FMEA that prevents failures, not paperwork that fills a binder

A Failure Mode and Effects Analysis exists to answer one question before production starts: what could go wrong here, how serious would it be, and what are we doing about it now. A design FMEA studies the product. A process FMEA studies how the product gets made. Both rank the ways a process can fail so a team spends its attention on the few risks that matter most. Used that way, it is one of the strongest prevention tools in quality. Filled in the wrong way, it becomes a form nobody reads.

Built early, or not really built

The most common problem I see is timing. Many FMEAs are written after the process is already running, to satisfy a customer or an auditor. By then the document only records decisions the team already made, so it cannot shape anything. A useful FMEA is built while the design and the process are still soft enough to change, so the analysis can actually steer them.

Severity, occurrence, detection

Every failure mode gets three scores. Severity asks how serious the effect is on the customer. Occurrence asks how often the cause is likely to happen. Detection asks how likely your current controls are to catch the failure before it escapes. For years teams multiplied the three into a Risk Priority Number and chased the highest totals.

The newer AIAG-VDA approach replaced that single number with Action Priority, which sorts work into high, medium, and low based on the combination of the three scores rather than their product. The change matters. A high-severity failure can hide behind a modest RPN, and a part that fails a safety requirement deserves attention even when it rarely happens and is usually caught.

Where FMEAs go wrong

A few patterns repeat:

• It is treated as a document instead of an analysis, filled in quickly and never opened again.
• Scores are chosen to land on a comfortable number rather than an honest one.
• Recommended actions are vague, such as monitor the process, instead of a specific change with an owner and a date.
• The links break. Risks found in the FMEA never reach the control plan, so the floor never hears about them.

Make it live

An FMEA earns its place when it drives real change. Its actions should produce design changes, mistake-proofing, or new controls, and those controls should appear in the control plan and on the floor. When a new failure shows up in production or in a return, it goes back into the FMEA so the next program starts smarter.

Some customers and auditors now ask for a review of effectiveness, which simply means showing that the actions you took actually lowered the risk. That is a healthy discipline. An FMEA that never changes after its first draft is usually a sign it was never really used. Build it early, score it honestly, connect it to the control plan, and revisit it whenever reality teaches you something new.