What Is Unit of Measure (UOM) in Product Data?
UOM errors silently corrupt pricing, matching, and AI search. Learn how unit of measure works, why it breaks, and how to fix it across supplier feeds.
When a supplier feed says a cable costs “$0.45” and a competing feed says “$45.00,” the catalog looks like an obvious pricing error — until you realize one is priced per meter and the other per 100-meter reel. Unit of measure (UOM) is the field that makes that distinction machine-readable, and when it is wrong or missing, it silently corrupts pricing comparisons, product matching, duplicate detection, and AI-generated answers across your entire catalog. Claro resolves this by normalizing UOM strings across every inbound supplier feed, attaching the conversion factors that link pack levels, and writing clean coded values back into your PIM or ERP — so the catalog stays trustworthy as suppliers change how they describe the same item.
Definition
A unit of measure is the standardized quantity — such as each, box, meter, kilogram, or liter — that defines how a product is counted, priced, packaged, or sold. The answer to “what is unit of measure” sounds simple, but the implementation is genuinely hard. A value of 12 means nothing until you know whether it is 12 each, 12 cases, 12 meters, or 12 kilograms.
Most SKUs carry several UOMs at once. The base UOM is the smallest sellable or stockable unit. The order UOM is how customers actually buy — often a box, case, or pallet. One or more packaging UOMs describe inner packs, cartons, and pallets. The relationships between levels are expressed as conversion factors: one case equals 24 each, one pallet equals 40 cases, one reel equals 305 meters.
Standards bodies maintain controlled code lists so these units are machine-readable rather than free text. The most common is UNECE Recommendation 20, which assigns codes such as MTR (metre), KGM (kilogram), EA (each), and LTR (litre). GDSN, most ERP platforms, and most PIM systems expect UOM values to map to such a code list rather than ad-hoc strings like “pcs”, “pc”, “piece”, or “stk”. When that mapping is missing or inconsistent, the unit becomes a silent source of cascading error.
Why unit of measure breaks product catalogs
UOM is where quantity, price, and identity intersect — so an error here propagates everywhere downstream. The same physical product sold as “1 EA” by one supplier and “1 BOX of 10” by another looks like two different items to a matching engine that has not normalized units first, even when both records share a manufacturer part number.
| Industry | UOM problem in the supplier feed | Downstream impact |
|---|---|---|
| Industrial distribution | Cable priced per meter by one supplier, per 100 m reel by another | Price comparison is off by 100x; cheapest source appears most expensive |
| CPG | Beverage listed as each vs. 24-pack case across retailer data pools | Duplicate listings created; inventory and reorder math breaks |
| MRO | Fasteners priced per 1,000 vs. per each depending on the supplier template | Buyers over-order by three orders of magnitude |
| Furniture | Upholstery fabric quoted in linear yards by some mills, square meters by others | Quotes and cut lists are silently wrong; margin is lost at fulfillment |
For enrichment and AI-readiness, UOM is just as load-bearing. An AI assistant answering “how much wire do I need for this run?” can only reason correctly when quantities, pack sizes, and units are explicit and consistent. Free-text units — “box”, “bx”, “carton/24” — are not citable by a generative engine. Coded, converted units are.
Messy vs. trusted: what UOM normalization changes
| Before UOM normalization | After UOM normalization with Claro |
|---|---|
| Same item appears as 3–5 records with different unit strings | One resolved entity with a canonical base UOM and pack hierarchy |
| Price comparison returns 100x errors across supplier feeds | Every offer converted to a shared base unit before comparison |
| Matching engine treats 'EA' and 'BOX/10' as different products | Conversion factors applied; records resolve to the same product |
| AI search returns inconsistent quantity answers | Coded units with provenance give AI a citable, consistent source |
| Manual remediation required every time a supplier changes their template | Automated normalization catches drift and writes clean values back to PIM/ERP |
How to normalize UOM across a supplier catalog
- 1Detect the raw unit string
Parse whatever the supplier sent — “pcs”, “M”, “case/24”, “reel (305m)” — including any embedded pack quantities. Flag records where the unit field is blank or free text with no code mapping.
- 2Map to a standard code
Resolve each raw string to a controlled value, such as a UNECE Rec 20 code. Maintain a crosswalk table that covers the supplier-specific variants you encounter most often so the mapping is reusable across future feeds.
- 3Capture conversion factors
Record base UOM, order UOM, and all packaging UOMs along with the numeric factors that relate them. A cable SKU, for example, should carry EA as the base UOM, MTR as the order UOM, and a factor of 305 for the reel-level packaging UOM.
- 4Normalize for comparison
Convert every competing offer to the same base unit before running matching, price comparison, or deduplication. A 100x pricing error that was invisible before normalization becomes immediately obvious once all offers are expressed in the same unit.
- 5Write back to PIM or ERP
Push the normalized UOM codes and conversion factors back into your system of record with full source attribution. Claro’s write-back keeps the catalog clean as new supplier feeds arrive, without requiring manual re-entry.
Related
Glossary
What Is Data Normalization?
The broader practice that UOM standardization fits inside — turning messy supplier values into consistent, comparable data.
Glossary
UNECE Rec 20 Unit Codes
The international standard code list that makes UOM values machine-readable across ERP, PIM, and GDSN systems.
Glossary
What Is GDSN?
The global data pool network where coded UOMs and pack hierarchies are mandatory for product data exchange.
Tool
UNECE Rec 20 Unit Code Validator
Check whether your UOM values map to valid standard unit codes before they reach your PIM or ERP.
Tool
Unit and Dimension Converter
Convert lengths, weights, and dimensions between common product units to verify conversion factors.
Playbook
Map Supplier Attributes to Your Schema
Step-by-step workflow for mapping supplier UOMs and attributes onto your canonical product model.
FAQ
What is the difference between base UOM and order UOM?
The base UOM is the smallest unit you stock or sell — typically each (EA). The order UOM is how customers actually purchase the item, which might be a box, case, or pallet. A single SKU can carry a base UOM of EA and an order UOM of CS (case) of 12, with a conversion factor linking them. Keeping both explicit prevents quantity and pricing errors in every downstream system that reads the record.
What is a UNECE Rec 20 unit code?
UNECE Recommendation 20 is an international code list that assigns short codes to units of measure — MTR for metre, KGM for kilogram, LTR for litre, EA for each, and hundreds more. Mapping free-text values like ‘pcs’, ‘M’, or ‘bx’ to these codes makes UOM data machine-readable and interoperable across ERP, PIM, and GDSN systems.
Why do duplicate products often have mismatched units of measure?
Different suppliers describe the same physical item using different units — one lists it as 1 EA, another as 1 BOX of 10. A matching engine that compares quantities or prices without first normalizing UOM treats these as two distinct products, creating duplicates. Converting every offer to a common base unit before comparison eliminates most of these false splits.
How does unit of measure affect price comparison?
Prices are only comparable when expressed in the same unit. ‘$45 per reel’ and ‘$0.45 per meter’ look wildly different until you apply the conversion factor (100 meters per reel). Automated sourcing tools that skip this step pick the wrong supplier every time — often by orders of magnitude.
Does unit of measure matter for AI search and generative answers?
Yes. AI assistants and generative engines can only answer quantity and sizing questions accurately when units are explicit and consistent. Coded, converted units with clear pack hierarchies are citable. Free-text values like ‘bx’ or ‘carton/24’ are not, which makes it much harder for AI to surface and quote your products reliably.
How does Claro help with UOM normalization across supplier feeds?
Claro detects raw unit strings from every supplier feed, maps them to standard UNECE Rec 20 codes, captures base-to-order conversion factors, and writes the normalized values back into your PIM or ERP. Because every enriched value carries a source trace, your team can audit exactly where each UOM came from and correct outliers without touching the rest of the catalog.
Claro
See how Claro handles this in production
This concept is one piece of keeping a catalog trusted. See how Claro resolves identity, enriches missing attributes, and validates every update before it reaches your PIM or ERP.
Learn more