GM Got Gamed

Following up on my post about my recent experience metrics and processes being distorted (and my less-than-perfect lean coaching efforts), I was thinking about to some first-hand experience I had at a GM engine plant, circa 1995. It’s the most blatant example of someone intentionally distorting data that I’ve ever seen.

Our engine block line was designed at a throughput goal of 92 blocks per hour, if everything ran perfectly (but it hardly ever did, at least not for an entire 60 minute stretch). Our plant superintendent, Bob, (the #2 guy in the plant) decided that 60 pieces per hour was an acceptable number (partly based on productivity numbers that were attributed to Toyota). Anything below 60 and you’d have to explain why. Now, he wasn’t really the listening, problem solving type. He managed by fear, yelling, and intimidation. There was more yelling involved than listening or problem solving.

Anyway, at the end of the engine block line was a mechanical counter that recorded the hourly production counts. The UAW workers who unloaded blocks dutifully recorded the number every hour.

The numbers might typically look like (for you RSS readers, you might not see the image):

That’s an average of 56.25 pieces per hour. Not quite up to Bob standards, although we exceeded the goal in 4 hours and came close to 92 in one hour.

At the end of the day, before our “4 o’clock meeting” where the plant salaried staff took its daily verbal beating from Bob, Scott, the production supervisor (technical title of “Team Coordinator” didn’t quite fit) would pick up the counts and do a little daily editing.

He’d take the numbers and turn them into, I kid thee not:

That’s still an average of 56.25 per hour. But it’s much more consistent, isn’t it?

Bad ole’ Bob never questioned these numbers. I know it’s hard to believe that he would believe those numbers, but when reviewing multiple departments, Scott’s fudgery helped avoid too much attention that a really bad hour would bring upon him. Rather than asking “why don’t we have more hours of 86 blocks?” the upper limit of expectations was set too low, at 60. I asked Scott once why he fudged the numbers each day, and his answer was:

“Bob wants 60 an hour, he gets 60 an hour.”

Other departments got more than their share of the daily beatings (I had a bet with a co-worker if Bob would say “pathetic” or “miserable” first). Bob always had the same pronouncement for our problems: we weren’t trying hard enough. And apparently, more yelling from Bob was what we needed to motivate us. But that never worked.

“Not trying hard enough” fell into two categories: 1) urgency and 2) intensity. We didn’t have a sense of urgency. We didn’t have the proper intensity. Like a shorter imitation of Mike Ditka (with a signature bad toupee rather than a signature mustache), Bob would yell and scream and spit. Sometimes we got “we need urgent intensity” or “we need intense urgency” if things were really bad. All of the yelling and screaming, all of the fear, all of the fudging of the numbers got in the way of true process improvement and true problem solving.

Obviously, situations like this are part of the reason our plant manager eventually got moved out of the way for a new, NUMMI-trained plant manager. That started our road to recovery, as a plant. It was never a worker problem, it was a management problem. That’s an important lesson of lean — what’s required is a change in management practices and management philosophy.

I’ll leave it for another post to talk about that “4 o’clock meeting” and what its goals were supposed to be. The meeting was designed by some internal lean consultants we had, but was co-opted for non-lean management methods. Why weren’t the lean consultants being listened to? Again, I’ll save that for another post.

Illustration from igotzillustration.com

About LeanBlog.org: Mark Graban is a consultant, author, and speaker in the “lean healthcare” methodology, focused on improving quality and patient safety, improving access, reducing costs, and fully engaging healthcare professionals. He is also the Chief Improvement Officer for KaiNexus.