
David Collins explains the concept of mistake proofing as applicable to Chinese factories, and why managers should never blame operators for mistakes.
TRANSCRIPT
Just the other day I was working with a contract manufacturer and they talked about what problems they were having with their operators. They were talking about how, “Oh, our employees aren’t so good.” That should never be an excuse. You should never say your employees aren’t good.
Ask yourself this: Why haven’t you mistake-proofed the job? Why haven’t you put in poka-yoke? Why haven’t you setup the job station so the operator cannot make a mistake? Your engineers, your production people, should definitely be smart enough to do that.
And there’s a lot of different ways to mistake-proof a job. If it’s dimensional, maybe putting go/no go devices. Maybe using some kind of vision system that’s cheap and easy to use (in electronics), can make the job so it can’t make a mistake. Barcoding or setting up the job so that only one type of part can be put on the position at a different time – another way of mistake proofing.
Good engineers and smart people can mistake-proof a job almost 100% so that the operator can’t put on the wrong part, can’t assemble it incorrectly, and cannot ship it to the next station in an improper way. You don’t expect your operators who don’t have college educations – maybe just barely finished high school or didn’t even go to school at all – to do the right thing, but you do have engineers, you do have production people who should understand how to mistake-proof a job.
Tooling, very simple. Tooling, fixtures, and other things are all great techniques for mistake proofing a job so that the operator has no chance but to make it correctly the first time. There’s much more sophisticated ways to do that through television monitors, bar coding but those are expensive and maybe not necessary for your factory.
Other things are good SOP’s [Standard Operating Procedures]. Are the SOP’s clear, written well? Are they in such a way that the operator can look up at them and understand the standard of work that they’re supposed to have. Most SOP’s I see are a joke. They’re black and white pictures of something with a circle around it. You can hardly tell what it is. Somebody just took a picture because somebody said they needed an SOP. It isn’t done properly. The job instructions aren’t done in such a way that the person can do it.
And then, another really good mistake proofing, and it appears often in China that supervisors think they’re too good to do this, is to actually walk on the line and watch the people do the job, and are they actually doing it to the SOP? If they’re not doing it the SOP, either the SOP is wrong and the operators found a better way to do the work, and you should be very open-minded and see if you can be taught something, or two, the operators getting lazy in not doing the SOP.
I don’t usually see good supervision on the line managing to the SOP, managing to the poka-yoke, managing to what is important coming off of each job station; instead what I see is leaders who just want to stand around and yell at the people and tell them they’re not doing a very good job. It’s not really a very good way to make your manufacturing run well.
The other things with mistake proofing are: Do you get your operators involved in building the SOP? One of the most important things Toyota has ever done is build their teams and their teams actually help the industrial engineers and others build the SOPs on the line because the people are the ones doing the job eight to ten hours a day. They know how that job works better than anybody.
Don’t be arrogant. Don’t think that you know everything. Your operators, who do the single job every day constantly, know probably more about that job after a few weeks than the engineer who designed it. Listen to them. Help them design it so it can’t make a mistake. But really, mistake proofing is something that every factory should be working towards so that it can deliver high quality product in a very good production manner.