An effectiveness check is how you prove a corrective action worked. Decide before the fix goes in what evidence you will accept and how long you will watch, then collect it from your own data: complaints, defect rates, audit results and process measurements. If the evidence meets the criteria, close the action. If it does not, reopen it and go back to the cause. The actions being finished is not proof. A closed action means the problem stayed gone.
Done is not closed
Most corrective actions get closed the day the last task is finished. That tells you the team did what they said they would. It does not tell you the problem is gone. The gap between those two is where repeat complaints come from. An effectiveness check is the step that closes the gap.
Two different checks
- Implementation. Did we do what we said we would? The dividers are in, the instruction is updated, the people are trained.
- Effectiveness. Did it work? Has the problem stopped? This takes time and data.
You need both. Many plants stop at the first and call it verification.
Set the criteria before the fix ships
If you decide what counts as proof after the fact, you will find proof. Write it down at the time the action is planned: what you will measure, where the data comes from, what number means success, and how long you will watch. “No further complaints” is a start, but “no scratch complaints or tote damage found in 12 weekly pack-out audits and the next three months of shipments” is a criterion.
What counts as evidence
- Defect, scrap and complaint data. The same numbers that showed the problem should show it gone.
- Audit results. A layered process audit question that checks the new control, passing over and over.
- Process data. A control chart or a capability study showing the process is stable and capable.
- Direct observation. Go and see the new control working on the floor, on different shifts.
- Customer feedback. Confirmation from the customer that parts have been clean since the change.
Records and data beat statements. “The team says it’s fixed” is not evidence.
How long to watch
Long enough that the problem would have come back if the fix did not hold. If it showed up once a month, one clean week proves nothing. A rule of thumb: watch at least several cycles of however often the problem appeared, and longer for safety or special characteristics. Set the period in the criteria, not on the day you feel ready.
How to run it
Write the criteria into the action plan
What, where from, the target, the period, and who checks.
Confirm the actions are in place
Check each one on the floor and in the documents. Record the date.
Collect the evidence over the period
Put review dates on the calendar. Do not rely on remembering.
Have someone other than the owner check
The person who did the work should not be the only one who judges it. A second pair of eyes catches what the owner has stopped seeing.
Compare the results to the criteria
Yes or no, with numbers. Not “mostly.”
Close it or reopen it
If it met the criteria, close it, with the evidence attached. If it did not, reopen it.
Record the outcome
Where the evidence lives, who checked and when. An auditor and your future self will both want it.
When it did not work
That is information, not failure. Go back to the cause: was the root cause wrong, was only one leg of the 5 Why fixed, or did the action never fully reach the floor? Asking people to try harder or retraining them again rarely changes anything. Change how the work is done, then check again.
Mistakes that make the check meaningless
- Closing on completion. The actions finishing is a milestone, not proof.
- No criteria set beforehand. Without a target, any result looks fine.
- Watching too briefly. A short clean spell proves nothing about a rare problem.
- The owner checking their own fix. It is hard to see your own blind spots.
- No route to reopen. If failing the check has no consequence, it is just paperwork.
- Using the wrong data. If you improved packing but measure machining scrap, you will see nothing.
Field Book 1, chapter 8 follows the scratched housing case through its check. The Effectiveness Verification Tracker keeps the criteria, the review dates and the decision in one place, and the CAR Tracker holds the action itself.
Questions people ask
What is the difference between verification and validation of a corrective action?
Verification confirms the action was carried out as planned. Effectiveness checking, sometimes called validation, confirms it solved the problem. People use these terms loosely, so say which you mean.
How soon should I check effectiveness?
Set it in the action plan, based on how often the problem appeared. Review at points along the way, and make the final decision after the full period.
Who should do the effectiveness check?
Someone other than the action owner, such as quality or an auditor, so the person judging is not judging their own work.
What if the problem comes back during the check?
Reopen the action. Look again at the root cause and whether every leg was addressed. Do not wait for the period to end.
Does an 8D need an effectiveness check?
Yes. The step that verifies the permanent corrective action works is part of a complete 8D, and customers often ask to see the evidence.