MRO Catalog Normalisation: UNSPSC Coding, UOM Standardisation, and Maximo Item Master Field Mapping
The three technical disciplines that determine whether your spare parts catalog survives an EAM migration — and how to get them right before you load a single record.
Most EAM migrations fail at the parts catalog, not the equipment register. The equipment register is messy but bounded — a mine site has a finite number of assets. The parts catalog is unbounded, accumulated over decades, and touched by every contractor, every OEM, and every storeroom manager who ever worked the site.
Key Takeaways
- UNSPSC commodity coding provides a vendor-neutral classification hierarchy for MRO parts — essential for cross-site spend analysis and catalog rationalisation.
- UOM standardisation is a prerequisite for accurate inventory valuation — a catalog with five representations of the same unit of measure will produce incorrect reorder calculations.
- Maximo Item Master has strict field constraints: DESCRIPTION is 100 characters, ITEMNUM is 30 characters, and ORDERUNIT must match a valid UOM in the MEASUREUNIT domain.
- Description standardisation using noun-modifier format (BEARING, BALL, 6205-2RS, 25X52X15MM) makes parts searchable and enables accurate duplicate detection.
- The combined equipment + parts workflow — linking normalised spare parts to the assets that use them — is the foundation of condition-based maintenance in Maximo.
Why the Parts Catalog Is the Hardest Part of an EAM Migration
Most EAM migrations fail at the parts catalog, not the equipment register. The equipment register is messy but bounded — a mine site has a finite number of assets. The parts catalog is unbounded, accumulated over decades, and touched by every contractor, every OEM, and every storeroom manager who ever worked the site.
A 50,000-line catalog is not unusual. A 200,000-line catalog is not rare. And within that catalog, the same physical part will appear under dozens of different descriptions, part numbers, and OEM names — each one a separate record, each one carrying separate stock, each one driving separate reorder calculations.
Normalising that catalog before migration requires three distinct technical disciplines: UNSPSC commodity coding, UOM standardisation, and Maximo Item Master field mapping. This guide covers all three.
UNSPSC Commodity Coding
UNSPSC (United Nations Standard Products and Services Code) is an eight-digit commodity classification code that provides a vendor-neutral hierarchy for MRO parts. The hierarchy has four levels: segment (two digits), family (two digits), class (two digits), and commodity (two digits).
For MRO parts, the relevant segments are primarily Segment 31 (Manufacturing Components and Supplies) and Segment 40 (Industrial Production and Manufacturing Services). Within those segments, the codes that appear most frequently in mine-site catalogs are:
| UNSPSC Code | Description |
|---|---|
| 31161500 | Ball bearings |
| 31161600 | Roller bearings |
| 31161700 | Thrust bearings |
| 40141700 | Hydraulic filters |
| 40141600 | Air filters |
| 31351500 | O-rings |
| 31351600 | Gaskets |
| 26111700 | V-belts |
| 40141900 | Lubricants and oils |
| 39121400 | Electrical cables |
Applying UNSPSC codes to a parts catalog enables three things that are otherwise impossible: cross-site spend analysis (how much are we spending on bearings across all sites?), catalog rationalisation (which commodity classes have the most duplicate records?), and supplier consolidation (which commodities are we buying from too many suppliers?).
The challenge is that UNSPSC coding requires understanding what a part actually is — not just what it is called. A record described as "BRG 6205 ZZ 2RS DEEP GRV" needs to be understood as a ball bearing before it can be assigned code 31161500. This is where automated classification with a parts-specific taxonomy produces materially better results than generic text classification.
UOM Standardisation
Unit of measure inconsistency is the most common data quality problem in mine-site MRO catalogs — and the most consequential. A catalog with five representations of the same unit of measure will produce incorrect inventory valuation, incorrect reorder calculations, and import errors in both Maximo and SAP MM.
The most common UOM variants found in mine-site catalogs are:
| Canonical UOM | Common Variants |
|---|---|
| EA (each) | Each, each, 1, PCS, Pcs, pcs, UNIT, Unit, NO, No, NOS |
| L (litre) | LTR, Litre, litre, LT, ltr, LITRE |
| KG (kilogram) | Kg, kg, KGS, Kgs, KILOGRAM |
| M (metre) | MTR, Mtr, METRE, Metre, metre |
| M2 (square metre) | SQM, Sqm, M2, m2 |
| M3 (cubic metre) | CUM, Cum, M3, m3 |
| BOX | Box, box, BX, Bx |
| PKT (packet) | Packet, packet, PACK, Pack, pack |
Normalisation maps all variants to the canonical form. The canonical form must match a valid entry in the EAM platform's unit of measure domain — in Maximo, this is the MEASUREUNIT domain; in SAP MM, it is the base unit of measure (MEINS) field.
One complication: some UOM variants are not simply alternative spellings — they represent different quantities. "BOX" might contain 10 items at one site and 25 at another. Before normalising BOX to EA, the catalog needs to be reviewed for records where the UOM carries quantity information that would be lost in the conversion.
Try Struktive on your own data
Upload a raw asset CSV and get back a normalised, DCIM-ready file in minutes. No account required.
Maximo Item Master Field Mapping
IBM Maximo Item Master has specific field constraints that determine whether a parts record will import cleanly. The key fields and their constraints are:
| Field | Constraint | Common Issue |
|---|---|---|
| ITEMNUM | 30 characters max | Legacy part numbers sometimes exceed 30 chars |
| DESCRIPTION | 100 characters max | Long descriptions need truncation without losing meaning |
| ORDERUNIT | Must match MEASUREUNIT domain | Non-standard UOM values fail validation |
| COMMODITY | Must match COMMODITY domain | UNSPSC codes not pre-loaded in domain |
| ITEMTYPE | ITEM, TOOL, MATERIAL, or SERVICE | Blank or non-standard values fail |
| ROTATING | Y or N | Must be explicitly set; blank causes issues |
| LOTTYPE | NOLOT, LOT, or SERIALIZED | Blank defaults to NOLOT but should be explicit |
The DESCRIPTION field constraint is the one that causes the most problems. A 100-character limit sounds generous until you try to fit a full noun-modifier description with OEM name, part number, and key specifications into it. The discipline of noun-modifier format — BEARING, BALL, 6205-2RS, 25X52X15MM — is partly a response to this constraint: it packs the maximum information into the minimum characters.
The COMMODITY field requires pre-loading the UNSPSC codes you intend to use into the Maximo COMMODITY domain before import. If you apply UNSPSC codes during normalisation but do not pre-load them into the domain, every record with a COMMODITY value will fail validation.
Description Standardisation: Noun-Modifier Format
Noun-modifier format is the industry standard for MRO part descriptions. The structure is: NOUN, MODIFIER 1, MODIFIER 2, MODIFIER 3 — where the noun is the generic name of the part and the modifiers provide specificity in order of importance.
For a ball bearing: BEARING, BALL, 6205-2RS, 25X52X15MM For a hydraulic filter: FILTER, HYDRAULIC, RETURN LINE, 10 MICRON For a solenoid valve: VALVE, SOLENOID, 24VDC, NORMALLY CLOSED For a V-belt: BELT, V, B-SECTION, 1270MM
This format has three practical advantages. First, it makes parts searchable — a maintenance planner looking for bearings can search "BEARING" and find all bearing records regardless of how they were originally described. Second, it enables accurate duplicate detection — two records with the same noun-modifier description and OEM are almost certainly the same part. Third, it fits within the Maximo 100-character DESCRIPTION constraint while preserving the information needed to identify the part unambiguously.
Applying noun-modifier format to an existing catalog requires understanding what each part is — which is why description standardisation and UNSPSC coding are best done together. The same classification step that assigns a UNSPSC code also determines the correct noun for the description.
The Combined Equipment and Parts Workflow
The most valuable output from parts normalisation is not a clean catalog in isolation — it is a clean catalog linked to a clean equipment register. When both datasets are normalised to the same standard, Struktive can automatically link spare parts to the assets that use them: the 6Y3222 GET tooth linked to every CAT 793F haul truck on site, the hydraulic filter linked to every Komatsu PC5500 shovel.
This linkage — stored in Maximo as ITEMORGINFO records with ASSETNUM references — is the foundation of condition-based maintenance. When a work order is raised against an asset, Maximo can automatically suggest the spare parts associated with that asset class. When a part is consumed, the consumption is linked to the asset, building the failure history that drives predictive maintenance.
That linkage is only possible when both the equipment register and the parts catalog are normalised to the same standard. A clean equipment register with a dirty parts catalog produces a Maximo implementation where work orders are raised but parts cannot be found. A clean parts catalog with a dirty equipment register produces a storeroom that cannot be linked to the assets it supports.
The combined workflow — normalise both datasets, link them, then load them — is the only sequence that produces a Maximo implementation that works from day one.