Failure modes and effect analysis (FMEA) is a tool that works to prevent process and product problems before they occur.
I like to define FMEA’s as an organized brainstorm. The process examines a product or process and asks what could go wrong. Then the team systematically determines and rank orders for each failure mode:
– the severity of the problem when it occurs
– the probability of the problem occurring
– the ability to detect the problem before it occurs
Good design engineers think about how the design could fail and improve the design. FMEA provides a structured team approach to further improve the design.
The cost of an FMEA is generally easy to determine. It is the cost of the talent on the team and any specific experiments or expenses related to the FMEA study.
The cost of changing a design based on the study is not included, assuming the design team would be already working to improve the product. The FMEA results provide direction for improvement work, thus focusing the effort and not adding to the effort.
Let’s say the team is five people and they spend a day working on the FMEA study. Further, let’s say each engineering has a loaded cost to the organization of $1,000 / day. Thus the engineering time cost of the study is $5,000.
If the team does a few experiments or has specific expenses that would add to the cost. In this simple example, let’s say they used one prototype in a short experiment to answer a few outstanding questions. To make it simple, assume the additional cost is $5,000.
A total cost of $10,000.
So, what do we get for the investment of time and resources in a simple design FMEA?
The return on investment for an FMEA study largely depends on when it is done and how well the team implements the action items (the outcomes of the study).
It also depends on the existing knowledge about product or process problems. If we already have a long list of issues and an easy way to prioritize them, say by analyzing field returns for an existing product, then FMEA may not add any new information.
If on the other hand, the product is using a new design approach or the process is manipulating a new material, then we may not have a well-crafted list of issues.
FMEA is a great tool when there are unknowns or uncertainties. It is also commonly used by teams to organize and manage a large number of risks.
It is not appropriate to expect any value when the study is always done despite the existing knowledge. If the study only lists what we already know, it has little value.
When the tool is applied in a situation that is early in the design process of a new product or process, with new design approach/architecture or materials, and the risks of failure are either unknown or uncertain, then FMEA may provide significant value.
In my experience, I’ve seen FMEA’s create value in three basic ways.
- Identifying and removing faults
- Prioritizing design improvement work
- Cross department coordination
Let’s explore each one briefly.
Identifying and removing faults
This is obvious. Let’s say the team already knew about 10 problems that needed addressing in the design. Then the FMEA added an additional 1 issue to resolve.
An estimate of the impact of that issue if unresolved is part of the prioritization process in an FMEA thus provides a rough estimate of the percentage of products that would have failed in the hands of customers.
The number of failures times the cost per failure provides the avoided cost.
For a simple example, let’s assume we plan on selling 10,000 units of this design. Each failure costs our organization $500. And, in the FMEA we estimated the problem would cause about 1% of units to fail.
1% of 10,000 is 100 failures. At a cost of $500 each, that would be a cost of $50,000.
Normally, a well-done FMEA will find many previously unknown problems. The issues may impact most or very few products, yet the basic logic applies. Avoiding field failures avoids the cost of failures from occurring.
In this example, I’m only counting direct costs, such as warranty replacement or repair costs. In reality, the failure cost may include brand image or customer satisfaction reduction.
Prioritizing design improvement work
Even simple designs may have many hundreds of potential failure modes.
Solving the problems take time and resources and solving the wrong problems adds little value. Focusing on the most severe and likely to occur issues tends to provide the most value.
I’ve worked with hundreds of design engineers and have asked many of them what could go wrong with the current design. Each easily can provide 3 to 5 issues that should be addressed before finalizing the design.
A problem that FMEA helps to solve is a design team of five people all working on the same product will easily create a list of number one priorities that most likely will not overlap.
The design team generally cannot have 5 top priorities.
The larger the team and the more complex the product or process, the longer the list of problems to solve. Without some way to organize the priorities, the team easily could be working aimlessly to solve individually determined top priorities.
This generally is not an efficient use of design talent.
FMEA is not the only way to provide order, yet it does provide a well thought out and reasonable way to focus the design team on solving the top priority problems.
What should be prioritized?
In my experience, most designers quickly consider the problems that will affect all products or have dangerous failure modes. The balance and priorities become more difficult when the chance and severity are not significant.
The FMEA study lets the team articulate the priorities clearly.
The value is found by shifting focus to the most important issues.
Let’s say a team identifies 100 issues that could improve the design. The program has the resouces to solve 20 issues before product launch.
Which 20 do you focus on?
Statistics is your friend
Using the basic concept of 80% of field issues are caused by 20% of the design flaws, if we select the right 20 issues we can avoid the majority of the field issues.
If we select randomly, we may only address 50% of the field issues. This is not likely given simple engineering judgement will most likely indentify and solve the major issues.
So let’s say it’s in the margin, that the last four issues identified to solve in the FMEA study would have been randomly selected otherwise. The difference is avoided field failures depends on the difference in expected impact to the cost of failure, yet even a small difference provides enough return on the FMEA investment to provide the priortization to the team.
Of course the prioritization creates a list with diminishing return on the effort to solve it prior to shipping. Yet, the net effect adds value.
Cross department coordination
In most product or process design work, the complexity of the system requires teams focusing on mechanical, electrical and software elements of the design.
One of the dangers that each large team faces is the competition for scarce resources. If you’ve ever been part of a prototype allocation discussion you know what I mean.
Each team may fully understand the risks that may cause field failures, yet they often lack the understanding from the other teams. Each team may view the design challenges they face as the most important.
One benefit that is difficult to quantify is the ability of the study to establish cross team discussions about priorities for the available resources.
Reducing the cross team competition frees engineering and management resources to focus on and solve the most important problems. It makes for a better work environment, which is a real value in itself.
While difficult to quantify, it has been a regularly observed benefit in my experience.
FMEA is a tool.
Used wisely on projects that may benefit from the systematic identification and prioritization of potential failure modes, it will provide significant value.