Comments on Lessons Learned: How to conduct a Five Whys root cause analysis

Eric, Thank you for this article. Like some of th...

2010-03-03T19:32:38.021-08:00

Eric,
Thank you for this article. Like some of the other commenters, I see mainly obstacles to implementing this in my organization:

dispersed teams - in some cases not all people are possible to be gotten into the same room, and in many cases this involves working with vendors. I suppose we could ask "why does our relationship with this vendor allow these problems to continually arise?"

resistance/distrust to any sort of predetermined meeting structure--anything that seems canned or gimmicky. I feel like I would have to introduce this by stealth.

I'd be interested to see more examples of how you would start to get this off the ground in a company--as a counterpoint to these examples of how it works in a mature form.

Sam

Hi Eric..I enjoyed your article. I work for a majo...

2009-09-25T07:03:46.529-07:00

Hi Eric..I enjoyed your article. I work for a major pharma company. We use 5Whys (along with fishbone) as part of a 7Step root cause analysis approach. I am constantly amazed about how difficult people find the 5Whys tool. I am no expert myself and believe the tool IS difficult to use. In my mind it's so-called simplicity has been totally over publicised. The difficulty is that people tend to jump steps in the causal chain and start placing solutions as answers to whys. I have yet to see a 5whys that I would consider to be a good one - including anything I have generated myself. I have read extensively on the Toyota approach - they have root cause analysis sensai who challenge ever minute step of the analysis until it is correct - I feel we are a billion miles away from that. However your article makes good points about insufficiently robust systems - indeed most root causes lie in management failings - but all too often 'human error' is an easy fall back.

Hi, Eric. As usual, this is a great post. I'm...

2009-07-10T23:55:45.311-07:00

Hi, Eric. As usual, this is a great post.

I'm a big fan of doing a Five Whys analysis, and wanted to add one tip. Instead of telling people to look for *the* root cause, I have them search for *a* or *some* root causes.

My goal with this minor wording change is to avoid arguments over the right hierarchy and focus on the goal, getting some solutions to try. When people are wound up about a problem, and especially when dealing with engineers, it can be easy to analyze past the point of diminishing returns.

Eric, Our site went down for about 30 minutes this...

2009-07-06T07:48:00.200-07:00

Eric,
Our site went down for about 30 minutes this weekend. I'm going to do a Five Why's analysis with the tech team today. Will let you know what happens.

best
steven
TakeLessons.com

I have numerous ideas to work on, mostly related t...

2009-07-03T23:23:50.065-07:00

I have numerous ideas to work on, mostly related to solving the problems I face. But only a limited time.

I think I should take the similar approach of limited investment to my problems, and when the problem recurs I put another bit of effort to improve my solution. Great approach! A sad state that we don't have more time.

Eitan, that's a real issue. It makes sense to ...

2009-07-03T15:45:26.835-07:00

Eitan, that's a real issue. It makes sense to do prevention work in rational batch sizes, and teams that do routine root cause analysis get better at finding the right batch size over time. Luckily, in most prevention situations, even the first few steps in prevention can pay time-savings dividends quickly. So even if the total cost of a given solution is higher when done incrementally, this is offset by the reduced risk of preventing something unimportant.

Plus, the real driver of context switching overall is firefighting and other bugs.

Joseph, you might find these articles helpful: ht...

2009-07-03T15:35:25.971-07:00

Joseph, you might find these articles helpful:

http://entrepreneur.venturebeat.com/2009/06/05/palantir-keeps-it-lean-and-mean-on-five-year-journey-from-zero-to-150-employees/

http://startuplessonslearned.blogspot.com/2009/07/how-to-conduct-five-whys-root-cause.html

http://www.joelonsoftware.com/items/2008/01/22.html

If you have a question about a specific technique/situation, let me know. I'm hoping to find other examples to share.

Thanks for a great post. I have a question regard...

2009-07-03T14:27:46.850-07:00

Thanks for a great post.

I have a question regarding the "chipping at the problem" approach.

If solving the issue requires a "big" investment (Let's say 20 hours). But after each session of 5-whys I decide to invest only an hour in the solution. All the setup time and context swithcing may turn this 20 hours task into 40 or 60 hours task.
Batching would have saved this setup time, but I invest my time sporadically, so there's a waste involved in this "chipping at the problem" approach also.

What's your experience with this?

Eric, Could you please give some concrete example...

2009-07-03T13:27:08.473-07:00

Eric,

Could you please give some concrete examples of Five Why's anaylsis? I understand the theory, but I am curious what it ends up like in practice once your hit the third or forth why.

@Wyatt - if you are willing to suggest specific su...

2009-07-03T08:31:28.366-07:00

@Wyatt - if you are willing to suggest specific subheads for this article, I will add them. Hopefully I'll learn from the experience and will think of adding them in the future. Deal?

@Eric -

1) if the client is the cause of waste, that's good to know. You can ask "how can we serve this client while mitigating all this waste?" or even "is serving this client really worth it?"

For example, periodically somebody at IMVU suggests that we stop supporting Windows 98 as a platform. It's so old and so buggy that it's a real pain to ship 3D software on it. But since we have a really good idea of the problems it causes, we can do the ROI analysis whenever somebody suggests that. We just add up the revenue we've made in the past few months from Win98 users, and compare to the pain that Win98 has caused as identified in 5Ys. So far, it's been a net win to keep supporting it.

2) On the trust issue, you simply have to avoid punitive measures altogether. If you have somebody on your team who is not pulling their weight - everybody knows it. 5Ys is not going to change that, although it might give you an opportunity to help that person become more effective. You never know if somebody's incompetence is due to their skills/talents or due to a bad context. You'd be surprised how much more productive they can become if you invest in tools, process, and infrastructure.

@Anonymous - thanks! @Daniel - yes, I think it...

2009-07-03T08:25:11.396-07:00

@Anonymous - thanks!

@Daniel - yes, I think it's worth doing even at very small sizes. You can be a lot less rigid about it, of course, but the fact that we are blind to sources of waste is something we all share. I recently met with a two-person-no-funding team that, after some hesitating, jumped in and tried it. They seem to be liking it.

(oh, and on the book front, you'll be able to get all of my essays so far in ebook, print-on-demand, and kindle editions - coming soon. let me know if you're interested)

Great article, Eric. I have one comment on style:...

2009-07-03T07:49:21.299-07:00

Great article, Eric.

I have one comment on style: could you break up your articles with subheadings? It would make article more approachable initially, and easier to scan when I go back to reference it.

For instance, above the "Let’s imagine the ideal solution would ..." paragraph have the heading "Training New Hires", above the "Returning to the meeting itself ..." have the heading "Chipping away at the problem", etc.

It's just a thought.

Other than that, you consistently have good content. Keep it up.

I have not read the previous post(s) on 5 why'...

2009-07-03T05:13:04.323-07:00

I have not read the previous post(s) on 5 why's (about to now though...), but I wanted to comment and let you know that I thought this was really great stuff. Furthermore, I think most of what you are suggesting is applicable in companies/ industries that are not software development. I work in advertising and this system would work to analyze failures.

Two questions:
1) What happens when the root cause of the issue is clearly something that a/the client did?
2) How can you keep employees from being fearful of these post-mortems? Obviously if you have a sub-par employee who makes too many mistakes over and over he/she has to be terminated ...a couple times that employees see "5-Whys = John Got Fired" and employees may start fearing the 5-Why and its benefits.

Thanks again for great post.

Eric
http://www.pixelmaverick.com

Great article as always, Eric. Have you considered...

2009-07-03T03:11:21.138-07:00

Great article as always, Eric. Have you considered putting all these into a book?

One question - at what team size do you think that it makes sense to go through this process formally? In theory it works with any team size, but when you only have 2 or 3 people in the meeting, the formality of the 5-whys seems, intuitively, a bit stifling. Then again, perhaps it's worth it. What do you think? Would you apply this technique in a start-up that was just you and two other guys, for example?

"Problems are caused by insufficiently robust...

2009-07-03T00:29:30.268-07:00

"Problems are caused by insufficiently robust symptoms"
symptoms => systems