In product traceability, a pallet label tells you which pallet you are looking at. It does not tell you what is on it. That knowledge comes from a separate record created when the pallet was built, and the name for that record is aggregation. This article explains what it is, what it contains, and why it is only as good as the moment it was captured. For where aggregation sits among the other building blocks, see what is product traceability?
The plain-English definition
Aggregation is the recorded relationship between a container and the items packed inside it. In GS1 language the container is the parent and the items are its children.
- Case 395011010000004574 contains bottles A7K29Q, A7K29S and ten others.
- Pallet 395011010000012012 contains 48 cases, including case 395011010000004574.
The parent is normally identified by an SSCC, the 18-digit number for a logistic unit; see what is an SSCC? The children are whatever sits one level down: serialized units inside a case, or cases on a pallet.
A useful analogy is a set of nested address books. The pallet's book lists its cases. Each case's book lists its bottles. Look up the pallet, follow the entries, and you can find every bottle without unwrapping anything.
What an aggregation record contains
A useful aggregation record answers more than "what is inside":
| Element | Example | Why it matters |
|---|---|---|
| Parent identity | Case SSCC 395011010000004574 | The container being described |
| Child identities | 12 unit identities (GTIN + serial) | What it contains |
| When | 2026-09-14, 10:42 | Separates this packing from any later repack |
| Where | Packing line 3, factory GLN | Supports investigations and audits |
| What happened | Packing (added) or unpacking (removed) | Lets the hierarchy change over time |
Those last three columns are what turn a simple list into traceability evidence. In EPCIS, the GS1 standard for sharing supply-chain events, this kind of record is an AggregationEvent, and the CBV (Core Business Vocabulary) supplies the business step "packing". The event view is explained in what is an aggregation event?
Building a hierarchy
Chain the records together and you get a tree you can query in two directions:
- Downwards ("what is inside?"). A distributor scans pallet 395011010000012012 and needs to know which batches and serials it holds, for stock records or a recall check.
- Upwards ("where is it?"). A consumer reports a suspicious bottle with serial A7K29Q. The brand looks up which case it was packed in, which pallet that case went on, and which shipment carried the pallet.
Both directions only work if every level was recorded. A missing case-to-pallet link breaks the chain for every bottle in that case. The full packing sequence is described in how individual products become cartons and pallets.
How aggregation is captured
There are two broad ways a line knows which children went into a parent.
Read at packing. A camera or scanner reads every child's code as it enters the case, or reads all of them through the open case before it is sealed. The record reflects what was physically seen. This is the stronger method.
Inferred from tracking. The line software knows the order of units on the conveyor after the inline camera, and assumes the next twelve went into the next case. This works when nothing is removed by hand and every reject is physically confirmed. It fails silently when an operator picks a unit off the belt to look at it.
Many lines combine the two: tracking to anticipate, reading to confirm. Either way, only units that passed inline verification should be eligible to be packed, which is why aggregation depends on the factory steps covered in how cameras and barcode readers verify product identity.
When aggregation goes wrong
The common failures are mundane:
- Phantom children. A rejected or sampled unit is still listed in a case it never entered.
- Orphans. A unit was packed but never read, so it has no parent and cannot be traced upwards.
- Stale parents. A pallet was split, but its old record still lists all 48 cases.
- Double parents. A case appears on two pallets because a rebuild was recorded as an addition without the matching removal.
- Duplicate messages. The same packing record is sent twice to a partner after a network error, and counted twice.
Most of these are prevented by the same disciplines that make serialization reliable: one-way states, durable local records, and reconciliation of counts at the end of a batch. They are explained in exactly-once serialization in plain English.
Who uses aggregation data
- Manufacturers use it for recalls by batch, for investigating complaints, and to show which shipment a unit left in.
- Distributors and retailers use it to receive and dispatch whole cases and pallets by scanning one label; see why warehouses scan cartons and pallets.
- Brand protection teams use it to tell whether a unit found in one market was shipped to another.
- Regulators, in sectors where rules require it, may expect the hierarchy to be available on request.
The data stays with the companies that create it and is shared with partners as needed. GS1 sets the standards for how it is expressed; it does not store it.