Root cause analysis
Solve it once, for good!
Most problems come back because something was fixed that was not the cause. Four proven methods help you find the causes that actually explain the problem. Step by step, in your own browser, without an account.
Which method fits my problem?Sound familiar?
- Tuesday’s outage is the same one as last month. And the one in March.
- There is an extra check on it now. And it still goes wrong.
- We wrote a procedure for it. Nobody uses it.
Three different worlds, one condition: something got fixed, and the problem is still there.
Why the quick fix leaves the problem in place
A problem shows itself through a symptom. The symptom is where it hurts, not where it comes from. Removing that symptom feels like progress: it is visible, it is fast, and it works. Until the cause causes it again.
Often not even in the same place. The same cause turns up a quarter later in another system, another department, another form. And then nobody recognises it as the same problem, so the investigation starts from scratch.
There is almost never just one cause. A problem of any size usually has two or three acting together. Remove one and it happens less often - and that is the most dangerous outcome there is. It looks like it worked, so nobody looks any further, and the two that remain quietly get on with it.
On top of that, without a method you find the cause you already suspected. Not through incompetence: that is simply how attention works. You look where you expect to find something, and you find something, and it is even partly right. That is what these four methods are for. Not to be clever, but to stop you from confirming yourself.
The order that works
The methods are not separate exercises: each step feeds the next one. It is a practical order rather than a rule, and a problem of any size often has more than one root cause.
-
01
Define
Is / Is-not
Record what the problem is and what it is not. The difference between the two points you in a direction.
-
02
Prioritise
Pareto analysis
Work out which category does the most damage, so you spend your effort on the right problem.
-
03
Widen
Ishikawa
Collect possible causes per category, so you do not settle for the first one that comes to mind.
- 04
Not just for technical problems
Root cause analysis is about cause and effect. That turns up everywhere people are trying to get something done together.
- A school watches the same year group fall behind on reading comprehension three years running.
- A local authority receives the same question through three channels and answers it three times, differently each time.
- A construction project overruns handover by two weeks every time, whichever contractor is on it.
- A sports club loses a fifth of its junior members after every winter break.
- A restaurant jams up in the kitchen on Friday evenings, and only on Fridays.
- And at home you are late every Monday, while the rest of the week works fine.
Five of those six need no machine, no measurement and no incident log. These methods work on what you can observe and ask about. All they require is a willingness to keep asking past the first answer.
Your analysis stays yours
Everything you enter is held in your own browser storage. Not one line of your analysis reaches our server: not when you save, not when you share, and not when you print. No account, no email address, no cookies.
That is not a promise but a property of how the site is built. If you want to take an analysis to another computer or to a colleague, that goes through a file or a share link that carries the content itself.
Starting takes a minute
Four questions and you know which method to begin with. After that you create a case and work through it step by step.
Which method fits my problem?