Guide · Traceability & recall
The 4-hour traceability test: mass balance, with a worked example.
The trace is the easy half of the annual traceability test. The half that fails is the mass balance, proving that everything you received or produced is accounted for, kilogram by kilogram, within a stated tolerance. This guide runs one complete worked example and shows where the kilograms usually go missing.
Home › The 4-Hour Traceability Test: Mass Balance With a Worked Example
Anatomy of the test
The annual traceability test has three components: a trace in one direction (varied year to year), a trace in the other direction, and a quantity reconciliation, the mass balance. BRCGS expects the whole exercise to complete within 4 hours of the test lot being nominated, and the test report to show the timings.
A useful discipline is to treat the test like an exam someone else sets: have a director or an office colleague nominate the lot, note the time, and hand it to the person who would actually run a real trace. If only one person in the business can complete the test, write that down as a finding, it is one, and auditors increasingly probe for it.
The mass balance equation
For a finished-product lot, the reconciliation is: quantity produced = quantity dispatched + quantity in stock + waste + samples + rework out. For a raw-material lot it inverts: quantity received = quantity used in production + stock remaining + waste + returns. Every term needs a document behind it, a dispatch note, a stock count, a waste log entry.
Your procedure should state a tolerance and justify it. Many producers work to ±1–2% for finished product; weight-variable products (wheels of cheese, joints of meat) often justify a wider band because catch weights are recorded at dispatch, not production. The number matters less than having stated it before the test rather than after.
Worked example: batch B-1042, blue cheese
Scenario: batch B-1042, made Wednesday week 12, is nominated at 09:00. Production records show 214 wheels at an average 1.42 kg, 303.9 kg produced. The trace forward finds the following:
Accounted: 302.3 kg against 303.9 kg produced, a 1.6 kg gap, 0.53%, inside a ±1% tolerance. Reconciliation closes; the report records 99.5% recovery, the timings (trace complete 09:41, mass balance closed 10:52), and the exercise passes with time to spare.
Now the honest version: on the first attempt most producers find a 5–8% gap. The missing kilograms are almost always in the same four places, unlogged grading waste, retention samples nobody wrote down, a farm-shop fridge that sells production overrun, and rework that went into a later batch without a record. The corrective action is rarely 'trace better'; it is 'log the leaks'.
| Destination | Evidence | Quantity (kg) |
|---|
| Irish multiple, private label | Dispatch notes DN-8811, DN-8837 | 145.2 |
| UK distributor | Dispatch note DN-8845 + export docs | 98.6 |
| Dublin cheesemonger | Dispatch note DN-8852 | 28.4 |
| Cold-store stock (counted) | Stock count sheet 14/wk15 | 22.7 |
| Grading waste | Waste log wk13 | 4.3 |
| Trade samples + QC retention | Sample log wk13–15 | 3.1 |
| Total accounted | , | 302.3 |
Where the kilograms usually go missing
- Rework without a paper trail — Offcuts and seconds folded into a later batch. Fix: a rework log naming source batch, destination batch and weight, two minutes at the time, hours saved in the test.
- Catch-weight drift — Production records nominal weights, dispatch records actual weights. Fix: state it in the procedure and set the tolerance accordingly, or record actuals at both ends.
- Samples, gifts and staff sales — Trade samples, auditor lunches, the staff fridge. Individually trivial, collectively a percent. Fix: one sample/other-use log.
- Waste logged by bin, not by batch — A weekly waste figure cannot be attributed to the test lot. Fix: waste entries carry a batch code.
- Farm gate and market sales — Direct sales dispatched with no note. They are still one-step-forward quantities. Fix: a simple daily direct-sales record with batch codes.
What goes in the test report
- Lot, scenario and direction — Who nominated it, when, and whether the trace ran forward, backward or both.
- Timings — Nomination time, trace-complete time, mass-balance-closed time. This is the 4-hour evidence.
- The reconciliation table — Exactly as above, every destination, every document reference, every quantity, the percentage and the tolerance.
- Findings and corrective actions — Gaps, owners, due dates. A clean sheet two years running invites the question of whether the test is honest.
Making the test boring
The whole exercise above is spreadsheet-achievable, and the discipline of the logs matters more than the tooling. What a connected system changes is the assembly time: in Keystone the forward and backward chain for any batch, with quantities per dispatch line and current stock, renders in under 30 seconds, so the 4-hour test becomes a 90-minute documentation exercise and the vertical audit becomes a live demonstration instead of a scramble. The waste, sample and rework logs still need humans, no software fixes an unlogged fridge.
Key takeaways
In one glance.
- The mass balance, not the trace, is where traceability tests fail, every term needs a document.
- State your tolerance before the test (±1–2% is common; justify more for catch-weight products).
- First-time gaps live in rework, samples, waste-by-bin and direct sales, log the leaks.
- Record nomination, trace-complete and balance-closed times: that is the 4-hour evidence.
- A findings section with corrective actions is stronger audit evidence than a suspiciously clean pass.
FAQs
Frequently asked.
What is a mass balance in food traceability?
A quantity reconciliation proving that everything produced (or received) is accounted for: dispatched + stock + waste + samples + rework must equal the produced or received quantity, within a stated tolerance, each term evidenced by a record.
What tolerance is acceptable for a mass balance?
The standard does not prescribe one, your procedure defines and justifies it. ±1–2% is common for fixed-weight finished product; catch-weight products often justify a wider band. What fails audits is an unexplained gap or a tolerance invented after the fact.
How long should the traceability test take?
BRCGS expects full traceability including the mass balance within 4 hours of lot nomination. Record the timings in the report, the times are the evidence.
Should the test trace forward or backward?
Both, across the audit cycle. A common pattern is finished product backward one year and raw material (or packaging) forward the next, so both directions carry recent evidence.
Who should run the traceability test?
Ideally someone other than the person who built the records, from a lot nominated by someone else. If the test only works with one specific person driving, that is itself a finding worth recording and fixing.