How often should the item master be audited?

Run the internal-consistency tests monthly; they're cheap. Run the receipt-ratio test (test 7) quarterly, and after any vendor onboarding, packaging change notice or WMS migration. Items with recurring count variances should be checked against receipts the day they're flagged.

Who should own the item master?

One named function, usually inventory control or a master data team, with change rights restricted to it. Buyers and receivers should be able to request a change, with evidence, but not make one. Shared ownership is how a field ends up changed three times in a month with no record of why.

Should vendors supply our item data?

Vendor-supplied data (from GS1 data pools, EDI or item setup sheets) is a good starting point and a poor final authority. Vendors change packaging without updating what they sent you. Verify the pack and dimensions at the first receipt and whenever receipts start showing a constant per-case delta.

Service guide · WMS data quality

Item master audit:
the WMS fields that multiply every error

Most WMS data quality problems are item master problems. A wrong field doesn't cause one error. It causes the same error on every transaction that touches the item until someone changes it. This is how to find those fields from an export.

Updated 17 September 2026
Written by WMSAudit
A Ravenspire LLC company
10 min read

Short answer

An item master audit tests the product records your WMS converts, stores and plans against: base unit of measure, conversions, case and inner pack, pallet configuration, dimensions, weights and control flags. Checking the fields for internal consistency isn't enough. The decisive test compares them with what receipts, counts and vendor ASNs actually show, because a wrong pack field produces the same variance on every receipt of that item.

Why the item master comes first

Almost every inventory error has a size and a frequency. Most errors are large and rare: a pallet put away to the wrong aisle, a receipt keyed twice. Item master errors are the opposite. They're usually small per event and they happen on every event, because the field is read by every transaction that touches the item.

A case pack of 12 where the vendor ships 14 posts two units short on every case received. The receiver did nothing wrong, the count will find it, and the adjustment will correct it. Then the next receipt brings it back. That's why item master errors are the ones most likely to survive a good cycle count program. The count program is built to correct events, and this isn't an event.

It's also why the item master belongs at the top of any WMS audit. When findings are ranked by annual cost, a quiet pack error on a fast mover routinely outranks every dramatic one-off in the building.

The fields that matter, and what each one breaks

Base unit of measure: The unit every balance is stored in. If it's wrong, or inconsistent across similar items, every conversion built on it inherits the error. Watch for items stocked in eaches whose base is set to cases, and the reverse.
UOM conversions (each, inner, case, pallet): What turns "3 CS received" into a stored quantity. A wrong factor scales the receipt: see unit-of-measure errors for how one field becomes a factor-of-12 variance.
Case pack and inner pack: The most frequently wrong fields in most item masters, because vendors change them and rarely say so. Inner pack entered in the case pack field is a common and costly slip.
Pallet configuration (TI × HI): Drives putaway capacity, receiving by pallet and replenishment by pallet. A wrong TI or HI misstates every full-pallet receipt.
Dimensions and weights by level: Used for slotting, location capacity, cartonization and freight. Each-level dimensions copied into the case record make every case look the size of one unit.
Control flags: Lot, serial, expiry, catch weight, hazard class, temperature class. A missing flag means the system can't enforce the rule, so FEFO, segregation and recall traceability quietly stop working.
Status and replenishment settings: Inactive items holding stock, min/max in a UOM the pick face doesn't use, velocity classes that haven't been reviewed since go-live.
Identifiers by level: GTIN or UPC at each packaging level. The same barcode on the each and the case means a scan of either records the same unit, which converts a case scan into one each.

A worked example: inner pack in the case field

An item ships as cases of 24 eaches, each case holding four inners of 6. During setup, someone enters 6 in the case pack field. Receivers scan cases, and the WMS records 6 eaches per case. The shelf holds 24.

One item · case pack 6 recorded, 24 actual Illustrative arithmetic

Effect of an inner pack entered as case pack
Line	Recorded	Physical	Working
One receipt of 40 CS	240 EA	960 EA	40 × 6 vs 40 × 24
Understated per receipt	720 EA physically unrecorded		Ratio 24 ÷ 6 = 4, exactly the inners per case
Over 26 receipts a year	18,720 EA		720 × 26

At $1.85 a unit$34,632· of stock the system doesn't know it holds, each year, on one item

Substitute your own cost and volume. What matters is the signature. The ratio between physical and recorded is exactly 4, which is the number of inners in a case. A variance ratio that equals a pack relationship is not a counting problem. Understated stock like this rarely shows up as an alarm. It shows up as buyers reordering stock the building already holds, as overflow locations nobody can explain, and eventually as a large positive adjustment someone books as a "found" gain.

Twelve item master tests

Tests 1 to 6 check the item master against itself and can be run on a single export. Tests 7 to 12 check it against reality, meaning receipts, counts, ASNs and the change log, and they're the ones that find the expensive errors.

Test 01

Blanks, zeros and default-of-one

Active items with a blank, zero or 1 in case pack, inner pack, TI, HI, dimensions or weight. A pack of 1 is legitimate for some items and a default left over from setup for many others. List them and ask which is which.

Test 02

Hierarchy integrity

Inner pack must divide evenly into case pack, and case into pallet (TI × HI × case). Flag any item where it doesn't, or where a conversion is fractional for a unit that can only be whole.

Test 03

Dimensions that don't scale

Case cube should be at least case pack × each cube. Flag cases smaller than their contents, and items where each and case dimensions are identical.

Test 04

Weights that don't scale

Case weight should be roughly case pack × each weight, plus packaging. Large deviations point at a pack or weight error, and either one breaks cartonization and freight.

Test 05

Duplicate identifiers

The same GTIN or UPC on two items, or on two packaging levels of one item. Either way a scan can record the wrong item or the wrong quantity.

Test 06

Missing control flags

Items in categories that require lot, expiry, hazard or temperature control but carry none. Cross-check against your own segregation rules and vendor SDS data.

Test 07

The receipt ratio test

The most valuable test here. For each item and vendor, take received-and-counted variance and divide by cases received. A constant non-zero whole number across receipts means the pack field is wrong by that amount. Random scatter means it isn't.

Test 08

Variance ratios that equal a conversion

For count variances, compute physical ÷ recorded. Ratios of 4, 6, 12, 24 or a pallet quantity that match the item's own conversions are conversion errors, not miscounts.

Test 09

ASN pack vs master pack

Where vendors send ASNs with pack detail, compare the advised pack with the master. A disagreement is either a vendor change you haven't absorbed or a vendor error, and the receipt history tells you which.

Test 10

Change-log churn

Items whose pack or UOM fields changed more than once in a year, or were changed by someone outside the owning team. Churn usually means two people are "fixing" the same field toward different answers.

Test 11

Variance onset vs change date

For items with recurring variances, check whether they started on the date a pack or UOM field changed. If they did, find out whether the WMS restated existing on-hand when the field changed. Systems differ, and the answer is often "for some transaction types".

Test 12

Inactive items with stock

Items marked inactive, discontinued or blocked that still hold on-hand. Either the stock is phantom, or it's real and invisible to planning. See phantom inventory.

Running it from an export

None of this needs access to the live WMS. Three exports cover all twelve tests:

The full item master, active and inactive, with every UOM level, pack field, TI, HI, dimension, weight, identifier and control flag. Don't filter it. Blank fields matter as much as wrong ones.
Twelve months of receipts with PO line, vendor, quantity and UOM as entered, converted base quantity, and any receipt-time count or variance.
The item master change log with field, old value, new value, user and effective date. This is the export most often missing from a data request, and it's often the one that settles the question.

Add count history and ASN detail if you have them, for tests 8 and 9. A spreadsheet will handle tests 1 to 6 for a few thousand items. The receipt ratio test is easier in anything that can group and divide across a year of receipts. The logic is simple, and the volume is what makes it slow by hand.

Keeping it clean after the audit

One owner, restricted rights. Change rights on pack and UOM fields belong to one team. Everyone else requests a change with evidence.
Every change carries an effective date and a reason. "Vendor changed pack per ASN 7731, 14 → 12" is a record. A silent edit is how test 11 becomes impossible.
Verify at first receipt. New items and new vendors get a physical check of pack, dimensions and weight at the first receipt, before the data is trusted.
Watch the per-case delta. Test 7 run monthly catches vendor pack changes within a few receipts instead of at the next physical inventory.
Fix at the source. If the error came from vendor-supplied data, send the correction back to the vendor, or the next item setup will repeat it.

WMS data quality beyond the item master

The item master is where WMS data quality problems do the most damage per field, but it isn't the only master. The location master decides whether a balance points at a real slot (orphaned and unmapped locations are a common finding). The configuration decides what the system permits: negative on-hand, over-receipt tolerance, keyed confirmations, half-confirmed moves. The transaction data shows whether the process followed the rules the configuration was meant to enforce.

A full WMS data quality review covers all four. The WMS audit checklist sets out 48 checks across them. Start with the seven item master checks, then work outward.

The operational takeaway

An item master can pass every consistency check and still be wrong, because the pack field is internally consistent at 12 while the vendor ships 14. What proves an item master is the receipt history, not the item master itself.

If you only run one test, run test 7: variance per case, by item and vendor, over a year. A constant number is a field somebody can fix this afternoon, and it will keep costing money until someone does.

Questions

How often should the item master be audited?: Run the internal-consistency tests monthly; they're cheap. Run the receipt-ratio test (test 7) quarterly, and after any vendor onboarding, packaging change notice or WMS migration. Items with recurring count variances should be checked against receipts the day they're flagged.
Who should own the item master?: One named function, usually inventory control or a master data team, with change rights restricted to it. Buyers and receivers should be able to request a change, with evidence, but not make one. Shared ownership is how a field ends up changed three times in a month with no record of why.
Should vendors supply our item data?: Vendor-supplied data (from GS1 data pools, EDI or item setup sheets) is a good starting point and a poor final authority. Vendors change packaging without updating what they sent you. Verify the pack and dimensions at the first receipt and whenever receipts start showing a constant per-case delta.