What is Supply Chain Traceability?

What is Supply Chain Traceability?

Supply chain traceability is the ability to answer two questions about any part in any unit you have ever shipped. Where did it come from: which supplier, which lot or batch, which certificate of conformance, handled how? And, in the other direction, where else did that lot go: which other units consumed it, and who has them now?

The second question is the one that matters, and the one most organizations cannot answer. Backward traceability, from a unit to its origins, can usually be reconstructed from paperwork if you have a week. Forward traceability, from a supplier lot to every serial number that consumed it, only exists if someone captured lot consumption as each unit was built. That capture is the whole discipline. Everything else is architecture around it.

Traceability is product genealogy extended upstream past your receiving dock. Genealogy records what happened to a specific unit inside your walls and in the field. Traceability adds where its parts came from before they arrived, which means it depends on supplier data that most companies keep in a spreadsheet outside the product record.

Three levels of resolution

Level	Question it answers	System of record	Captured when
Part	Which suppliers and manufacturer part numbers may supply this part at this revision?	PLM, as the approved vendor list	Part release and supplier qualification
Lot or batch	Which supplier lot went into this unit, and which units consumed this lot?	MES, as lot consumption recorded at each operation	Point of manufacture
Serial	What is the complete history of this individually identified component?	MES and service systems, as the unit's genealogy	Manufacture and every subsequent service event

Part level is the foundation, and it lives in PLM. The PLM and supply chain integration guide makes the approved vendor list the first phase for exactly this reason: every downstream trace starts from knowing which suppliers were allowed to supply which part. If that relationship is not inside PLM, linked to the bill of materials, nothing built on top of it holds.

Lot level is where traceability becomes real, and it lives in MES, not PLM. The as-built record originates on the floor. Each time a unit passes an operation that consumes a lot-controlled material, the lot identifier is scanned against the serial number. That is a discipline and a scanning infrastructure before it is software, and it is the cheapest traceability investment you will ever make, because reconstructing it afterwards costs orders of magnitude more.

Serial level is the deepest form: individually identified components with their own genealogy. Aerospace mandates it for life-limited and critical parts under AS9100. Medical devices carry it through the FDA's unique device identification rules. Automotive is getting there through software-defined vehicles, where the version running on a unit is as important as the parts in it.

Discrete and process are different problems

Discrete manufacturers trace serials and lots against an assembly structure. The question is which physical parts went into which unit. Process manufacturers trace batches against a recipe. The question is which ingredient lots went into which batch, and where that batch was split, blended, reacted or packaged. The discrete vs process PLM article explains why these need different data models, and why mainstream PLM platforms built for assembly handle batch genealogy poorly without industry extensions. Process traceability also carries regulatory obligations, FDA 21 CFR Part 11 and EU GMP Annex 11 among them, that discrete serialization does not.

Why the floor is rising

Three pressures, and they are arriving together.

Regulation. The EU's Ecodesign Regulation is introducing the Digital Product Passport, which requires part-level knowledge of materials and origin for a growing list of product categories. REACH already demands substance-level declarations. Conflict minerals reporting requires smelter-level provenance for tin, tungsten, tantalum and gold. FDA UDI, AS9100 and IATF 16949 impose their own. PLM is the system of record for all of it, but only if the supplier data is inside it.

Reshoring. Companies rebuilding regional supply chains are onboarding new suppliers at a rate not seen in decades. Each one has to be qualified, documented and linked to parts. Doing that in spreadsheets builds the data debt that will slow the next crisis response.

Recall economics. A recall bounded by traceable lot and serial data is a service action. A recall that requires physical inspection of every unit built in a date window because nobody captured lot consumption is a crisis. The gap between those two outcomes is the business case, and it is not subtle.

How it works in the product record

The mechanics, in the order they have to be built:

1. Part-supplier relationships in PLM. Every purchased part carries its approved manufacturers and manufacturer part numbers, with qualification status and effectivity by revision.
2. Release events as triggers. When a part or an engineering change is released, the affected suppliers are notified from the system, not by email. The supply chain guide calls this closed-loop change management, and it is traceability in the other direction: it is how you avoid a supplier building the wrong revision in the first place.
3. Lot capture in MES. Lot-controlled materials are scanned against the serial number at each consuming operation. This is where forward traceability is born.
4. Where-used queries across the seam. A lot in MES resolves to serial numbers; serial numbers resolve in PLM to configurations, effectivities and customers. The PLM integration between the two is what makes the query possible.
5. Nonconformance and CAPA in QMS. When a supplier issue surfaces, the quality event is recorded in the QMS and linked to the engineering change it triggers, so the trace is auditable end to end.

The 737 MAX is the reference case for what happens when a safety-relevant change is not traced through this chain. It is usually filed under digital thread, and it belongs here too: traceability is the digital thread at the level of individual units and individual lots.

The recall query

Here is what working traceability looks like, in one scenario. A brake caliper supplier reports that one casting lot was heat-treated out of specification. The OEM queries MES for every vehicle that consumed the lot: 1,140 units across two plants. It pulls their configurations from PLM and their owners from the dealer system. It issues a targeted recall within 48 hours.

The same event without lot capture: every vehicle built in the six-week window during which that lot could have been consumed, at every plant that received castings from that supplier. Tens of thousands of units, most of them fine, all of them inspected.

If your systems cannot run the first query today, that is the gap to close, and the supply chain integration guide sequences the work.

What traceability is not

It is not blockchain. A shared immutable ledger can be useful when several parties who do not trust each other need to agree on a custody record, which is why it turns up in conflict minerals and ethical sourcing programs. It does nothing about the real problem, which is capturing lot consumption on the floor and holding supplier data in the product record. An immutable ledger of data nobody collected is not traceability.

It is also not a reporting feature. Traceability that is assembled after the fact, for the audit, from documents, is documentation. Traceability that is captured as the product is built and queryable in minutes is a capability. The configuration management discipline draws the same distinction, and for the same reason.

Related Concepts

• PLM and Supply Chain Integration: A Phased Implementation Guide, the four phases that build this
• What is Product Genealogy?, the unit-level record traceability extends
• MES vs PLM, where the as-built record originates
• QMS vs PLM, nonconformance, CAPA and the audit trail
• Discrete vs Process Manufacturing PLM, why batch traceability is a different data model
• What is PLM Supplier Integration?, the portal side of the supplier relationship
• Digital Thread Safety: What the 737 MAX Teaches Us
• Supply Chain Traceability, Lot Traceability and Approved Vendor List in the glossary

Frequently Asked Questions

What has to be traceable?

It depends on the product and the regulation, but the pattern is consistent. Raw materials by source and lot with their certificates. Purchased components by supplier, lot and test certificate. Subassemblies by manufacturer and configuration. The final unit by work order, station, operator and test results. Aerospace requires serialization of critical parts, medical devices require lot and UDI tracking, electronics must report conflict minerals, and the EU Digital Product Passport is extending part-level origin requirements to consumer products.

What is the difference between backward and forward traceability?

Backward traceability starts from a unit and asks where each of its parts came from: supplier, lot, certificate. Forward traceability starts from a supplier lot and asks which units consumed it. A recall needs forward traceability; a warranty investigation needs backward. Most systems can do backward from paperwork. Forward traceability, the where-used query across every serial number, only works if lot consumption was captured as each unit was built.

How does traceability relate to PLM?

PLM holds the product definition, the bill of materials and the approved vendor list that says which suppliers may supply each part at which revision. That is the as-designed side of traceability. MES captures the as-built side: which lot from which supplier actually went into serial number 4471. Traceability is the link between those two records. It breaks wherever supplier data lives outside PLM or as-built capture is skipped on the floor.

What happens when a supplier issue is discovered?

With traceability in place: query MES for every unit that consumed the affected lot, query PLM for the configurations and customers those serial numbers correspond to, and issue a targeted service action the same day. Without it: identify the affected population by date range and hope, or recall everything built that quarter. The difference in cost is measured in orders of magnitude, and it is the single strongest business case for capturing lot consumption at the point of manufacture.

Is traceability different in process manufacturing?

Yes, structurally. Discrete manufacturers trace serials and lots against an assembly structure, and the question is which physical parts went into which unit. Process manufacturers trace batches against a recipe, and the question is which ingredient lots went into which batch and where that batch was split, blended or packaged. Process traceability carries regulatory obligations such as FDA 21 CFR Part 11 and EU GMP Annex 11 that discrete serialization does not, and mainstream PLM platforms built for assembly handle it poorly without extensions.

Does blockchain solve supply chain traceability?

No. Blockchain provides an immutable shared ledger, which can be useful when several parties who do not trust each other need to agree on a record of custody, as in some conflict minerals and ethical sourcing programs. It does nothing about the actual problem, which is capturing lot consumption at the point of manufacture and holding supplier data in the product record. An immutable ledger of data nobody collected is not traceability.