Tricksereien mit Maßnahmen in der FMEA

A blog by Dr. Uwe-Klaus Jarosch, October 2025

Have you read the blog about actions in the FMEA?
It is about  types of actions and how to use language for adequate phrasing .

Here and today we will talk about ways to make actions effective.

Targets of the FMEA method is on one hand to properly understand the product or the process in development ( as a result of structure-,  function-, and failure analysis).
On the other hand we target to actively prevent potential week points, potential failures or at least to detect them.
In this blog I would like to focus on the second issue. It is about how to make actions not only paperwork or text line in a FMEA but real activitiy.
Only with the realization of the actions the risks will be mitigated.

Following the frequently used FMEA standards we need to go 2 steps:
First step: we collect what has happened until now or what what will be done for sure. We talk about the actions in Initial State.

The initial state will be rated: How good is the potential failure prevented by this set of prevention actions?
How safe will this failure be detected by the detection actions given in this initial state?

The second step will be necessary as soon as there is a change in development and the initial state is no longer sufficiently safe.

Or the rating oft he initital state action package is showing not to be sufficient.

Then we talk about Optimization.

Typically, we need additional actions for the optimization. And – in difference to the actions in initial state – these actions will be tracked.

Action tracking means: There is a responsible person and a due date. The action is tracked by its progress until it is either completed or refused.

As long as an action is not closed it will be tracked. This may become somewhat unpleasant for the responsible.

Cheating with action tracking

For this reason, three tricks are often used:

  • As much as possible is entered—if necessary, retrospectively in the initial status—and not tracked.
  • Even with new optimization measures, the status “Completed” is entered immediately.
  • A third trick is to enter measures in the FMEA, but then not actually track them.

Do not allow cheeting

At this point, for me it becomes obvious if a company is taking the FMEA method serious or better allows a placebo session.

When risks have been detected and have been seriously = correctly rated then ONLY actions are able to mitigate the risks and to design a safe product or process.

Long time ago I have learned:  “The assumption is the start of the catastrophe”

This is true as well in context of the FMEA methodology.

  • I place a not yet finished action in the initial state and hope it will be successfully closed.
  • I do the statement: The action is closed. But I do not have any evidence.

It seems to be nasty, petty-minded, like addition workload. But it is a neccessary step to transfer an assumption into a fact.

The list of proposals below will turn the FMEA into the evidence sheet not to let the risks occur:

    • Only list such actions in the initial state which are evidently closed with success. Add the proof in the action.
    • You may write an action into the initial state that will be performed by an independent department, reported by neutral experts, using a safe and approved procedure and ensuring to inform you in case of nonconformity.
    • Else, write ALL other actions into a optimization action state (dated state).
    ALL actions for optimzation firstly need a decision tob e done *.

    * Accordig to AIAG-VDA FMEA Manual (2019) there is an additional action status „in decision“. This status makes clear if an action is currently a proposal but not yet decided to be done. After the action is decided to be done it will turn into a „conventional“ action with status 0 – 100% or status „refused“.
  • ALL actions get a responsible person and a due date. Only persons may be nominated that are active participants in the FMEA sessions or nominated by the direct superior. Otherwise, a core team member is acting as care taker for this action.
    It is not reasonable to nominate a „role“ instead of a true person. In best case, your tool does not allow the nomination of a role only. Only true persons will track their actions.
  • All actions will be checked fort he current status frequently. It will not work if there are many open actions in status 0% until short before the due date and suddenly they will be finished.
  • Changing the action status to 100% shall request an evidence. Evidence can be all clear documents and expert reports, e.g. drawings incl. drawing no and revisio, DV results, lab reports by third party, MSA or similar.

Conclusions: 

This will take effort. Just to follow up numerous actions in detail is workload.
But without this approvals the rating of occurrence and detection is only hopes and wishes, wishes and hopes.
This would be too little to derive a reliable risk assessment.

If you decide to do a FMEA then make it properly.

Stay curious

Yours 
Uwe Jarosch

Leave a Reply

Your email address will not be published. Required fields are marked *