Procurement

UNSPSC voor MRO: Classificeren van Onderhoud, Reparatie en Bedrijfsvoering Spend

MRO procurement is de moeilijkste categorie om consistent te classificeren. Dit is waarom UNSPSC werkt voor onderhouds-spend, en wat er nodig is om zuivere MRO data te krijgen uit een systeem zoals SAP of Maximo.

Procurement5 May 20267 min read

Why MRO spend is the hardest to classify

Elke procurement categorie heeft classificatie-uitdagingen. MRO spend — de onderdelen, materialen, gereedschappen en diensten die installaties en assets draaiende houden — heeft er meer dan de meeste.

Dit maakt MRO classificatie moeilijk:

Onderdeelbeschrijvingen zijn geschreven voor engineers, niet voor category managers. "Seal kit 47mm OD — pump bearing housing" is precies genoeg voor de maintenance ingenieur die het bestelde. Het vertelt een classificatiesysteem heel weinig over of dit valt onder Industriële Machines (Segment 23), Industriële Apparatuur (Segment 24), of Elektronische Componenten (Segment 32).

Hetzelfde onderdeel heeft tientallen beschrijvingen. Verschillende locaties, verschillende engineers, verschillende supplier catalogi — één enkel fysiek component verschijnt onder veel verschillende benamingen in de inkoopgeschiedenis van hetzelfde bedrijf.

Onderdeelnummers zonder beschrijvingen zijn gebruikelijk. Buyer typt "47234-B" in de PO. Het systeem heeft supplier context, GL account en historische

MRO spend is decentralised. In manufacturing and FM environments, maintenance buyers often have significant local autonomy. Category management at headquarters may not see the spend detail until months later.

The result is an MRO spend file where 30–50% of lines are either unclassified, inconsistently classified, or parked in catch-all categories that make analysis impossible.


What UNSPSC segments cover MRO spend

Most MRO spend for manufacturing, FM, and infrastructure companies falls across these UNSPSC segments:

SegmentDescriptionTypical MRO spend
23Industrial Machinery and EquipmentPumps, compressors, motors, conveyors
24Power Generation and DistributionElectrical components, switchgear, cables
26Electronic Components and SuppliesSensors, controllers, relays
31Manufacturing Components and SuppliesFasteners, seals, gaskets, bearings
47Cleaning and Janitorial SuppliesMaintenance cleaning products
72Construction and Maintenance ServicesLabour and contractor services
73Industrial Production and ManufacturingFabrication and specialist work

Getting classification right at the commodity level — 8 digits — within these segments requires either extensive manual lookup against the UNSPSC codeset or an automated engine trained on MRO-specific data.


The case for UNSPSC over manufacturer part numbers

Some procurement teams argue that manufacturer part numbers (MPNs) are a better way to track MRO spend than UNSPSC codes. MPNs are precise and enable exact duplicate identification. But they have significant limitations:

  • MPNs are supplier-specific. The same component from a different manufacturer has a different MPN, making cross-supplier analysis impossible.
  • MPNs go obsolete. When a manufacturer discontinues a product, its MPN disappears. UNSPSC commodity codes are stable across supplier changes.
  • MPNs do not enable category-level aggregation. You cannot ask "how much did we spend on bearings last year?" from MPN data alone. UNSPSC enables exactly this.

The answer for most companies is both: enrich your spend data with MPNs where possible, and apply UNSPSC codes for category-level spend analysis. These are complementary, not competing, approaches.


How automated MRO classification works in practice

A good automated classification engine for MRO spend uses multiple signals beyond the line item description:

Supplier context. If a supplier is known to supply only bearings and seals (from historical purchasing patterns), even a bare part number triggers a high-confidence classification in the right commodity area.

GL account routing. Maintenance spend coded to GL account 6400 (or its equivalent) signals a different probability distribution across UNSPSC segments than capital expenditure coded to GL 0700.

Historical priors. If the same part description has been classified as 31161500 (Plain bearings) 95 times in the past twelve months, the 96th occurrence is classified automatically at high confidence.

LLM augmentation. For genuinely ambiguous descriptions, a large language model brings broad product knowledge. "Seal kit 47mm OD — pump bearing housing" resolves to the correct commodity code because the model understands what a pump bearing housing seal kit is, even if it has never seen this exact string before.


Enriching to manufacturer part number level

For MRO categories where part number precision matters — typically for companies that want to go direct to manufacturer and cut out distributor margin — UNSPSC classification is the first step, not the last.

Once spend is classified at commodity level, the next step is part number enrichment: matching line items to a manufacturer's official catalogue to identify the exact OEM part, the manufacturer's direct price, and alternative sourcing options.

This is the approach used in Pearstop's MRO work. Classification gives you the category view. Part number enrichment gives you the commercial lever — knowing that the bearing you are buying from a distributor at margin is available direct from the manufacturer at 30% less.

The combination of UNSPSC classification and part number enrichment is what makes the "go direct to manufacturer" strategy operationally feasible at scale.


Getting started with MRO classification

The fastest route to clean MRO spend data is a Data Stability Baseline:

  1. Export your spend data from SAP, Maximo, or your procurement system — typically 12–24 months of purchase history
  2. Pearstop classifies the export using the four-layer engine, returning UNSPSC codes at commodity level
  3. Your team reviews the flagged items — typically 5–10% of lines — via a review interface
  4. The classified dataset becomes your spend baseline for category analysis and negotiation

From that baseline, ongoing classification of new MRO purchases runs automatically each month.


The downstream value of clean MRO data

A classified MRO spend dataset unlocks several commercial decisions that were impossible with unclassified free-text data:

Supplier consolidation. You can see how many suppliers you are using in each commodity category, and whether consolidating to fewer preferred suppliers would produce better pricing or service terms.

Price benchmarking. With consistent category codes, you can compare the price you are paying for the same UNSPSC commodity across suppliers, sites, and time periods.

Spend baseline for contract negotiation. A classified spend baseline is the foundation for any meaningful negotiation. Without it, suppliers know more about your spend than you do.

De identificatie van de directe inkoopmogelijkheid bij de fabrikant. Artikelnummerverrijking bovenop UNSPSC classification identificeert welke MRO-goederen de grootste marge beschikbaar hebben door direct in te kopen.


Verder lezen: Nauwkeurigheid van UNSPSC classification — Wat 90–95% werkelijk betekent | Hoe UNSPSC classification in SAP te implementeren | Problemen met de asset register die slim onderhoud voorkomen

Free Tools

Not sure which UNSPSC code to use?

Paste any product or service description and get the correct 8-digit code instantly — or explore the full taxonomy tree to understand the hierarchy.

Richard Wallace

Richard Wallace

Co-founder, Pearstop

Richard brings deep commercial experience in hard services and FM. He works with clients to design data quality programmes that translate directly into procurement performance and contract accuracy.

LinkedIn →

Further reading

Laatste inzichten

Procurement

Hoe Betere Inkoopdata Facilitaire Bedrijven Helpt Marges te Beschermen bij Olieprijsvolatiliteit

Wanneer olieprijzen stijgen, vreten brandstoftoeslagen stil de marges weg in facility management. Be…

Lees meer
AI & Digital

Hoe U de AI-race in Procurement Wint: Begin met Schone Data

AI in procurement werkt alleen als uw data schoon, geclassificeerd en betrouwbaar is. Leer hoe Pears…

Lees meer
AI & Digital

Wat Is Procurement Data Readiness en Waarom Is Het Belangrijk voor AI?

Procurement data readiness is de basis voor AI in procurement. Leer wat het betekent, waarom datakwa…

Lees meer