Fixing Inventory Discrepancies: KPIs, SOPs, and Controls for Accurate Stock
An inventory discrepancy is the gap between what your system says you have and what is physically on hand. Inventory discrepancy control is critical for service levels, cash flow, and accurate financials.
What Is an Inventory Discrepancy?
Definition and how it differs from shrinkage
An inventory discrepancy is any variance between recorded stock and actual count at the SKU-location level. Discrepancies can be positive (overstock vs system) or negative (understock vs system).
Shrinkage is the subset of discrepancies caused by loss, theft, damage, spoilage, or unrecorded consumption. Not all discrepancies are shrinkage; many are timing, data, or procedural issues.
Overstock vs understock variance examples
Overstock: System shows 90 units, physical finds 100. Likely double-receipt, unit-of-measure error, or unrecorded return.
Understock: System shows 100 units, physical finds 90. Possible picking error, theft, damage, or mis-located inventory.
Location variance: Total warehouse quantity matches, but the bin is wrong, driving pick failures.
Where discrepancies surface in systems and reports
Cycle count variance reports by SKU/bin
Negative inventory and backorder exceptions
Pick/ship short reports and order fill rate drops
Perpetual vs GL valuation tie-out differences
Aged open RMA or ASN mismatches
Work order variances and BOM usage deltas
Why Discrepancies Matter
Customer and service-level impact
Stockouts and late shipments reduce fill rate, OTIF, and CSAT.
Incorrect availability causes order cancellations and channel penalties.
Higher safety stock to mask inaccuracy inflates costs.
Financial and cash flow impact
Overstated inventory inflates assets and misstates COGS.
Shrinkage raises COGS and compresses gross margin.
Cash is trapped in overbuys driven by faulty signals.
Operational ripple effects across purchasing and production
MRP/MPS plans over- or under-order.
Line stoppages from missing components.
Labor wasted on re-picks, recounts, and expedites.
Common Root Causes
Counting errors and procedural gaps
No double-counts or witness controls
Counting with movement allowed (no freeze)
Inconsistent units (each vs case vs pallet)
Rushed counts during peak windows
Receiving and shipping mistakes
Missed lines vs ASN, short receipts, or over-receipts
Cross-docking without system confirmation
Mislabeling inbound pallets or mixed-SKU cartons
Shipments closed in system before physical load-out completion
Data entry and master data issues (UOM, SKUs, barcodes)
Incorrect UOM conversions and pack sizes
Duplicate or retired SKUs still active
Barcode mismatch to SKU or lot
Inconsistent standards between ERP, WMS, and 3PL
Damage, spoilage, and uncontrolled returns
Breakage not written off promptly
Expired product still in pickable bins
Returns restocked without inspection or wrong condition code
Theft, fraud, and access control failures
High-value SKUs without cage or camera coverage
Segregation of duties gaps in adjustments
Collusion during receiving or shipping
Supplier errors and documentation mismatches
Incorrect pack counts or short packs
ASN quantities not matching BOL or actual
Substitutions without prior approval
How to Detect and Measure Discrepancies
Key KPIs and formulas (accuracy, shrinkage, variance)
Inventory accuracy = 1 − (sum absolute variance units ÷ sum book units). Example: 300 variance units on 20,000 book units = 98.5%.
Location accuracy = correct-location SKUs ÷ total SKUs checked.
Shrinkage rate = inventory write-offs ÷ average inventory value. Track by period.
Count hit rate = counts within tolerance ÷ total counts.
Order fill rate = shipped in full on first attempt ÷ total orders.
GL tie-out delta = perpetual value − GL inventory account. Investigate if outside tolerance.
Common targets:
Inventory accuracy: 97–99.5% depending on complexity.
Shrinkage: 0.2–0.5% for distribution/light manufacturing; 0.8–2.0% in retail categories with higher risk.
Cycle counting strategies (ABC and risk-based)
ABC by velocity/value: A (top 20% value/throughput) count weekly; B (next 30%) monthly; C (remaining 50%) quarterly.
Risk-based overlays: Increase frequency for pilferable, high-dollar, high-defect, and returns-heavy SKUs.
Sampling size rule of thumb: Count enough SKUs per cycle to cover 4–6 full passes per year on A items.
Freeze locations during counts. Use blind counts and recounts when outside tolerance.
Exception reporting and variance codes
Use standardized variance codes to drive actions:
RC: receiving count mismatch
SH: shipping short/over
CC: cycle count error
DM: damage/spoilage
MS: mislocation/bin error
MD: master data/UOM issue
TH: theft/loss suspected
RT: returns/restock error
Each code triggers a specific workflow, documentation, and owner.
Audit trails, logs, and video/time stamps
Tie each movement to a user, timestamp, device ID, and source document.
Capture scan logs for receive, move, pick, pack, and ship.
Link CCTV timestamps at docks and high-value areas to transaction times for evidence review.
Sampling methods and tolerance setting
Tolerances by item class and value. Examples:
A items: units ±0.5% or value ±$50, whichever is tighter.
B items: ±1.0% or ±$100.
C items: ±2.0% or ±$200.
Monetary approvals: require supervisor sign-off above thresholds and finance sign-off for high-value adjustments.
For serialized items, tolerance is zero variance.
Rapid Response: Reconcile Today
Triage checklist and safety considerations
Stop movement in affected bins/locations.
Verify safety: secure forklifts, ladders, and traffic lanes.
Confirm count team and witness are trained and independent.
Pull last 7–30 days of transactions for the SKU/location.
Physical recount procedure and witness controls
Blind count by counter A; record units and condition.
If outside tolerance, recount by counter B with supervisor.
Validate packaging and UOM conversions during recount.
If still outside tolerance, expand to adjacent bins and staging.
System adjustments, approvals, and documentation
Post inventory adjustment with variance code, notes, and evidence links.
Route for approvals per threshold. Require finance approval for write-offs over set limits.
If GL is integrated, ensure automatic journal or post manual JE with reference numbers.
Update KPIs and variance dashboard same day.
Quarantine suspect items and locations
Move damaged, expired, or unknown-condition goods to quarantine.
Disable picking from suspect bins until resolved.
Place holds on related open orders if risk of short ship is high.
Post-mortem notes and immediate containment actions
Record immediate root cause hypothesis and quick countermeasures.
Schedule targeted counts for related SKUs.
Notify purchasing or supplier if inbound is implicated.
Tighten access or add camera coverage if loss is suspected.
More time, More Sales
Root Cause Investigation Playbook
Diagnostic decision tree
If variance is positive and recent return activity exists, check RMA restock and condition codes.
If variance is negative and pick density surged, analyze short picks, substitutions, and picker error rates.
If location accuracy fails but total on-hand matches, investigate putaway and move confirmations.
If only one lot/serial range is impacted, verify lot merges/splits and expiration pulls.
If variance correlates with a supplier or ASN, audit receiving counts and pack sizes.
If variance spikes on weekends or shifts, review access logs and supervision coverage.
If UOM conversions vary by channel or system, validate master data sync and barcode mappings.
Process mapping of the transaction path
Map from PO/WO creation to receipt, putaway, moves, pick/pack/ship, and adjustments.
Identify where confirmations, scans, and approvals should occur.
Compare expected vs actual timestamps and handlers.
Evidence collection (receiving docs, scans, CCTV)
PO, ASN, packing list, and BOL
Receipt logs, scale/weight slips, and pallet labels
Pick tickets, wave logs, and carrier manifest
Scan logs per user/device and bin history
CCTV snapshots aligned to event times
5 Whys and fishbone analysis
People: training, fatigue, segregation of duties
Process: SOP clarity, approvals, handoffs
Technology: scanning, latency, integration failures
Materials: packaging, labeling, UOM
Environment: congestion, layout, lighting
Ask “why” iteratively until a controllable cause is identified.
Corrective and preventive actions (CAPA)
Corrective: adjust stock, rework labels, reissue pick waves, supplier chargebacks.
Preventive: revise SOPs, add mandatory scans, adjust layout, implement tolerance workflows, strengthen master data governance.
Prevention and Controls
Receiving controls and supplier scorecards
Count verification by risk: 100% for A items or new suppliers; sampling for proven vendors.
Weight/measure checks for standard packs.
Require ASNs with SSCC labels; reconcile variances before putaway.
Scorecard vendors on receipt accuracy, documentation quality, and damage rate; escalate and apply corrective actions.
Barcoding, RFID, and mobile scanning standards
Use GS1 standards for GTIN, SSCC, lot/serial, and dates.
Enforce scan-to-confirm for receive, putaway, move, pick, and ship.
For high-shrink SKUs or fast-moving pallets, consider RFID portals at dock doors.
Validate labels at receiving; quarantine noncompliant barcodes.
Standard work, training, and segregation of duties
SOPs with photos and step-by-step scans for each process.
Train and certify counters; refresh quarterly and after errors.
Separate receiving, adjustments, and approvals; rotate counters to reduce bias.
Physical security and loss prevention
Cage high-value SKUs; limit keys and log access.
CCTV coverage for docks, returns, and high-loss aisles.
Conduct random bag checks within policy and local law.
Seal trailers; record seal numbers in outbound documentation.
Returns/RMA and reverse logistics controls
Inspect, grade, and disposition returns before restock.
Separate salvage, refurb, and resaleable stock.
Link RMA to original order and lot/serial; adjust only after inspection.
Tolerance thresholds and variance approval workflow
Define per-class tolerances and approvers.
Auto-approve variances within tight limits; route exceptions to supervisors and finance.
Require root cause notes for all adjustments; block “misc” codes.
Advanced Tactics and Technology
Serial/lot tracking and expiration management
Enforce lot/serial capture at receive, move, and ship.
FEFO for dated goods; auto-exclude expired lots from picks.
Perform targeted cycle counts by lot nearing expiration.
Kitting, BOM accuracy, and production reporting
Validate BOM quantities and scrap factors quarterly.
Backflush only with reliable scanning and yield tracking.
Count kit components and finished goods post-assembly; reconcile variances before releasing.
Multi-location, 3PL, and omnichannel alignment
Standardize SKU, UOM, and barcode conventions across all facilities and partners.
Require 3PLs to share scan logs, variance codes, and timestamps.
Align omnichannel ATP rules with inventory accuracy; buffer risky SKUs.
Integration checks and master data governance
Run daily reconciliation: PO receipts, shipments, and adjustments across ERP/WMS/3PL.
Use master data change controls with maker-checker approval.
Monitor failed messages and retry queues; alert IT on spikes.
AI and anomaly detection alerts
Flag sudden usage spikes, negative on-hand, repeated adjustments by user, or lot gaps.
Rank SKUs by risk and recommend targeted cycle counts.
Visual recognition or weight-sensing can validate counts on fast lanes.
Role-Based Responsibilities
Warehouse operations and cycle count owners
Execute cycle counts, recounts, and bin audits.
Enforce scan-confirm and quarantine processes.
Maintain variance logs and immediate containment actions.
Purchasing and supplier management
Enforce ASN and labeling standards.
Review supplier accuracy scorecards; issue corrective actions.
Align order multiples and pack sizes to reduce UOM errors.
Finance and accounting tie-out to the GL
Reconcile perpetual to GL monthly; explain deltas by code.
Post adjustment journals with references; review write-offs vs thresholds.
Validate valuation method (FIFO/LIFO/weighted) aligns with operational dating and lot moves.
IT and systems ownership
Maintain integrations, device management, and user access.
Govern master data, barcode standards, and audit logging.
Provide dashboards and automated alerts to operations and finance.
Management dashboards and cadenced reviews
Review KPIs weekly: accuracy, shrinkage, count hit rate, and open variances.
Track corrective actions to closure and verify effectiveness.
Sponsor projects for layout, technology, and SOP improvements.
30-60-90 Day Plan
Quick wins in the first 30 days
Freeze bins during counts; add blind recounts.
Implement variance codes and approval thresholds.
Quarantine damaged/expired returns; stop restock without inspection.
Run ABC classification and start weekly A-item cycle counts.
Lock down access to adjustments and high-value cages.
Stabilize and standardize by day 60
Publish SOPs for receive, putaway, move, pick, ship, and returns.
Standardize barcode and UOM; fix top 20 master data offenders.
Launch supplier scorecards and receipt verification tiers.
Add exception dashboards for negative on-hand, repeated variances, and GL tie-out gaps.
Optimize and automate by day 90
Expand risk-based cycle counts and location audits.
Pilot RFID or vision for high-risk lanes; quantify ROI.
Integrate CCTV timestamps with WMS events for investigations.
Implement anomaly alerts and role-based dashboards.
Conduct a CAPA review and reset targets for next quarter.
FAQs
What is the difference between an inventory discrepancy and shrinkage?
An inventory discrepancy is any difference between system and physical counts. Shrinkage is the portion caused by loss, theft, damage, spoilage, or unrecorded consumption. Many discrepancies stem from process or data errors rather than shrinkage.
What is an acceptable inventory variance percentage?
Targets vary by industry and complexity. Common goals are 97–99.5% inventory accuracy and shrinkage below 1%. High-risk retail categories may see higher shrinkage; controlled distribution often targets 0.2–0.5%. Set tighter tolerances for A items.
How often should we run cycle counts versus a full physical inventory?
Run continuous cycle counts: A items weekly, B monthly, C quarterly, with risk-based increases for pilferable or problem SKUs. Conduct a full physical annually if required by audit or policy, but robust cycle counting can reduce its scope and disruption.
How do you reconcile an inventory discrepancy in the GL and financial statements?
Post inventory adjustments in the WMS/ERP with variance codes and documentation. Ensure the GL reflects these via integrated postings or manual journal entries referencing adjustment IDs. Reconcile perpetual to GL monthly and explain differences by cause and approval.
What causes negative inventory and how do you fix it?
Common causes are timing gaps (ship before receive), backflushing without receipts, data errors, or missed confirmations. Fix by receiving and dating transactions correctly, enforcing scan-confirm, blocking picks below zero, and running exception alerts for negatives.
Which is better for reducing discrepancies: barcodes or RFID?
Barcodes with disciplined scan-confirm are cost-effective and sufficient for many operations. RFID adds value for high-volume, high-shrink, or fast-moving pallet flows where hands-free reads reduce misses. Pilot, measure read rates and exceptions, then scale based on ROI.
How should discrepancies be handled with a 3PL or dropship partner?
Use shared variance codes, ASN and labeling standards, and daily reconciliation of receipts and shipments. Require access to scan logs and timestamps, align approval thresholds, and include accuracy SLAs and corrective actions in the contract.
Which KPIs best track inventory accuracy and shrinkage over time?
Track inventory accuracy, location accuracy, shrinkage rate, count hit rate, order fill rate, and perpetual-to-GL tie-out. Segment by site, SKU class, and supplier to target improvements.
Use Verve AI for Shopify to forecast demand and create smarter purchase orders
Or install Verve AI for WooCommerce to manage inventory planning without spreadsheets
Both versions are built specifically for ecommerce inventory workflows.