Inventory accuracy formula: how to measure record accuracy

Inventory record accuracy tells you how often your system’s stock quantity matches what is on the shelf. It is the number that decides whether you can trust every plan built on top of it. Here is how to calculate it properly.

The formulas

Inventory record accuracy (IRA) = counts within tolerance ÷ total counts

Within tolerance when: |counted − system| ÷ system ≤ tolerance

Dollar accuracy = 1 − (sum of |counted − system| × unit cost) ÷ (sum of system × unit cost)

Without a tolerance, a count must match exactly. That is right for expensive or critical items, and too strict for cheap items counted by weight or in bulk, which is why tolerances are usually set by ABC class.

A worked example

ItemClassToleranceSystemCountedOff byWithin tolerance?Variance at cost
A-100A0%4003961.0%No$8.00
A-200A0%1201200.0%Yes$0.00
B-100B2%50492.0%Yes$1.00
B-200B2%80756.3%No$20.00
C-100C5%100955.0%Yes$15.00
C-200C5%30300.0%Yes$0.00
MeasureCalculationResult
Record accuracy with tolerances4 ÷ 666.7%
Record accuracy, exact match2 ÷ 633.3%
Dollar accuracy1 − $44 ÷ $2,68598.4%
Net varianceevery difference was a shortfall−$44

The same counts give three very different pictures. Dollar accuracy looks excellent because the errors were on cheap items, while record accuracy shows that a third of the records were wrong. Report record accuracy as the main measure, and the dollar figures beside it.

In Excel

With system quantity in B, counted quantity in C and the tolerance as a decimal in D, starting in row 2:

Within tolerance (E2): =IF(B2=0,C2=0,ABS(C2-B2)/B2<=D2)

Record accuracy: =COUNTIF(E2:E7,TRUE)/COUNTA(E2:E7)

The IF handles items the system shows as zero: a count of zero is a match, and anything else is not.

What a good figure looks like

Many operations aim for 95% or better within tolerance. Below that, reorder points and shortage lists are working from stock that may not be there, which is how a shortage arrives with no warning.

Why records drift

  • Receipts put away before they are booked, or booked before they arrive.
  • Picks from the wrong bin or the wrong item.
  • Scrap, samples and damage that are never reported.
  • Units of measure mixed up, such as cases booked as units.
  • Returns put back on the shelf without a transaction.

Track accuracy by location and by class to find which of these is the problem. Read our guide to cycle counting for how to set up the counts, or use the free cycle count schedule calculator, which also works out record accuracy.

Measuring it in the Lab

The Cycle counts page in the Statadventure Optimization Lab applies a tolerance by ABC class, reports record accuracy by class and location, and lists the adjustments to post.

Measure record accuracy on sample data

The live demo needs no account. Use the sample company or load your own files, which stay in your browser tab.

Frequently asked questions

What is inventory record accuracy (IRA)?

The share of counted items whose system quantity matches the count, within an agreed tolerance.

Should IRA be measured in units or dollars?

Use the share of counts within tolerance as the main measure, and report dollar variance beside it. Dollar figures alone can hide many small errors.

What tolerance should we use?

A common start is 0% for A items, about 2% for B items and about 5% for C items. Tighten them as accuracy improves.

Try the Statadventure live tools

Inventory tools for US manufacturers: reorder point, safety stock, EOQ and landed cost, with the inputs and assumptions shown.

Open the live tools

Contact us

Scroll to Top