understanding variation

By Merit Smith
Vice President
Director, Health Care Practice

When I met the Dalai Lama he was wearing an orange blanket. I talked with him for a few minutes, and the experience was so unusual that I can’t remember what we talked about.

The word I would use to describe him is “jolly.” He had a twinkle in his eye and a rolling, booming laugh. You could not help but feel his humor. He knew some important things, and he enjoyed talking with people about them. But just because he was talking about things he knew we re important was no reason he couldn’t laugh.

When I met Dr. W. Edwards Deming, he was wearing a business suit. I asked him during a seminar to autograph some books during a break. I don’t think anyone would describe Dr. Deming as jolly, but he seemed to me to be a warm individual who enjoyed talking, one on one, about things that he knew we re important.

I asked Dr. Deming a classic rookie question: “Where should I start?”

He said, “Reading the books might do the trick.”

“Oh, I’ve read them,” I replied.“

Understanding what you’ve read could be useful,” he said, smiling, and I knew he didn’t mean it in a dismissive way. He was enjoying the conversation.

I then asked, “What I really mean is which of your principles is the best one to start on?”

He answered, “Now that you’ve rethought your question maybe you should take a look at ‘scientific management.’ There’s a lot to it.” And that was our conversation.

One way or another, in the 18 years since that conversation I’ve been involved in thinking about scientific management. I’ve read a lot about it, I’ve talked to a lot of smart people about it, I’ve tried to practice it myself and I’ve worked with clients as they practice it.

I’ve come to a point in which I have a fairly clear idea of what Dr. Deming meant when he talked about scientific management.

I think scientific management comes down to two simple ideas that are powerful tools for solving service, cost and quality problems. The first idea is understanding variation. Any process will display variation. Some of the variation is normal and natural, the healthy operation of the interaction of dozens of factors within a process. The normal level of variation that we can expect from a process falls in a predictable range. Let me provide an example.

My cousin, Laura, lives in Dublin, and she takes the Number Four bus to work every morning. It arrives at her doorstep between 8:03 and 8:05. Traffic might be a bit heavy or light. The bus driver may get stopped at a traffic signal, or the weather can affect the speed of the bus. Typically, however, the bus arrives between 8:03 and 8:05.

The arrival has a process characteristic that is the interaction of the design of the process (route and bus), the operation of the process and the operating environment (traffic, weather). It all works together so that on average the bus arrives between 8:03 and 8:05. There is some fluctuation (variation), but it is normal and can be anticipated. But it can’t be managed in any real sense, other than by changing the design of the process or, perhaps, how the driver drives.

The normal bus arrival re p resents process and its characteristic performance and variation. Understanding the nature of variation is the starting point of understanding scientific management.

An important lesson of riding the Number Four is that if you a re part of the system, you can’t really influence the performance of the system. You are truly along for the ride. You can’t reduce the traffic, change the weather or change how the driver drives . In fact, trying to reduce normal variation, other than by thoughtful redesign, will only increase variation. (More about that “thoughtful” part later.) Rookie managers (and unthoughtful executives) try to manage normal variation — and they only make things worse.

One day Number Four was six minutes late and chaos reigned. What happened? The card reader that riders use when they got on the bus malfunctioned and they had to use cash. Riders had to dig into their pockets for coins, causing a bit of a s low line at each stop. This slowed the bus along the entire route. Everyone involved with that bus on that day knew that there was some “special cause” variation involved in their daily commute.

What is important about special cause variation is that it is unusual and explainable and can be fixed by specific action. The driver could provide insight about how to reduce the special cause variation in his process, and management intervention into the system is needed.