At this year's Quality Expo in Chicago, two unrelated experiences left me first delighted and then dismayed.

First my editor at Quality tracked me down and told me that my prior blog posts (here, here and here) on data and data value have had a high blog readership in the magazine. I am delighted. Thanks for your interest.

Now the dismay: I ran into an old friend who described a business transaction that left me shaking my head.

My friend told me about a supplier who went to their customer and negotiated a 10% price increase. The customer agreed to it because the supplier was using the increase to fund new high-speed vision inspection equipment. The new equipment would enable the supplier to 100% inspect the product they’re supplying and guarantee that the customer would receive only good parts.

Guarantee! That’s hard to argue with.

Evidently the customer had been very frustrated with this supplier because they had had to put up with a lot of defects. They seemed eager to shift the effort to inspect and sort good from bad to the supplier, and were even willing to share in the cost. They must have felt 10% was a pretty good deal. (Given all the estimates that total cost of poor quality is 30% to 40% of sales, I can see how they could reach that conclusion.)

So why am I shaking my head?

If your definition of good quality is “no bad parts” then this is a perfect solution.

But "no bad parts" is only one definition of quality. And it is the wrong one.

Don’t get hung up on the "parts" language. At the risk of over-simplifying, it doesn’t matter if you’re talking about the diameter of a metal part, the time it takes to close a call in a support center; the amount of yellow ink printed on a magazine cover, the weight of peanut butter in a jar, or the sales of a particular product by a sales rep. All of these are processes, and all have targets and acceptable limits (specifications).

Regardless of the product or the service, "no bad parts" is a poor definition.

Maybe a picture would help:

What the graphic shows is that the green dot is our target and when our output is on target, it is the best that it can be. It also says that the further you get from target, the worse the quality.

"No Bad Parts" says that "Best" and "Fair" are the same. They’re not.

Think about this: A product that is at the yellow dot is closer to the red dot than the green one. It is closer to Unacceptable than it is to Target. When you’re out there in the boondocks of your specifications, you’re a long ways from target, and it doesn’t take much to push you over the edge.

A far better definition of good quality is "on target with no variation."

Twenty-seven years ago W. Edwards Deming published his 14 Points to guide businesses "Out of the Crisis."

Point 3 of Deming’s 14 points was "Cease dependence on inspection to achieve quality." My friend’s story tells me that we’re still depending on inspection to achieve quality. Is there any doubt that we’re still in crisis?

How about you? What examples do you have of inspecting quality into a product? How is it working for you? Leave a comment below and lets get a conversation going.