Chapter 30: Why 'We Fixed It' Is Never Enough: What Effectiveness Verification Actually Means

ISO 9001 doesn't ask if you fixed something. It asks if your fix worked. Clause 10.2.1(e) is specific: you must "review the effectiveness of the corrective action taken." That word—*review*—implies assessment. It implies measurement. It implies evidence gathered over time, not a signature on a form.

Here's where the confusion starts for most teams: you complete a corrective action request and document the action taken. Maybe you replaced a worn tooling insert that was causing dimensional drift. Maybe you retrained your inspection staff on a caliper reading procedure. Maybe you implemented new material receiving criteria to prevent contamination. Then you declare the CAR closed because the action is done.

But ISO 9001 wants to know if the worn tool was the actual root cause, or if you got lucky and the problem just hasn't come back yet. It wants proof that retraining stuck. It wants to see that the new material criteria are actually catching defects before they reach the floor.

The distinction between output verification and outcome verification is where most auditors draw the line between a pass and a major finding. Output verification answers the question: "Did we implement the action we said we would?" A signed-off work order for tooling replacement is output verification. A new procedure document is output verification. A training attendance roster is output verification. These are necessary but not sufficient.

Outcome verification answers a harder question: "Did implementing this action eliminate the root cause and prevent recurrence?" That requires measurement. That requires follow-up. That requires you to actively monitor whether the nonconformance is gone and stays gone.

Consider a real example from a Mississauga automotive supplier. They were producing a fixture component with dimensional creep—tighter tolerances were drifting out of spec on the 340th and 341st piece in a 500-piece run. The team's root cause analysis identified worn tooling. They replaced the tool and ran a pilot lot of 100 pieces. All 100 were in spec. CAR closed.

Three weeks later, the same dimension started drifting on production lot number 47. Major finding in their next audit. The root cause wasn't the tool alone—it was the tool *plus* a cooling system pressure that had dropped by 2 bar, which they hadn't caught because they only measured the tool wear in isolation.

What they should have done: Monitor the dimensional drift metric across the next 10 production runs (500 parts minimum), establish statistical control, and document that Cpk ≥ 1.33 was sustained. That's outcome verification. That's proof.