An auto component supplier near Chennai, two plants, around ₹160 crore, most of it going to two tier one customers. They passed their IATF surveillance audit and the finding that came out of it was the one I see most often: the traceability records and the production records did not agree.
Nobody had falsified anything. They had a good ERP running production and a separate, carefully maintained quality register in a spreadsheet, and over eighteen months the two had drifted the way two hand-maintained things always do.
What does IATF 16949 actually require for traceability?
The parallel register is the problem, not the solution
Almost every plant that struggles here started from a reasonable place. The ERP did not capture inspection results the way quality wanted, so quality built a spreadsheet. The spreadsheet was better at the quality job. It also became a second source of truth, updated by different people, at different times, with different definitions of what counts as a batch.
An auditor does not need to understand your process to find this. They pick a despatched part, ask for its record, then ask production for the same part's record, and compare. The gap is the finding. It takes them about ten minutes.
Four links that have to hold
Part to batch. The despatched part has to name the production batch that made it, and that name has to be the one production used rather than one assigned afterwards for the record.
Batch to material. The batch has to name the raw material lots consumed, with the heat number where the customer requires it. This is the link most often reconstructed rather than recorded, and reconstruction is visible.
Material to supplier. The lot has to name who supplied it and against which incoming inspection. If a supplier problem surfaces, this is the link that tells you how far it spread.
Batch to customer. Forwards from a suspect batch to every customer who received output from it, as a query rather than an afternoon's work. A containment exercise that takes two days has already failed.
Record inspection where production happens
The design principle that fixes this is unglamorous: the inspection result attaches to the batch inside the same system that ran the batch. Not exported to it, not reconciled with it nightly. The same record.
That has a side effect worth more than the audit outcome. First pass yield per operation, per shift, per machine stops being something somebody assembles the week before a customer review and becomes a number the plant already has. PPM by customer and part comes out of the same place. We go into the measurement side in improving OEE and cutting machine downtime.
"The spreadsheet was better than the ERP at the quality job. That was true, and it was also why we got the finding. Being better at your own job and being the second version of the truth are not incompatible."Auto component supplier, Chennai
Layered process audits and the document trail
The other requirement that catches people is showing your own layered process audits happened on schedule with actions closed out. Plants usually do the audits and lose the evidence, because the schedule lives in one place, the findings in another and the closure in somebody's email.
The fix is the same shape: one record, with the finding attached to the audit and the action attached to the finding. If you are preparing for a first certification, our auto component page sets out what we configure first for an automotive supplier.
Facto records inspection against the batch production ran, which makes the traceability record and the production record the same data by construction. If you have an audit coming and a parallel register you are nervous about, talk to our team.
