Why do so many businesses find it so difficult to provide good quality to their customers on every single transaction?
We've been working on improving quality for a long, long time. In the last century we even decided that quality should be treated as a science and created the profession of quality management. Since then we've developed tools like quality circles, total quality management, statistical process control, and Six Sigma, all designed to help us reach the nirvana of zero defects.
Despite all that, my dry cleaner brings me someone else's clothes, my substitute mailman brings me someone else's mail, my order for a new IBM T42 laptop computer was mistakenly filled with a used one, and my surgeon removed the wrong kidney. Just kidding about the kidney, but unfortunately hospitals make errors that serious every day. Why do so many businesses find it so difficult to provide good quality to their customers on every single transaction?
Poor quality typically results from variability. The dry cleaner probably has a process for handling clothes that, when everything goes perfectly, gets the garments cleaned correctly and back into the hands of the rightful owner on the day promised. The same can be said of most businesses. The challenge is in getting the laundry out when things don't go perfectly. The challenge is repeatable reliable performance.
To control quality requires controlling variability, or to design a sufficiently robust process that it can delivery quality despite high variability. What does that really mean?
My dry cleaner may have high employee turnover, which in turn means high variability in who is handling the laundry. He has two choices. The first is to find ways to control employee variability by focusing on reducing his turnover. His second choice is to design a process for handling laundry that can be successfully executed by someone very new to the company, even on his first day on the job. He may want to do both.
Mistake-proofing (a.k.a. "poka-yoke" if you Google) is a methodology to address variability of a process. And please don't call it idiot-proofing. Despite all my college degrees, I can assure you that more than once I have tried to order something on the Internet for which I didn't enter the data correctly. Up pops up a note, usually in red lettering with an asterisk, telling me to correct the information before I can continue. I much prefer the term mistake-proofing to idiot-proofing, don't you?
Mistake-proofing recognizes that people make mistakes. We always have and we always will. The goal is to keep those mistakes from becoming defects.
For any given process, you will need to decide how far to take mistake-proofing. At one end of the spectrum, called prevention, you can eliminate the possibility of human error becoming a true defect. That can be very expensive. If you work in a nuclear plant or a transplant center, that may be money well spent. In a dry cleaner, it may be a poor investment unless you're talking about Environmental Protection Agency- or safety issues. At the other end of the spectrum, called detection, you may want mistake-proofing to ensure that all defects are caught before they leave the building. The extent of mistake-proofing that you decide to develop for a given type of quality problem is a business decision, no different from many others you make.
So how do you mistake-proof? Start by understanding that humans make mistakes and it doesn't mean that they don't care or that they're idiots. But also understand that you can find ways to keep those mistakes from becoming defects. Select a process where you have a quality problem. Look at the potential opportunities for error. Consider how errors are most likely to be made and the cost of those errors becoming defects. Many people begin by focusing on detection mistake-proofing, and then move toward prevention, but there is no one size fits all approach. Look at the nature of the problem and determine the right approach for your situation.
Many products you use are mistake-proofed:
Gas caps tethered onto cars to prevent driving off without them
Gas cap fitted with ratchet to signal appropriate tightness
Hole in bathroom sink to reduce chance of overflow
Color coding of vials for blood tests
Lawnmower handles that shutoff the power when released
Most forms of detection oriented mistake-proofing are simple and inexpensive. Start looking at your house, your car, the grocery store -- all the places where you work and play. Notice what is, and what could be mistake-proofed. You will get ideas of how simple and how valuable mistake-proofing can be. Then turn your attention to your own business. It's time to get the laundry out.