigation.

Your retail inventory system keeps the quantities and posts the adjustments, and the store team keeps the routines around it. The guide belongs to our wider overview of retail operations.

What retail inventory accuracy measures

An inventory record is one line in your system: a product in one store, with the quantity the system says is on hand. The record is accurate when a count finds the same quantity. It is judged at one moment against one count, so a record can be right on Monday and wrong on Tuesday. Some guides call the result inventory record accuracy.

Software acts on the records. A record that shows four when the shelf holds none tells the system the product is still in stock, so a reorder or a refill may never be requested. A record that shows none when the shelf holds four can do the opposite and order stock the store does not need. The study’s authors make the same point: retailers rely on these records for the tools that forecast demand, plan assortment and replenish shelves, and wrong records weaken all three. Our guide to retail out-of-stocks shows how to tell whether an empty shelf comes from a wrong record, a late order or stock that never reached the shelf.

The study counted a record as inaccurate when the recorded quantity did not match the quantity found on the shelf. In the journal’s summary note, the authors estimate the potential lost revenue at more than 1% of the retailer’s North American retail sales and more than 3% of its gross profit. You will sometimes see 65% quoted as an accuracy rate. It is the other way round: 65% of the records in that study were inaccurate, and they came from one retailer. It is not an industry average.

Two findings help a store team more than the headline. Inaccuracy varied far more between product categories than between stores (26.4% of the variance against 2.7%), so look at categories, not only at store averages. Auditing practices lowered inaccuracy, while a complex store environment and the distribution structure raised it.

Pet store employee holding a tablet and pointing at a lower shelf of dog food bags while she checks the shelf against the stock record

The inventory accuracy formula: two numbers from one count

To calculate inventory accuracy, take two numbers from the same count. The first is how often the record is right. The second is how far off the wrong ones are. They answer different questions, and neither is enough alone.

Set the standard before the count. A line is one product in one place, with its system quantity and its counted quantity. Count every line on the list, including the lines the system shows at zero, because a product on the shelf with a zero record is a miss. Then decide whether a line must match exactly or may be off by a small tolerance, such as one unit on a cheap item. Poskio’s cycle counting guide notes that practitioners often demand an exact match on the highest-value items and allow a small margin on cheap ones, and that some organizations drop tolerances altogether. Whichever rule you pick, write it down and keep it, or this month’s result cannot be compared with last month’s.

The first number is the one our retail cycle count process guide calls count accuracy. Here it is record accuracy, because it shows how often the record was right:

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

The second number shows how far off the misses are. On each line, take the difference between the counted and the system quantity and drop the sign, so a shortage on one line cannot cancel an overage on another:

Variance rate = (sum of absolute differences ÷ sum of system quantities) × 100

Here is an invented example, not typical results. A store counts 40 lines against an exact-match rule, and 30 of them match, so record accuracy is 30 ÷ 40 × 100 = 75%. The ten misses are five shortages worth 20 units and five overages worth 10 units, so 30 units are off against 500 on the system and the variance rate is 30 ÷ 500 × 100 = 6%. One of the shortages is 12 units. Netted, 20 short and 10 over would leave only 10 units, or 2%, and hide two thirds of the problem.

Read the two together. Three lines in four were right, yet the 12-unit shortage alone is 40% of the units that are off. A high record accuracy with a high variance rate means a few large misses, so look for one event, such as a delivery posted twice, that hit several lines at once. A low record accuracy with a low variance rate means many small misses, which often points to a skipped step, such as returns that were never posted.

Where retail inventory records drift

A record changes only when someone posts something, and the shelf changes whether or not anyone posts it. Drift starts wherever the two can move apart: receiving, transfers, storage, damage and holds, sales and returns, and counts and adjustments. The table names the usual failure at each place and the guide that covers it.

Where stock changes How the record drifts Read next
Receiving A delivery is posted late, twice or in the wrong unit, or a short case is accepted as a full one. Retail goods receiving checklist
Transfers Stock leaves one store or arrives at another with only one end posted. Retail stock transfer process
Storage One product lives in several places, or in a place nobody checks, so counts miss units. Retail stockroom organization
Damage and holds Broken or expired goods leave the shelf with no record, or sit in a hold area and still count as stock. Retail damaged goods procedure
Sales and returns The wrong variant is scanned at the checkout, or a returned item goes back on the shelf without a posting. Retail inventory discrepancy investigation
Counts and adjustments A count is taken while stock moves, or an adjustment is posted with no reason, and the real cause stays hidden. Retail cycle count process

Several causes follow the checks in Poskio’s cycle counting guide.

The table shows where to look, and your own counts show which row matters most. Sort every confirmed difference into one cause from a short list: a wrong entry, a skipped step, a mixed-up product or unit, stock found elsewhere, a real loss, or unexplained. A few rounds of counts will show which row of the table causes most of your misses.

Clothing store employee at the service counter lifting a returned denim jacket out of a clear plastic return bag

What to change first

Counting more is rarely the first fix, because a count finds drift but does not stop it. Work in this order:

  1. Take a baseline. Count one aisle against the standard you set and work out both numbers. There is no standard rate to aim for, because the result depends on the tolerance you set, the product category and the time since the last count. Shopify’s cycle count guide says improvement over time matters more than hitting a fixed benchmark, so the first number only has to be an honest one.

  2. Sort every miss by cause. Use the short cause list above and write the cause on the count sheet. A cause you cannot name is a result too, so log it as unexplained.

  3. Fix the biggest source first. If most misses trace back to one door, such as deliveries posted late or twice, tighten that step before you add counts. More counts would only report the same drift more often.

  4. Count where it pays. Count fast-moving and high-value lines more often, include the lines the system shows at zero, and put any line that missed twice in a row on the next list.

  5. Keep every change traceable. Each adjustment gets a reason and a named approver. A missing or duplicate entry is corrected as that entry, not buried in a count adjustment.

A baseline does not need a full store count. One aisle is enough to get the first two numbers and to see which row of the table to look at first.

Compare stores on the same terms

Across several stores, the two numbers help only when every store measures them the same way. Check four things before you compare.

Use one standard for each product group. A store that allows a tolerance and a store that demands an exact match cannot be compared, so set the rule centrally, for example an exact match for fast sellers and one unit for cheap lines.

Compare on the same count interval. A store that counts a line every week will usually show a better result than one that counts it once a year, because less time has passed for drift. Datascan, an inventory counting provider, says that across the counts it runs for its clients, inventory accuracy percentages range from 20% to 97%, depending on the time between counts, the market segment and company culture.

Area manager and store manager comparing three printed count result sheets side by side on a staff room table

Show coverage next to both numbers. Coverage is the share of scheduled counts that actually happened, and a store that skipped half its counts has measured half its lines.

Cut the results by category as well as by store. In the study, far more of the variation lay between categories than between stores, so a good store average can hide one category that drives the problem.

Keep the routines running across many stores

Accuracy depends on routines that are easy to skip: the weekly count, the receiving check and the transfer sign-off. Bitreport task management can set each one as a recurring task for every store, with a due date and a required photo such as the signed count sheet or the delivery note. Managers can follow completion by location and export a report for the weekly review. On store visits, an item in an audit can check that the latest count sheet is filed, and a failed item can create a corrective task.

Bitreport is not an inventory system or a scanner-based count tool, and it does not calculate record accuracy or post stock adjustments, so the counts, both numbers and the adjustments 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 a good inventory accuracy rate?

There is no standard rate to aim for, because the result depends on the tolerance you set, the product category and the time since the last count. Shopify’s cycle count guide says improvement over time matters more than hitting a fixed benchmark, so the first number only has to be an honest one. A baseline does not need a full store count.

How do you calculate inventory accuracy?

To calculate inventory accuracy, take two numbers from the same count. The first is how often the record is right. The second is how far off the wrong ones are. Here is an invented example, not typical results. A store counts 40 lines against an exact-match rule, and 30 of them match, so record accuracy is 30 ÷ 40 × 100 = 75%. The ten misses are five shortages worth 20 units and five overages worth 10 units, so 30 units are off against 500 on the system and the variance rate is 30 ÷ 500 × 100 = 6%.

What causes inventory inaccuracy in retail stores?

A record changes only when someone posts something, and the shelf changes whether or not anyone posts it. Drift starts wherever the two can move apart: receiving, transfers, storage, damage and holds, sales and returns, and counts and adjustments. Sort every confirmed difference into one cause from a short list: a wrong entry, a skipped step, a mixed-up product or unit, stock found elsewhere, a real loss, or unexplained.

Is 65% the average inventory accuracy of retailers?

In a 2008 study by Nicole DeHoratius and Ananth Raman of nearly 370,000 inventory records from 37 stores of one retail chain, 65% of the records did not match what was on the shelf. You will sometimes see 65% quoted as an accuracy rate. It is the other way round: 65% of the records in that study were inaccurate, and they came from one retailer. It is not an industry average.

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

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