Day 15 - Poka Yoke

Today was spent in a learning session on mistake proofing or Poka yoke. The course focused on the concept of removing the possibility of errors occurring in the first place, to support not only delivery of defect-free products to our customers but also to support single piece flow (or just improved flow in general). 

Why do errors occur? Is it because people don't care? Or don't try hard? 

No. 

Errors occur because people are human, and poorly designed processes and unclear communication make it easy for errors to occur. We often design products, services or processes in such a way as to virtually ensure errors. I see it every day in our organization. No wonder people get frustrated... Employes and customers...

 We did an exercise to try and error proof putting into a cup from ~8' away. Most groups came up with SOPs, visual aids, and the basic funnel design - we tried to eliminate the passive or "don't care" employee factor with a starting chute. But the most interesting solution to this was another group's solution: they removed the human almost entirely by creating a ramp and "gutter" that pointed directly at the cup. Gravity ensured sufficient power to reach the target and the gutter provided accuracy. Human effort required? Virtually nil: simply push the ball onto the top of the ramp. Very clever solution!

So how can the human factors be minimized if perhaps even eliminated? I think to one of our products, Te50, which has a number of point-to-point wiring connections, any of which can be incorrect or loose. Some of the wiring is between a PCBA and a terminal block attached to a controller (essentially). The PCBA provides little to no actual functionality - so what if we could modify the design to eliminate the entire PCBA? Likewise, could other wiring be done with a keyed connector, which would simply snap into place, ensuring correct wiring every time? If a connector costs $2 more to the cost of the wire, as an example, but reduces the labour to wire from 2 minutes to 20 seconds, at $40/hr that 1:40 saved is worth $1.113 (40$/60min x 1.67min = 2/3 x 5/3 = 10/9 = 1.11$). So we are paying maybe $0.49 to reduce the labour time and eliminate a possible defect. Let's also consider that we can now eliminate the time to inspect this wiring - let's be conservative and say that we save 30 seconds - that works out to $0.33. We are down to adding $0.16 to the cost of the product. 

Now if that defect were to make it to our customer, how much does that cost? Thousands and thousands of dollars is what it costs. And that's if the customer's unit doesn't fail in such a way as to require replacement/return for repair. All it takes is one assembler maybe rushing or losing concentration for that defect to show up.

Isn't that worth $0.16?

I think so.

AMac