Every audit I have ever conducted includes the same ritual. The quality manager opens the training matrix — a spreadsheet listing every employee, every procedure they need to be trained on, and the date the training was completed. Green cells mean trained. Red cells mean not yet trained. The auditor checks a few records, confirms that training was delivered, and moves on. Finding closed. Requirement met.
Except it is not met. Because training records and competence are not the same thing, and the fact that we treat them as interchangeable is one of the most persistent failures in quality management.
The attendance fallacy
Training is the input. Competence is the output. We measure the input because it is easy to document — someone attended a session, a signature was collected, a record was filed. We do not measure the output because it is hard. Competence is the demonstrated ability to apply knowledge and skill to achieve a defined result. It requires evaluation, not just delivery.
I have seen operators with perfect training records who could not perform the task they were trained on. I have seen auditors who were trained on a standard but could not interpret its requirements. I have seen quality engineers with certificates from prestigious training organisations who could not run a capability study. The records said they were competent. The reality said they were present.
A training record proves attendance. Only performance proves competence. The gap between these two is where quality failures live.
What competence actually looks like
When I set up the quality system for a 900-employee greenfield plant, I designed a competence model that had nothing to do with training hours or attendance sheets. For every role, we defined the specific tasks that the person needed to perform and the standard of performance they needed to achieve. Then we evaluated each person against those standards.
For an operator on a press line, competence meant: can you set up the tooling, run a first-piece check, interpret the SPC chart, and respond correctly to an out-of-control signal? Not in a classroom. On the line, under production conditions, observed by a qualified evaluator.
For a quality engineer, competence meant: can you run a 5-why analysis that reaches a systemic root cause, write a corrective action that addresses that cause, and present it to a customer in a way that inspires confidence? Not on a test. On a real nonconformance, with real stakes.
The evaluation was documented. The standard was clear. And the result was a competence assessment that actually meant something.
Why companies use training records instead
Training records are easy. Competence assessment is hard. Training records can be generated by scheduling a session and collecting signatures. Competence assessment requires defined standards, qualified evaluators, observation time, and the willingness to tell someone they are not yet competent. That last part is the killer. Most managers do not want to look an operator in the eye and say "you have been here for three months and you are not yet qualified to run this machine independently." It is uncomfortable. It requires a conversation about performance that most managers would rather avoid.
So the training record becomes the proxy. The signature becomes the proof. And the audit finding becomes impossible to write, because the organisation has met the letter of the requirement while completely missing its intent.
The cost of pretending
I investigated a serious nonconformance at a plant where a critical sealing operation had been performed incorrectly for three months. The operator had been trained. The training record was on file. The procedure was posted at the workstation. And the operator had been doing it wrong since day one.
The training had been a two-hour classroom session covering fourteen procedures. The operator could not possibly have absorbed all fourteen procedures in two hours. There was no practical demonstration. There was no observation period. There was no competence verification. There was a signature on a sheet that said "trained."
The cost of that signature: 18,000 parts recalled, €220,000 in sorting and rework, a customer line-down charge of €45,000, and an 8D that took six weeks to close. The cost of a proper competence assessment: approximately four hours of an evaluator's time.
What I recommend
Stop equating training with competence. Keep your training records — they satisfy the auditor and they document the inputs. But add a layer that actually matters:
Define competence standards for critical tasks. Not every task needs a formal competence assessment. Focus on the tasks where errors have the highest consequence: safety-critical operations, inspection of critical characteristics, activities that cannot be verified downstream.
Evaluate practically. Observe the person performing the task under real conditions. Use a checklist. Document the result. If they are not competent, say so — and provide the coaching they need to get there.
Re-evaluate periodically. Competence is not permanent. People forget. Processes change. New equipment arrives. Annual re-evaluation of critical tasks catches the competence drift that training records never will.
ISO 9001:2026 will almost certainly strengthen the competence requirements, moving further toward demonstrated ability and away from documented attendance. The companies that wait for the standard to force this change will be playing catch-up. The companies that adopt it now will have a quality system that actually prevents defects instead of one that just documents training.
