Lesson 6 of 6
Traceability testing for recalls
10 min
Listen to this lesson
Traceability is a promise: when a supplier lot or a production run is suspect, Contoso can name every affected unit and every customer quickly, and prove that everything else is unaffected. A promise like that must be tested the way it will be used, with a recall drill on realistic data. This lesson designs the drill for Contoso's rotor-to-customer trace and lists the evidence that quality and the process owners sign.
Two directions
| Direction | Question | Start point | Contoso example |
|---|---|---|---|
| Bottom-up (forward) | Where did this lot go? | A supplier or component batch | Lakeside lot LKS-26-311 → rotor batch 0000001042 → production orders → alternator serials → deliveries → customers |
| Top-down (backward) | What went into this unit? | A finished batch or serial number | Alternator serial reported by Great Lakes → production order → rotor batches consumed → Lakeside lot |
In SAP S/4HANA, the classic tool for batch traceability is the batch where-used list: MB56 analyzes it, and the Batch Information Cockpit (BMBC) offers top-down and bottom-up views together with stock and batch data. The Manage Batches app shows a batch's history. For serialized units, the serial history (IQ03) links units to production orders and deliveries. SAP also sells a separate product, SAP Global Batch Traceability, for tracing across several systems; Contoso's single-system trace does not need it.
Preconditions for a trace that works
- Every goods movement of a batch-managed material carries the batch: receipts, transfers (311 from 171S to 171L), issues to orders and receipts from orders.
- Batch where-used data is written for those movements. After batch management is first activated, or after data migration, check whether the where-used list must be compiled (MB57) so that history is complete.
- Production orders record which component batches they consumed, including splits.
- Serial numbers are assigned at every step of the alternator flow, including EWM and returns.
- Material documents and serial history are not archived inside the period recalls may need.
The recall drill
Wei's test data recipe creates a realistic history in CMQ: three Lakeside lots received into 171S and moved to 171L, five production orders with batch splits, goods receipts with serial numbers, putaway in EWM, deliveries to Great Lakes and a stock transfer to Toledo, plus one customer return. Then Priya hands the drill team only the suspect lot number and starts the clock. The team must produce the list of affected serial numbers, their location (in stock, in transit, at which customer), block what can be blocked, and prove the trace top-down on a sample.
| Drill check | Expected result | Evidence |
|---|---|---|
| Bottom-up trace of lot LKS-26-311 | Exactly the production orders that consumed batch 0000001042 | MB56 or BMBC output |
| Units affected | All serials from those orders; none from others | Serial lists per order |
| Customer list | Great Lakes ship-tos and the Toledo transfer, with quantities | Deliveries with serials |
| Stock containment | Affected stock in 171A and 171B set to blocked or restricted | Material documents or batch status |
| Top-down sample | 10 units traced back to the lot correctly | Serial history and order component list |
| Unaffected proof | Orders without the lot shown as not containing it | Where-used for the other lots |
Recall drill checks and evidence.
Package the drill like a real recall file: the trigger (the suspect lot and why), the trace outputs, the list of affected serials with locations and customers, the containment postings, the top-down confirmation and the time log. Add a draft customer notice that names exact serial numbers, because writing the notice shows whether the data is good enough to send. The same file structure is then used in a real event, so nobody invents the format under pressure.
What goes wrong in testing
| Pitfall | Consequence | Prevention |
|---|---|---|
| Drill on hand-picked clean data | Trace looks perfect; real gaps hidden | Seed realistic history with splits, transfers and returns |
| Where-used list not compiled after activation | Older movements missing from the trace | Check completeness against material documents |
| Repetitive manufacturing noise | Where-used shows many irrelevant batches | Know the limits of the tool for your production type |
| No time target | A trace that takes two days passes as "working" | Agree a target time with quality and measure it |
Typical traceability test failures.
What would you do?
The recall drill produces a correct list of affected serial numbers, but it took 11 hours because the team exported material documents to spreadsheets and matched them by hand. The target is 4 hours. What should Priya record?
- Test both directions: bottom-up (where did the lot go?) and top-down (what went into this unit?).
- MB56 and BMBC analyze batch where-used; IQ03 shows serial history; Manage Batches shows batch history.
- Every movement must carry batches and serials; check where-used completeness after activation.
- Run a timed recall drill on realistic seeded data and repeat it after changes.
Check your understanding · not graded
Lesson check
1. Which direction answers "where did this supplier lot go?"
2. Which transaction analyzes the batch where-used list?
3. What should be checked after batch management is first activated?
4. Why does a recall drill need a time target?