FMEAs are like diets. There is the ‘5-bite diet’ (no need to explain this one too much) that is so simple to understand, easy to initially say ‘yes’ to, but demonstrably impossible to maintain (bordering on dangerous). Then there is the ‘werewolf diet,’ which appeals to our inner hunter-gatherer that we think we still are, but beyond that, it makes no nutritional sense. There are diets that work well for one person and not well for another. And finally, there is usually (barring medical conditions) a perfectly feasible, sensible, sustainable diet that can work for you if it is tailored to your needs, physiology, and goals.
The reality is that we cannot be truly healthy without a ‘perfect’ diet that works for us. There could be more than one ‘perfect’ diet. There are plenty of ‘mediocre’ diets. And lots of ‘terrible’ diets.
If you are with me so far … welcome to the world of FMEAs!
What is a Failure Modes and Effects Analysis or FMEA?
You can find lots of wrong definitions about anything by jumping onto the internet. For example, an artificial intelligence (AI) tool might say that a ‘diet’ is something like …
… the total sum of all foods and drinks a person or animal habitually consumes …
This is not wrong. It is also not helpful. The reality is that when we tell others (or ourselves) that we are going on a diet, what we actually mean is that we are coming up with a food and drink strategy to achieve a specific outcome. It could be to lose fat, gain muscle, improve energy, or some other tangible end state.
When it comes to food and drink, there are numerous well-established principles and universal guidelines that help us create a balanced diet. However, if coming up with a diet were straightforward, then all we would need to do is go online and download ‘the one and only diet,’ and simply follow it without too much discomfort, confusion, or frustration. But we all know that there is no such diet, which is why there are hundreds of thousands of highly paid nutritionists worldwide, individually helping clients to work towards their own attainable goals.
Now, let’s stop talking about diets and start talking about FMEAs. The same AI tool we asked to come up with a definition for a diet might say the following when we ask it to do the same for a FMEA:
… a FMEA is a systematic approach to identify potential failures in a system, process, or product, and to analyze their potential effects …
Again, this is not wrong. And again, it is also not helpful, for the same reason that the definition of ‘diet’ above is not helpful. Good FMEAs are those that achieve a specific outcome. And that outcome is always to improve the product, system or process they are focusing on. Analyzing problems is very different from solving problems.
And this is one of the many reasons FMEAs have a ‘PR problem.’
Snake Oils
‘Snake oil’ was routinely sold in the then-unregulated Middle America of the 1800s and early 1900s. The first generations of the product were genuine rattlesnake fluids (of some sort) but were soon replaced without much fanfare with easier-to-extract mineral oils. Salesmen masquerading as doctors argued that adding these fluids to diets would cure everything from rheumatism to palsy. Perhaps reassuringly, no claims were made regarding any curative powers as they pertained to gullibility. The salesmen had a tried-and-true approach. They arrive at a town, plant an accomplice in any gathering crowd to attest to the amazing powers of ‘snake oil,’ sell the product to as many people as possible, and leave town before the customers realize they were the subject of a fraud.
Modern-day equivalents of ‘snake oil’ still abound. Eben Byers was a professional golfer of the 1920s whose jaw fell off his head due to the consumption of ‘Radithor,’ a radioactive solution containing radium that gave him a ‘toned-up feeling.’ Cigarette manufacturers throughout the 1930s and 1940s said the ‘purifying effects of heat’ of their products improved health and vitality. ‘Fat-free’ foods are often foods that replace fats with large amounts of equally unhealthy sugar while still being marketed as healthier alternatives.
All these products were successfully marketed to human beings (just like you and me) who desired certain end states, but wanted a shortcut on the pathway to health and vitality. In some cases (like cigarettes), we just wanted to feel like we were being healthy while still engaging in unhealthy vices. Whether it’s victims of ‘get rich quick’ pyramid schemes, bodybuilders living on human growth hormones who die of heart attacks in their early 40s, or people consuming poison that tastes good, there will always be a market for people who want to shortcut their way to success while pretending there will never be a price to pay.
Which also applies to FMEAs … sort of.
FMEA Oils (A Short History)
Unlike snake oil, the very first FMEAs of the 1950s were quite amazing. They helped designers, engineers, and manufacturers (particularly in space and aeronautics) come up with what we call ‘corrective actions’ for designs of new products and the manufacturing processes that created them … at the earliest stages of production.
Before long, people wanted to replicate the success of FMEAs across their industry or company. But there was a problem. Conducting a FMEA effectively requires significant organizational ‘philosophical’ investment. It requires organizations to do things differently from how they have always done them in the past. It requires buy-in from the CEO (or whoever is in charge). It requires people to understand that FMEAs help you find dozens of ‘fast, simple and free’ ideas to implement in your first design without being able to ‘prove’ that any of them will improve reliability. It requires participation from the senior representatives of different organizations. Those representatives need to be trained before they partake in the FMEA. And they also need to be done early in the production process (which is unfortunately the time when most designers, engineers, and manufacturers are so ‘excited’ by the new product they don’t want to waste time thinking about how it might fail).
So while plenty of organizations wanted the benefits of FMEAs, they weren’t prepared to take investing in them really seriously. A well-executed FMEA consistently yields a significant return on investment by preventing problems, crises, and customer complaints in the future. But humans are very good at compartmentalizing and convincing themselves that future problems aren’t today’s issues.
Nonetheless, many governmental organizations, militaries, and companies demanded that FMEAs be completed by their suppliers without ever committing to them ‘philosophically.’ So FMEAs are mandated in contracts, and clients demand evidence that they have been completed. But what does this evidence need to look like? Standards started to emerge, starting with the US Department of Defense, and then most of the other engineering societies.
These standards are incredibly helpful for people who want to conduct FMEAs properly. They emphasize tailoring your FMEA approach to ensure you come up with those incredibly important corrective actions. They contain tools to help you and your team evolve your approach. They give you guidance on what you need to focus on, and what you don’t.
But these standards also contain definitions, examples, and templates that people who don’t want to take FMEAs seriously can simply use as mandated artefacts. The advice in these standards can be cherry-picked to become non-negotiable criteria of excellence.
So FMEAs started to become ‘snake oils.’ FMEAs have been diced up, re-branded, packaged, standardized, converted to software and AI outputs, all to be able to simply and quickly say that a FMEA has been conducted without ever doing it properly, all the while printing out tables and documents that ‘comply’ with standards.
And it doesn’t work.
So what is a ‘real’ FMEA?
A real FMEA is a group brainstorming activity with between four and eight fully trained cross-functional representatives. Any fewer than four means not enough of your product, system, or process is covered. Any more than eight means that not everyone gets heard. Having lots more than eight means that the FMEA is based on a couple of dominant (loudest) personalities.
A real FMEA is not guaranteed to cover the entire product, system, or process. In fact, the best FMEAs rarely do. The focus is on analyzing part of the item really well, and not the whole item really poorly. To do this, the facilitator needs to be amazing.
The facilitator is constantly managing personalities and ‘what’ is being talked about. The facilitator is constantly ensuring that brainstorming focuses on the riskiest parts of the product, system, or process.
And how does he or she do this? By starting from the start. That is, the facilitator brainstorming the requirements of the item with the team – even if the ‘requirements are already done.’ In this context, a requirement is what a customer, operator, or user would say they want to be able to do with your product, system, or item. This is best done with character avatars representing real humans, which helps make sure we cover things like usability, energy consumption, and all the other things we forget if we rush past this step. By the end of this stage, the team can really empathize and understand how people will interact with each item, meaning that all the remaining brainstorming focuses on their challenges (not ours).
Importantly, the facilitator only allows this brainstorming to go on for a distinct amount of time … even if it means they stop the FMEA team mid-sentence. Everyone then has a break, while the facilitator (who will be exhausted by the end of the FMEA) organizes and prioritizes the requirements that the team brainstormed.
When the team comes back from the break, the facilitator will then focus on understanding what needs to happen for the most important requirement to be met. If this is a design FMEA, then the facilitator will start brainstorming the item functions needed to support the most important requirement (flow rate, force, calculations, cutting, grinding or whatever else). If this is a process FMEA, the facilitator will start brainstorming through process steps (transport, polishing, conveying, measuring or whatever else).
This process continues, with the facilitator stopping each brainstorming step at a fixed time, asking the team to take a break, prioritizing the topics discussed, and then brainstorming the next step in priority order when the team returns. Each step terminates after a fixed time, with its elements prioritized before moving on to the next.
The end of this structured brainstorming process is a lot of great ideas called ‘corrective actions’ that address the riskiest parts of the product, system, or process that influence the riskiest functions or process steps that in turn support the highest priority requirements.
But I don’t want to ‘not’ do everything …
This applies to most engineers, designers, and manufacturers. We tend to want to do an ‘amazing’ job, which means we brainstorm ‘everything.’ But … this never works.
If (for example), we don’t stop brainstorming requirements until we believe we have got them all, then we have less time to do the next steps. If we then brainstorm the functions that support each requirement, we will almost always have no time to do anything else afterwards. So, to convince ourselves we have done a good job, we throw in a couple of token ‘corrective actions’ in the last five minutes allocated to our FMEA because we didn’t budget our time. All because we want to do ‘everything.’
But if everything is important, then nothing is important.
For example, let’s say we are designing the next generation of a two-way radio. The thermoset plastic case will be changed in shape or form, but will otherwise be made of the same material using the same process. We will probably have a team that is very experienced in making these cases, knows exactly what can go wrong (and how to prevent it). Further, we might also know that our robust plastic cases rarely fail in the field (because of the knowledge of our manufacturing team). There is absolutely no point in brainstorming how our plastic case might fail as part of a formal FMEA. The electrical, mechanical, and software engineering representatives will simply have their time wasted if we brainstorm how a well-known, very reliable plastic case might fail, and come up with corrective actions that our manufacturing team was likely going to do anyway.
Engineering judgment is amazing
We can’t design anything if we need to brutally analyze everything before we make a decision. And this is one of the reasons people don’t like FMEAs. They are often seen as endless discussions about every conceivable way something that hasn’t been designed might fail. Many textbooks and standards say that FMEAs are ‘bottom-up’ analysis tools, which implies we brainstorm every way something can break and then think about how we can do something about it.
The people who mindlessly write these phrases in textbooks and standards have clearly never facilitated a FMEA properly …
A real FMEA helps us focus on the ‘troublesome’ parts of our new product, system, or process, leaving the rest for engineering judgment. By continually prioritizing our brainstorming effort as we go, we (by definition) let our team do what our team does best. FMEAs should only focus on the issues that are going to ‘keep us up at night,’ and nothing else.
How do I avoid FMEA snake oils?
Simple. Ensure you are ‘philosophically’ committed to FMEAs as an organization.
If you are looking for a service provider to help you ‘do’ a FMEA as long as it doesn’t change how you ‘do’ things … then you aren’t committed. If you want AI to do your FMEA entirely … then you aren’t committed. If you only have the ‘junior’ people attend FMEAs … then you aren’t committed. If you are skeptical that an approach of preventing problems versus rectifying is possible or valuable … then you aren’t committed. If you want people to provide individual inputs into a FMEA via email for someone to collate and ‘turn into’ a FMEA … then you aren’t committed.
Don’t rush it. If you aren’t committed, don’t pretend you are. But this can change if you research and understand how FMEAs are supposed to go. This can also change if some of the misconceptions about FMEAs are corrected (like having to do everything). Or … you can keep eating a diet rich in snake oils.
Ask a question or send along a comment.
Please login to view and use the contact form.
Leave a Reply