Une étude de cas

As a consulting practice, we take a practical approach to managing safety: what works, how it works, why it works the way it does. Management theory? That’s someone else’s business, not ours. So, while you’re leading and managing safety, we’re busy studying what it takes – in the real world – to send everyone home alive and well at the end of every day. 

Useful to understanding that are cases where things went wrong; aka, failures. Those involving safety go by familiar names: near-miss, close call, the “good catch.” There are vehicle accidents, collisions, property damage, fires, explosions. Harm to humans comes with labels like first aid, reportable, recordable. 

Failures deemed serious are investigated, and occasionally an investigation report becomes public information, such as those written by regulatory agencies. On rare occasion leaders running a business see it as their duty to share their reports. A great teacher, failure is.

Still, talking about failure is never an easy thing to do. In my book, doing that is performing a public service. If you are someone who’s been there and done that, speaking on behalf of the rest of the world, thank you for helping make the world a better and safer place. 

If only the rest of the world would take time to read and understand these case studies of failure. The thing is, who’s got time to do that?

I do.

A Recent Report

Having been a report reader for five decades, I can attest that a large part of them are devoted to describing the “what” of what went wrong: things like O-rings and joint rotation; foam shredding; valves and piping; storage tanks and methyl isocyanate; cement and spacers for drill pipe. In one notorious case involving those last three things on the list, by my count there were a half-dozen investigations.

Things are the stuff of tragic failure. For those in the business, things take center stage, but for the rest of us, those things aren’t typically our problem. But don’t think for a moment that because you don’t have “that” you won’t have “that problem.” 

A perfect example is found in the investigation report I’m currently sifting through. A high-profile construction incident (easily found on the web). Prominently displayed are the two things at the epicenter of the failure: a cone screw and bracket. That’s how a working platform was supposed to be securely connected. When the two weren’t, a working platform tipped under load and five people fell; three died.

Why weren’t they properly fastened? There were issues with training and poor execution of the mentoring process for new hires. Performing high-risk work like this on midnight shift raised all kinds of potential problems: lighting, fatigue, limited presence of supervision, and proper verification that a critical job task such as connecting cone screws with climbing brackets was properly executed. 

Working at elevation required fall protection. The two wearing their fall arrest devices properly survived the fall. Had the other three done the same, this failure would probably have been labeled a serious near-miss; no report published for someone like me to read.

Would the causes have been any different?

A very sad case, indeed. I’m sure you’re not the least bit surprised to hear that multiple processes failed simultaneously. Rarely are the most serious failures the product of a single cause. 

As to those described in this report, for a quarter century we’ve been kicking off every safety leadership course we teach asking leaders about their toughest safety challenges. I’ve yet to stand in front of a room full of leaders anywhere in the world and not hear about lack of training, inexperience, limits of supervisory coverage, lack of time, non-compliance, and improperly wearing PPE.

I always make it a point to follow up: If you don’t successfully manage these challenges, sooner or later will someone get hurt? Easy to predict the answer; this case proves it correct.

Not that you needed any.

The Function of Investigation

Since this edition of the NEWS deals with reading investigation reports, it would help matters to understand the function of an investigation. A question for you: “What’s the purpose of an investigation?” Is it primarily to find root cause? Or what might be better stated as understanding the causes? 

That’s certainly the textbook answer. How can you fix a problem if you don’t know what caused it to happen in the first place?  

In practice, sometimes you can do exactly that, merely by eliminating the “thing” that caused the problem. NASA eventually solved the problem of joint rotation, seal leakage, and foam shredding by permanently grounding the Shuttle fleet. Years ago, a good client who’d had one too many hand injuries from people using knives decided to institute an “Open Blade Policy.” Knives were no longer allowed on the premises. Period. End of discussion.

As you might expect, checking personal knives at the gate did not go over well. But solve the problem, it did, no matter what its root cause.

That case nicely makes a bigger point: the goal of an investigation is to make sure something doesn’t happen again; at best, understanding causation is necessary but not sufficient. Unless you’re in the research business, studying the problem is not the point of the process.

So, understand cause? Sure. Fix the problem? Absolutely. Now, please take careful note as there’s a third function an investigation can perform: insight as to what’s really going on. We call it Performance Visibility: the degree to which a leader knows reality for what it really is. If every leader really knew what was going on, there’d be a lot fewer things going awry; low Performance Visibility has proven fatal. 

Sometimes it takes an investigation report to get a healthy dose of Performance Visibility. 

Safety Culture

Picture an important job done at night, staffed by a crew not particularly well trained, experienced, or closely supervised: execution and compliance were far from flawless. That’s what those in charge of the project described in leur report to the public. What does that say to you about their safety culture? 

Before you answer, we would do well to define that popular term of art. Culture is one of those words thrown around in conversation as if everyone knows exactly what it means. It’s been my experience that the opposite is more likely. So, let’s not be guilty of perpetuating poor communication. 

My definition of culture: what most people do, most of the time. 

By that definition, if you want to know what the culture is, all you have to do is to look: culture – what people do – is observable and measurable. See what I mean about being totally practical? No psychological profiling necessary. Don’t be fooled by the outliers – at either end of the distribution curve – and don’t fool yourself thinking the culture is something it is not. 

Culture is what it is.

It’s easy to jump to a conclusion about an organization’s culture; given serious thought, the potential for differences becomes apparent: an organization will have certain behaviors in common, and experience variability in others. In a company with multiple sites, is the safety culture at site A identical to B, C and D? Within site A, is the culture the same in operations as it is in maintenance? On day shift the same as night shift?

I suppose it’s possible. That’s not at all how I found safety culture. The differences between the best sites and the worst have been huge; within some sites, differences are stark as well. In practice, there is culture, subculture, and individuals.  Company culture is simply a homogenized mixture of all three.

As to the culture found in this case: “… for this particular work, our culture was not applied consistently from senior management all the way to the workforce.” 

Reaching that conclusion requires a different definition of culture than mine; perhaps they have one. By my definition, culture – what most people do most of the time – doesn’t get “applied” – consistently, or otherwise. It is what it is. A lot of times it’s not what management wants it to be. Sometimes it isn’t what management thinks it is, either.

That’s where the Performance Visibility function of an investigation becomes very useful. Now the executives know what they have on their hands, at least on one project. What are the odds it’s a one-off case? Or do things like this go on elsewhere all the time?

If I were them, I’d be seeking information. A lot more information.

Études de cas

Studying cases where others have failed comes with two huge advantages for a leader like you. First, you aren’t forced to deal with the challenges of getting people to be honest about their failures – or you, for yours.

Even better, you’re spared the pain of the particular failure. If only your counterparts were so fortunate.

On the other hand, what you’re missing out on is Performance Visibility, the third product of the investigation process: one that in my opinion is the equal of finding root cause. Reading any report, you could assume their problems are theirs – not yours. You may very well be right. 

What if you’re not? 

Were I you and reading any number of these case studies about failure, I’d be seeking more information. 

A lot more information!

Paul Balmert
Avril 2026