Skip to content

Inventory record accuracy: formula and how to improve it

Inventory Record Accuracy (IRA) shows how well ERP/WMS data matches actual stock in terms of quantity, locations, batches, and value. High accuracy is the foundation of reliable replenishment and OTIF/Fill Rate service.

What the metric measures

Discrepancies between records and physical stock at the SKU/location/batch level, and the quality of receiving, picking, transfer, production write-off, and returns processes.

Calculation

  • IRA, % = Accurate items / Checked items × 100
  • Quantity accuracy, % = Items with correct quantity / Checked × 100
  • Location accuracy, % = Items in the correct location / Checked × 100
  • Abs. deviation, % = Σ|Actual − Recorded| / Σ Recorded × 100
  • Adjustments, % of value = Σ|value of adjustments| / Average inventory value × 100
  • Cycle count coverage, % = Active lines counted / All active lines for the period × 100
  • Count effectiveness, % = Lines within tolerance / Counted × 100

Tolerances under the ABC approach: A ±1 unit or ±0.5%; B ±1%; C ±2–3%. For liquids/bulk goods, use weight/volume standards.

Calculation example

A sample count checked 500 items and in 470 the book quantity matched the physical count within tolerance. Accuracy = 470 / 500 × 100 = 94%. The variances on 30 items total 85,000, of which 60,000 come from ten expensive products.

Data sources

ERP/1C and WMS (stock levels, movements, adjustments), scanner logs/barcodes/RFID, ASN/receiving, shipments, production write-offs, transfer logs, cycle count reports, returns, and disposal.

Interpretation

  • IRA↓ and abs. deviation↑: receiving/picking errors, unrecorded transfers, mixed batches, write-offs "by standard rates," no scanning
  • Adjustments%↑ with stable IRA: inaccuracies in high-value items
  • Improvements: mandatory scanning, "one-touch" flows, cycle count by exception, movement freezes, packaging/labeling standards, RCA/CAPA

Useful breakdowns

  • ABC/XYZ; zones/racks/locations
  • Processes: receiving/picking/transfers/production
  • Batches/serial numbers/expiration dates; responsible shifts/crews
  • Suppliers; high-value/high-risk goods

Dashboard visualizations

  • Trend of IRA and absolute deviation
  • Heatmap of inaccuracies by zone/location
  • Pareto of SKUs and adjustment reasons
  • Cycle count coverage panel and count backlog
  • Chart of adjustments% of value and "no scan" incidents

How to set targets

IRA target: A ≥ 98–99%, B ≥ 97%, C ≥ 95%; 100% cycle count coverage per quarter (A: weekly/daily, B: monthly, C: quarterly); adjustments ≤ 0.2–0.5% of value. Fix tolerances, freeze rules, mandatory scanning, and SLAs for investigating and fixing root causes.

Common mistakes

  • Counting during active movements; no "freeze"
  • Mixed batches/UOM; duplicate stock-keeping units
  • A one-time "annual inventory" instead of regular cycle counts
  • Focusing on recounts without RCA
  • Different tolerances across warehouses; lumped adjustments without reasons

Questions for discussion

  • Where do 80% of deviations occur, and what are the top 3 causes?
  • What effect would 100% scanning at receiving/picking have?
  • Does the cycle count cover all class A items/high-risk zones?
  • Which CAPA will reduce adjustments by X% in 30 days?

Questions and answers

How do you calculate inventory record accuracy?

Accuracy, % = number of items where records matched the count (within tolerance) / number of items counted × 100. Also measure variances in money.

Why do records differ from physical stock?

Receiving and shipping errors, mixed-up items, unrecorded write-offs, theft and documents posted late.

How do you improve record accuracy?

Scan barcodes on every transaction, run regular cycle counts, especially for expensive items, and investigate the cause of every variance.