How do I learn more about root cause analysis?

April 22, 2011 5:23 PM

Root cause analysis. What exactly does it mean in the context of an engineering job and how can I prepare for this interview?

So I have a job interview coming up for a hardware mechanical engineer position. One of the requirements is the ability to do root cause analysis, documentation, and recommendation.

I can't say I've done formal root cause analysis beyond general troubleshooting (this thing is broken so I should fix it). How can I best prepare myself and demonstrate that I can do the job?

Are there any books/articles/websites on this topic? I want to show that I'm proactive in my willingness to learn.

posted by just.good.enough to work & money (8 answers total)

Root cause analysis is all about finding the root of the problem, not just the proximate cause.

A quick fix might be rebooting some computer. When it comes up, everything works fine. Root causing the problem means finding out why the computer failed in the first place, and understanding why the reboot was a valid band-aid fix. Then you talk about whether the cost of fixing the root issue is worth incurring, or just live with the band-aid.
posted by b1tr0t at 5:29 PM on April 22, 2011

Reading about The 5 Whys might be a great start.
posted by xil at 5:56 PM on April 22, 2011

Find a quick and dirty course and learn the basics of the Six Sigma program. Maybe this one. Should give you enough info to get through an interview.
posted by raisingsand at 6:14 PM on April 22, 2011

Try googling "Root Cause Analysis". One of the first hits I found was this presentation -- it covers a lot of the basics.
posted by elmay at 6:33 PM on April 22, 2011

I conduct RCA's fairly often. It's a pretty intuitive process. The idea is that you gather a necessary team of people together, state your problem, and do your best to follow all logical paths through contributing causes to possible root causes. You'll often find that 5 different people walked into a room with 5 different ideas of what the real cause was and by the end of the analysis, something somewhat unexpected was uncovered as the root cause. When I conduct them, I use Visio and a projector to visually map out the chain of causes. So, for example, let's say you came home and wanted some ice cream, but you opened the freezer and found it had melted. A simple RCA might go something like this:

Ice cream melted ----> freezer temp too warm -----> freezer was not running ------> power is out in the kitchen -----> circuit breaker tripped -----> living room and kitchen outlets are on same circuit

So, without thinking about it in this way, you may have just gone and flipped the circuit breaker switch without thinking about how your kitchen should be on its own circuit, leaving the possibility of more power failures in the future.
posted by TrialByMedia at 7:35 PM on April 22, 2011

I had Root Cause and Corrective Action (RCCA) training when I worked at AlliedSignal. I could swear the slides elmay mentions above were used in my class. Googling around I found the AlliedSignal Process Improvement/Problem Solving Model (pdf) which I think would help you out.

As someone who has performed and written root cause analysis reports, it's OK that you haven't done any in the context of an engineering job. I never did any before my first job in industry. My company trained me in how they wanted the analysis performed and how they wanted the RCCA written. After I did my first analysis, my Lead Engineer reviewed and redlined it, I edited it, and we presented it to my manager. After she was OK with it we instituted some changes, and that anomaly never occurred again.

IMO if you personally are smart enough to get a ME, you're smart enough to learn an arbitrary company's way of root cause analysis, do the analysis, write a report for your manager in that company's format, and recommend a fix.

Now go land that job proactive new ME! You're better than just.good.enough.
posted by Rob Rockets at 8:13 PM on April 22, 2011

I work for a forensic engineering firm and we do this by trade.

To me it's just basic engineering analysis. I mean our cases obviously come from the failures, so it's generally pretty easy to know where to start.

It's really just common sense engineering. You look at other's designs, do your calculations, and just think, "hey, is this right?" It's really just what ME's do by trade, or should have done more likely. The important concept you want to keep in your mind isn't "This thing is broken, how can I fix it" it's "why did this thing break, and what can I do to refine the design so that it doesn't".
posted by sanka at 8:37 PM on April 22, 2011

Thanks for all the help, everyone. It didn't even come up in the interview.

I did find this helpful:
Root Cause Analysis for Beginners PDF
posted by just.good.enough at 6:02 PM on April 25, 2011