We're accustomed to thinking of poka-yoke (error-proofing) as something for a manufacturing assembly line, or at the very least, for a machine. In this standard conception, there are fail-safe devices (some cool, some pretty basic) to ensure errors are prevented. Electric eyes in elevators keep doors from closing on people. Some hotel rooms are equipped with a room key holder that turns off the power when the key is removed to prevent electricity from flowing to the room when it's vacant. Gas caps on cars are attached with a cord to prevent drivers from leaving it on the roof. (I left a cap near Grants Pass, OR, if anyone happens to see it….)
But why not institute poka-yoke for management? Why not create systems that prevent bad management practices from taking hold?
I thought of this yesterday when reading the NYTimes interview with Kevin Sharer, chief executive of Amgen. When he ascended to the big chair, he talked to the top 150 people in the company one at a time for an hour.
I seek feedback, because that’s the only way that you can grow as a C.E.O., which is a very isolating job. And so if you don’t create mechanisms to get authentic feedback, you won’t.
Sharer sees this isolation as a major "error." It might lead to, oh, I don't know, a $6000 shower curtain. Or a $1400 garbage can. So why not institute some poka-yoke to prevent the problem?
Every year I have the head of human resources at Amgen who reports to me conduct an evaluation of me done by my team that they write up and then present to me and the board. And that’s an uncomfortable process, of course. You hope, as C.E.O., for the team to say, “Boy, boss, you did a great job this year, are we lucky to have you.” And, by gosh, every year they’ve come up with three or four things that are quite authentic that I ought to do better. So you’ve got to create those kinds of feedback loops.
But wait! There's more!
There are probably four things I do [to stay in touch with employees at all levels]. First of all, we take an all-staff survey every two years that is very comprehensive and very broad, and I really read that closely. One of the questions is: What kind of job do you think Kevin Sharer is doing? So you want to get some direct feedback? There it is. I read all the write-in comments.
Second, I’ll go out on rides with sales reps, just me and the sales rep calling on doctors. I try to do that four times a year and spend a day with them. I visit every factory every year, and have small focus groups with people.
I give probably 10 town hall meetings a year where 150 or 200 people come in, and the Q. & A.’s are robust. We have a very effective human resource department here that I listen to broadly, and so you just try to get feedback every way you can. This is fundamentally an isolating kind of job. And so if you don’t take affirmative action to break the isolation, you will be isolated.
What's striking (to me, anyway) about Sharer's approach to management is that he doesn't whine about the exigencies or constraints of the job. He just finds a way to poka-yoke the aspects he finds dangerous.
By contrast, I work with a lot of managers who complain that their schedules don't allow them to spend enough time with their staff. Or that their email burden is so overwhelming that they don't talk — really talk — to people. And while I won't deny that managing isn't easy (lord knows I've screwed up enough communication on the job), Sharer's example is pretty powerful. Why not create a system that guarantees you'll get the results you want?
So the question to you: what would you poka-yoke? And how would you do it?
I agree CEO’s need to work to keep real feedback available to them. Getting trapped in the center of a bunch of yes-men (especially if the organization “doesn’t want to hear problems just solutions…). But is it poka-yoke? It doesn’t really mistake proof I don’t think. It provides some additional information that can be used wisely to avoid mistakes.
That would seem closer to explaining to workers that a USB cable must be plugged in with the side that says top on the top. Which is not the same as designing the USB cable to only plug in if you have the top in the right orientation.
Granted this distinction might not be that great. I have just thought about the idea of applying mistake proofing thinking to problems (but not actually implementing counter-measures, other than awareness) that prevent mistakes. To me this seems like something less than actual poka-yoke.
I agree CEO’s need to work to keep real feedback available to them. For management poka-yoke, you could design a system that didn’t allow a meeting to be scheduled until an agenda was posted
John,
You raise a good point. I suppose whether or not it’s “true” poka-yoke depends on whether the internal surveys and review are a routine and necessary part of the CEO’s annual review. If, for example, he had to submit the feedback to the board of directors each year and list his action steps based on that feedback, that would get a whole lot closer to true poka-yoke.
But no, it doesn’t actually prevent him from sucking, or treating his employees like dirt, or running the company into the ground. Given that humans have free will, my guess is that true human poka-yoke is impossible.
What do you think?
Well unless there are constraints then poka-yoke probably is impossible. With a USB cord you can design it to essentially be impossible to put in wrong (though really not totally impossible, I mean I think I could get it to fit wrong if I really tried). But a much easier way to not plug it in correctly would be just not to plug it in at all.
For management poka-yoke, you could design a system that didn’t allow a meeting to be scheduled until an agenda was posted (I can see doing that with software – of course people could just put in fake agenda stuff – but I can push in the USB cable the wrong way too). Of course that can create problems, I sometimes want to schedule a meeting and having set the details yet. But I am just trying to think how you could enforce less obvious (less physical often) poka-yoke solutions.
You could require things like an A3 report before x is done. Of course the A3 report could be done in a sloppy lame manner. I think the non-physical poka-yoke attempts do have challenges, sometimes I can imagine creating physical like (software…) bottlenecks… Which then could act as preventing certain types of mistakes…