# Lot, Serial or Item: Which Identity Level Does an Operational Use Case Need?

Source: https://activatedigital.ai/knowledge/product-data/lot-serial-item-operational-identity
Last verified: 4 September 2026
Summary: Choose lot, serial or item-level identity from the operational action and evidence needed, not from an assumption that every product needs serialisation.

## In 60 seconds

Choose the **lowest identity granularity that still lets the operation take the required action and support the evidence it needs**. A lot or batch is enough when every affected unit can be treated together. Item-level identity becomes useful when the action depends on the history, condition, custody or status of one exact unit.

- This is an operational design rule, not a universal standard requirement. DPP granularity remains product-rule specific.
- Persistent item-level identity is not universally required for circularity or every operational use case.
- Congress directed FDA not to enforce the Food Traceability Rule before **20 July 2028**. Its traceability model centres on Traceability Lot Codes.
- SBB reports 123 monitoring installations, about 11,000 trains recorded per day and RFID tagging on 50% of recorded vehicles, so measurements attach to the right vehicle.
- One RFID field study found inventory-record inaccuracy fell by about 26%. A second found strongly heterogeneous effects, including categories with no statistically significant improvement.
That is an operational design rule, not a universal standard requirement. It is also separate from the legal question of how many Digital Product Passports a product range must have. [DPP granularity remains product-rule specific](https://activatedigital.ai/knowledge/guides/granularity).

So persistent item identity isn't the default answer to every traceability, circularity or service problem.

## Start with the action, not the identifier

Teams often begin this decision with the technology: QR code, GS1 identifier, serial number, RFID tag. That reverses the problem.

Start with the action:

- Do you need to stop, quarantine or recall a defined group?
- Do you need to know the condition or history of one exact unit?
- Do you need to connect a measurement to one physical asset?
- Do you need to preserve custody or service state across a hand-off?
- Can the next decision be made from a model, part number, lot or measurement taken at the time?
Only then choose the identity level.

The same separation applies to product facts. A colour or composition fact may belong at model, variant or batch level while the operational event that matters happens to one individual unit. [Catalogue fact scope and operational identity level are different questions](https://activatedigital.ai/knowledge/product-data/model-variant-sku-facts).

GS1 identifiers and application identifiers can represent different data elements and levels of identity, but the identifier does not decide the business need. [The encoded data element has to match the operational job](https://activatedigital.ai/knowledge/fields/gs1-application-identifiers-dpp).

## A practical class, lot, serial and item decision tree

ActivateDigital's design principle: use the coarsest level that still preserves the action and evidence you need. This is architecture guidance, not a legal rule.

Work through the decision in this order.

### 1. Can every unit in the class be treated the same way?

If yes, class, model or product-type data may be enough. There is no operational benefit in distinguishing individual units if the action and evidence are identical for all of them.

If no, go finer.

### 2. Can a lot or batch isolate the affected population?

If one production run, shipment or traceability lot is the meaningful unit of action, lot identity may be sufficient. A recall or containment decision does not become better merely because every item has a unique serial if the operational action is identical for the whole lot.

If the lot still contains units that need different treatment, go finer.

### 3. Does the action depend on one exact unit's state, history or custody?

If yes, serial or item-level identity may be necessary. This is the point where one physical object needs to stay distinguishable from otherwise identical objects.

### 4. Can the relevant state be measured at the decision point instead?

Sometimes a live measurement is more useful than persistent identity. If the decision depends on current condition and the condition can be measured directly, the history of one item may add little.

That last branch matters because identity and information aren't the same thing. A unique number does not automatically create a useful decision.

## Where lot or batch identity is enough

Lot-level traceability is not a compromise waiting to be replaced by serialisation. It is a legitimate operating level where the action is defined by a shared production or traceability history.

The US FDA's Food Traceability Rule provides a useful example of the logic. Its traceability model centres on Traceability Lot Codes and records tied to Critical Tracking Events. FDA guidance says a Traceability Lot Code generally persists unless a qualifying transformation occurs, while businesses may also maintain separate internal lot codes.

That proves shared identifiers and internal identifiers can coexist, and that continuity rules matter. It does not prescribe one warehouse technology or EPCIS implementation.

The rule's current enforcement timetable also shows why legal adoption and operating maturity should not be collapsed. Congress directed FDA not to enforce the Food Traceability Rule before **20 July 2028**. FDA has separately documented unresolved distribution-centre concerns around mixed lots and accurate lot-level tracking. The lot model is real, but implementation difficulty is real too.

A similar principle applies to recalls. If a defect can be isolated reliably to a lot, the operational question is the lot. If individual units within that lot have different histories or statuses that change the action, the identity threshold moves lower. [The recall page owns the legal process and the need to identify the affected population precisely](https://activatedigital.ai/knowledge/guides/dpp-after-product-recall).

### When one material identity becomes several

A transformation can change the operational identity even when some parent material facts remain valid. A steel coil may be slit into narrower coils or cut into saleable pieces. The children can need new operational IDs while still retaining a controlled relationship to the parent coil and heat evidence. If that relationship is lost, the certificate has not become false, but its binding to the child has become unresolved.

Aluminium follows the same pattern through rolling, slitting, extrusion and downstream fabrication. A parent coil or billet can become several children, while a blended or remelted output can have several contributing parents. So the identity record needs to preserve the relevant parent-child relationship rather than assuming one persistent lot number will survive every operation. For the aluminium-specific distinction between nominal material identity and delivery evidence, see [alloy, temper and delivered-batch identity in aluminium](https://activatedigital.ai/knowledge/digital-product-passport/aluminium/alloy-temper-batch-identity).

This does not make item-level identity a universal DPP requirement. Use the level the operation needs, then preserve enough lineage to support the facts carried into the child record. The full method for deciding what can be inherited, what needs re-binding and what becomes UNKNOWN belongs with [how parent evidence should survive a split or transformation](https://activatedigital.ai/knowledge/product-data/material-lineage-after-cutting-splitting-transformation). If the problem is wider than one transformation and identifiers or evidence repeatedly stop joining across hand-offs, use the separate diagnosis of [why traceability chains break](https://activatedigital.ai/knowledge/guides/why-supply-chain-traceability-breaks).

## Where item identity becomes necessary

Item or serial identity becomes more defensible when the next action depends on one exact object's state.

US pharmaceutical traceability provides one regulated example. Under DSCSA, package-level verification is connected to suspect or illegitimate product investigation, quarantine and disposition. Persistent package identity supports an action against the specific package being investigated. That does not quantify time saved or prove every product category needs package-level serialisation.

Swiss Federal Railways provides a different production example. SBB says its trackside monitoring system uses RFID vehicle identification to assign measurements to specific vehicles. The page reports 123 monitoring installations, about 11,000 trains recorded per day and RFID tagging on 50% of recorded vehicles at the current snapshot. The operational reason for identity is clear: a measurement only becomes useful for maintenance when it can be attached to the right vehicle.

Deutsche Bahn provides a component-level example. Its technical delivery terms require globally unique serialised, machine-readable identification for contractually in-scope parts and components, with GS1-based marking recommended. This is a real supplier requirement where agreed in contract. It is not a rail-sector-wide mandate.

Retail evidence also shows that item-level visibility can improve a specific operational outcome without proving a universal optimum. A controlled field study by Hardgrave, Aloysius and Goyal found RFID-enabled visibility reduced inventory-record inaccuracy by about 26% in one study. A second study found strongly heterogeneous effects across categories, including categories with no statistically significant improvement. The useful conclusion is not "RFID improves accuracy by 26%". It is that finer identity can create value where the operation can use it, and the effect depends on context.

Item identity is also what makes reuse of one identity visible, because two units cannot honestly share a history. Where that matters commercially, ActivateDigital can [issue one identity per unit and watch how it behaves](https://activatedigital.ai/product-protection).

## Serial and item are not always the same design choice

For this guide, use **serial** for a unique instance identifier and **item-level identity** for the broader operational requirement that one physical unit remains individually distinguishable in the system.

In many implementations the two coincide. In others, the persistent key may be carried through a different identifier, device identity or asset record. What matters isn't the label on the identifier but whether the identity remains stable enough for the operation that depends on it.

So "give everything a serial" is weak architecture. If the serial is not captured at the point of action, not linked to the relevant state or not preserved across the hand-off that matters, the presence of a unique number solves very little.

## What legal DPP granularity still owns

Operational identity and legal DPP granularity can overlap, but neither should be used as a shortcut for the other.

For textiles, the final DPP granularity remains product-rule dependent. Separate pages cover [the model, variant, batch and item data structure for textiles](https://activatedigital.ai/knowledge/digital-product-passport/textiles/product-variant-batch-item-data) and [the unresolved legal granularity question](https://activatedigital.ai/knowledge/guides/granularity).

A business may need item identity for a service workflow even when the future legal DPP requirement is coarser. Equally, a legal passport requirement at a particular level does not mean every internal operation should use that same level.

## The two failure modes

### Over-serialising

Finer identity creates more identifiers, more state to preserve, more capture points and more exception handling. If the action is still taken at lot or type level, that extra complexity may not change the decision.

ActivateDigital's view is to treat that as a design cost, even where the technology itself is inexpensive. Do not create an item-level continuity obligation unless a real action or evidence requirement needs it.

### Under-granularity

The opposite failure is more obvious. If one unit's condition, custody or service history changes the next action, a lot identifier cannot tell you which unit is which. The system has lost the distinction the operation needs.

So the right question isn't "what is the most detailed identity we can create?" It's "what distinction must survive for the operation to work?"

## What is NOW, EMERGING and only POSSIBLE

| Maturity | What the evidence supports |
|---|---|
| **NOW** | Operations can run at class, lot/batch or instance level. Current standards and deployments demonstrate more than one useful granularity. |
| **EMERGING** | More item-level lifecycle and service applications where the economics, capture process and persistence justify them. |
| **POSSIBLE** | Wider item-level continuity across service and resale where identity and event capture survive across organisations. |
| **NOT_SUPPORTED** | Persistent item-level identity is not universally required for circularity or every operational use case. Sector examples demonstrate feasibility, not a mandatory optimum for all products. |

When item identity is justified, the next architecture question is what history should follow it. [EPCIS event data and product master data solve different parts of that problem](https://activatedigital.ai/knowledge/product-data/epcis-events-vs-product-master-data).

When the decision is specifically about repair, reuse, remanufacture or recycling, use the lifecycle threshold instead. [Ask which lifecycle state genuinely has to follow the individual product](https://activatedigital.ai/knowledge/guides/when-lifecycle-data-needs-item-level-identity).

And if the business case is resale, do not jump from persistent identity to a value claim. [Resale has several different frictions and identity only addresses some of them](https://activatedigital.ai/knowledge/guides/digital-product-identity-resale).

## What would change this page

This answer should be revisited if:

- product-specific DPP acts make a finer identity level legally mandatory for particular products;
- GS1 materially changes identifier guidance relevant to these operating patterns;
- new scaled implementations show that a currently item-level job can reliably be done at a coarser level, or the reverse;
- durable cross-company service and lifecycle workflows make item continuity materially more common.
**Sources as at 4 September 2026.**

## Keep exploring

The questions this page usually raises next.

- [CompareNext questionWhat Do EPCIS Events Add to Product Master Data?→](https://activatedigital.ai/knowledge/product-data/epcis-events-vs-product-master-data)
- [What to do nextNext questionWhen Does Lifecycle Data Need to Follow the Individual Product?→](https://activatedigital.ai/knowledge/guides/when-lifecycle-data-needs-item-level-identity)
- [What to do nextNext questionDoes Digital Product Identity Make Resale Easier?→](https://activatedigital.ai/knowledge/guides/digital-product-identity-resale)
- [What to do nextNext questionChoosing between a barcode and a digital twin→](https://activatedigital.ai/knowledge/guides/how-much-digital-state-product-needs)

## Sources

- [U.S. FDA, Title II of the Drug Quality and Security Act. Used for package-level verification, investigation, quarantine and disposition under DSCSA.](https://www.fda.gov/drugs/drug-supply-chain-security-act-dscsa/title-ii-drug-quality-and-security-act)
- [U.S. FDA, Frequently Asked Questions: FSMA Food Traceability Rule, updated through 9 June 2026. Used for Traceability Lot Code continuity and the distinction from internal lot codes.](https://www.fda.gov/food/food-safety-modernization-act-fsma/frequently-asked-questions-fsma-food-traceability-rule)
- [U.S. FDA, Discussion Paper: Identifying Additional Flexibilities for Satisfying the Food Traceability Rule’s Lot-Level Tracking Requirement, May 2026. Used for the documented mixed-lot and distribution-centre implementation concerns. The discussion paper is not binding guidance.](https://www.fda.gov/media/192696/download)
- [U.S. FDA, FSMA Final Rule on Requirements for Additional Traceability Records for Certain Foods. Used for the lot-level model and current non-enforcement-before-20-July-2028 position.](https://www.fda.gov/food/food-safety-modernization-act-fsma/fsma-final-rule-requirements-additional-traceability-records-certain-foods)
- [Swiss Federal Railways, Train monitoring and measuring systems. Used as a production example of identity-linked asset measurements. It does not establish cross-industry ROI.](https://company.sbb.ch/en/railway-development/services-rus/railway-environment/train-monitoring-systems/measuring-systems.html)
- [Deutsche Bahn, Supplementary contract terms for serialised marking of products to the GS1 application standard (EVB Serialisierung), valid from 1 January 2025. Used for a contractually bounded serialised-component implementation.](https://lieferanten.deutschebahn.com/resource/blob/13200236/337db97261ccde31ade89dcc6b3b941b/-EVB-Ergaenzende-Vertragsbedingungen-Serialisierung-Stand-01-01-2025--data.pdf)
- [Hardgrave, Aloysius and Goyal, RFID-Enabled Visibility and Retail Inventory Record Inaccuracy: Experiments in the Field, 2013. Used for controlled outcome evidence and its heterogeneous category effects.](https://doi.org/10.1111/poms.12010)