Once upon a time, in a real, live U.S. corporation, top management decided that Six Sigma was a good idea. They trained up many Black Belts to lead projects to produce documented savings as defined by the existing accounting system. To encourage people to be successful, the Black Belts—not the project teams—were slated to receive substantial bonuses for producing these savings.One such project involved a Black Belt who found that many of the tests on one particularly complex product were duplicates of one another. He eliminated the duplicate tests and was able to claim $800,000 in annual savings. He was a happy man. He got a nice bonus.

Here’s the rest of the story. It just so happens that the eliminated tests were automatically run in parallel with other tests. Production was based on a constant week-to-week contract, so running an inventory ahead was very risky. Labor was unchanged. Scrap was unchanged. Production was unchanged. Our happy Black Belt got a nice bonus for savings that didn’t actually occur. Not his fault. Systems error.

Two lessons come to mind. When I was with Ford Motor Co. and we started our Employee Involvement/Participative Management effort many years ago, we struggled with how to make cost-reduction projects really work to improve the bottom line. We could focus on reducing scrap, because then we could purchase less material and really save money. But when we got to labor savings, it got a little sticky.

You see, the people who really knew where the opportunities were, the hourly workers, were the very people who would have to be laid off to save real money. The dilemma was obvious. No one is going to help save money for the company if it were to result in their own job loss.

In those days, we had a natural attrition rate of about 6 percent. We figured if we were willing to retrain the workers whose jobs would be lost from a savings project and simply take advantage of our natural attrition rate, we could get a 6 percent annual profit-improvement rate and a more productive workforce. As a matter of fact, we committed to lay off no one as a result of an improvement project. Well, it worked.

Today the world is different. The rules we made up may no longer work, and the rules in this Six Sigma story surely aren’t working too well. So you’d better figure out a way for everyone to benefit from your Six Sigma efforts if you want them to work effectively.

The second lesson is a broader, more articulate version of the first—Myron’s Perversity Principle:

"If you try to improve the performance of a system of people, procedures, practices and machines by setting goals and targets (and incentives) for the individual parts of the system, the system will defeat you every time and you will pay a price where you least expect it."

Be careful how you encourage Six Sigma success. People are complex creatures.