I Hate These Straight Lines

Root cause analysis helps you solve a problem by helping you discover its most important causes. Many different methods and tools are used for this, but some of them are painfully inadequate. 5 Whys is one of those tools.

Pretend that the image below shows a problem and its causes. The problem is at the top, in the tip of the yellow area. Now follow the various branches down from the problem. The deeper you get, the more branches you have, and at the tip of every branch is something that acts as a cause. The problem actually happened because a whole bunch of causes came together in exactly the wrong way. This is what real problems look like if you investigate them thoroughly. There is no single, straight line of causation.

How will you solve this problem? One dumb way would be to try cutting off one or both of the major branches at a point just before the problem i.e. the "immediate cause" approach. Another less-than-smart way would be to eliminate one major cause at the end of the most obvious branch i.e. the "straight line" approach. Either of these might offer relief for a little while, but neither is likely to last long. All you really did, essentially, was try to dam the Mississippi River at the worst possible point or cut off one of many inputs. In other words, you'll accomplish nothing or force the river to find a new way to the Gulf... oh yeah, I thought this looked familiar.

So how could you solve the imaginary problem i.e. stop the Mississippi from draining into the Gulf?

• Remove or change some of the major causes i.e. endpoints?
• Place barriers to stop adverse cause flows before they become unstoppable?
• Change the system so that there are no adverse cause flows?
• Change the system so that separate adverse cause flows can't merge together? 
• Change the system so that the major flows can't head for New Orleans?!

Most of us don't have the ability to permanently stop the rain, dig major canals or tilt the North American plate so all runoff heads to Canada, so those last three solutions aren't going to work. That's not really the point of this analogy, though. The real lesson is that a single straight line cause analysis (like 5 Whys) is completely inadequate in many real world situations i.e. anything beyond the utterly simple.

If a problem looks simple enough that 5 Whys seems adequate, then you basically have one of the following scenarios:

1. you were taught that 5 Whys is adequate and you haven't questioned that claim, or
2. you have totally underestimated the difficulty of the problem, or
3. the problem never was very difficult to understand, or
4. you really don't care if the problem gets solved or not i.e. you're just doing a "root cause analysis" because somebody forced you into it.

In scenarios 3 and 4 you don't need a root cause analysis method that actually works, in scenario 2 the problem is more complicated than 5 Whys can handle, and in scenario 1 you are the problem. Scenarios 2, 3, and 4 show you that 5 Whys is generally unnecessary or inadequate. No amount of faith on your part is going to change that. You should ditch it and find something better.

Who can be Dangerous

It's not a question and it's not permission. It doesn't refer to the band or the Doctor. It is a word sometimes used in problem solving, safety investigation and root cause analysis, and it can be dangerous. It can also be powerful, but with power comes responsibility.

I have talked about ditching the word why and replacing it with what, where, when and how. I didn't include the word who because who must be used sparingly, with extreme caution, and always in a generic sense i.e. anonymously. If you use it in a non-generic sense, you will be perceived as chasing blame, and you will poison the environment you're trying to improve. Once you do that, you can forget ever again being granted good will, cooperation and honesty in your root cause analysis work.

Who, in a generic and anonymous and blame-free sense, can help find issues that need more attention.

• Mechanics that seem to cause more rework.
• Operators on one shift having a noticeably higher error rate.
• Laborers at one job site getting injured more frequently.
• Server techs/admins experiencing more service interruptions .
• Nurses in one ward having more adverse patient care outcomes.
• Engineers in one department releasing more designs with errors.
• Technicians in one lab contaminating more samples.
• Baggage handlers in one terminal misplacing more bags.
• Programmers on one project missing more deadlines.
• Sales partners in one office not meeting quotas.

See the pattern? Asking who in a generic and anonymous sense points the way to other problems. Asking who for learning and not for blame uncovers issues closer to the root causes. Use the word who sparingly and for the right reasons and you will be rewarded with valuable information.

On the other hand, asking who for the purpose of blame and scapegoating is dumb. When you find a person to blame, and maybe to fire or let go, you send a message, and that message is fear. People will receive that message and act upon it, and you won't like the results. Asking who in order to place blame is a stupid practice. You need to ditch it.

Delete the Word "Why" From Your Vocabulary

Root cause analysis is a process that loves to have people ask questions like: why did that exist, why did that occur, why was that done, why why why why why. I think it's time to stop using the word why. It's played out, over with and done for. It has too many negative connotations and it promotes sloppy thinking.

Replace why with what, where, when and how. These will get you to a better place as an analyst because they force you to dig into details with more energy. They will also improve every model or diagram or chart you make: replace unsupported things that just sort of affect each other with solid facts describing what factors were present, where they were, when they were there and how they did what they did.

Adults ask what, where, when and how. Kids ask why. Ditch it (the word not the kid).

Root Cause Contrarian?

It's really simple. I disagree with the self-appointed "experts" i.e. the academics and the consultants. On one hand you've got the old school single root cause dinosaurs, on the other you have the au courant multipath causation nerds, and on the gripping hand are the holistic hipster complex systems dweebs that say there are no root causes. They're all right, wrong or fuzzy depending on the situation, but always annoying.

Do you want to know what a root cause is? It's really not that difficult, but you couldn't tell that from listening to the "experts" whine about how this, that and the other thing should or shouldn't be included in the formal definition. They're taking a simple concept and turning it into a philosophical discourse on the nature of the universe. Who the hell cares? We have problems to solve, dangit!

Sweep away all the cruftiness, bloviation and ivory tower drivel and this is what you get.

A root cause is something that's important to the creation and
impact of a problem and which doesn't result from
something more important.

Was that so hard? Oh sure, every "expert" is going to find a problem or five in this. Fine. That's their choice. I'll keep doing what I've been doing. Oh, I've read their books and taken their courses and heard them speak at conferences and seen what they write online. What I learned from all this is that these guys think they have to mess with definitions and tools and methods. "We have to differentiate ourselves from the competition" ... barf puke spit, may I never say those words in that order ever again.

Root cause analysis is about solving problems - saving lives, preserving assets, increasing quality, minimizing downtime - in the most effective ways possible. All else is bunk. Period.