Six Sigma Principle One: Recognize Determinism - dummies

Six Sigma Principle One: Recognize Determinism

By Craig Gygi, Bruce Williams, Neil DeCarlo, Stephen R. Covey

A fundamental element of the Six Sigma approach is the ability to recognize determinism. If you’re like most people in Western society, you’re results-oriented. You ask questions such as “How did it turn out?”, “What finally happened?”, “What was the final score?”, “How long did it take?”, and “What’s the bottom line?” Everyone is always looking at the results.

But how did the results happen? Why did they happen? What specifically caused them to happen? You want to know the answers to these questions because, if good things happen, you want to know how to make them happen again. And if bad things happen, you just as surely want to know how to prevent them next time. This focus on the process stems from determinism.

Determinism is the principle that you can create a desired outcome by configuring and controlling the inputs in a specific manner. In Six Sigma, you analyze the inputs and the process and then implement the best possible combination to achieve your objective. By doing so, you exercise direct control over your environment instead of allowing your environment to control you. You’re deterministic, not reactionary, in your thinking.

Many companies and organizations want to improve their performances. They recognize intuitively that their performance results are the outcome of all their business and work processes. These processes are quite literally “the way” business is done.

So, to improve the outcomes, a company has to change the way it does business. It wants to change the processes — the function f in the breakthrough equation — and combine the business inputs in a way that produces a better outcome. It’s not a wishful notion and not a trick for getting everyone to work harder.

Cause and effect

To know how you came to the result you have, you have to examine your deterministic process. Look behind every result and examine the inputs, the process, and the error that combine to produce it. When you know what causes the outcome, you can begin to position yourself to control the outcome next time. Understanding the root cause-effect relationship is the first step to controlling outcomes.

Here’s a simple example: The guy gets the girl. That’s the outcome. How did that outcome happen? Well, as all razor companies would have you know, it’s because he has a smooth, sexy face. What caused that?

It’s the result of his shaving process and choice of ingredients — combining hot water, shaving cream, a mirror, a particular razor, and a steady hand to get a close shave. If anything’s wrong with the outcome — if the girl thinks the guy’s face is too scruffy despite the shave — you examine the ingredients and shaving process to determine the root cause of the problem.

Regardless of complexity, literally every result has one or more causes. The more you can single out these causes and understand them, the better your opportunity to change them for the better. In Six Sigma speak, you’d say that knowing the Xs, the function f, and the amount of uncertainty ε means you know what caused the outcome Y. Cause and effect.

Correlation doesn’t imply causation

Be careful not to confuse coincidence with cause and effect. Just because two events happen together doesn’t mean that one caused the other. People often assume that events that are closely connected — either spatially or temporally — are somehow also connected causally. These mistaken assumptions are called superstitious delusions (the Latin term is non causa pro causa, which means “non-cause for the cause”).

Superstitious delusions are why a football coach who once won an important game while wearing red socks now wears them for every game. Did the socks cause his team to win, or did the win stem from some other input or inputs?

Businesses are just as likely as superstitious coaches to confuse coincidence with causation. Think about the company that ramps up capacity after just a single great month of sales because it believes this sales boost indicates a market expansion. Only later does the company discover that no expansion was forthcoming and that the increased sales were correlated to a different factor.

Even if two variables are legitimately correlated, they don’t necessarily have a causal relationship. One may fluctuate in relation to the other due solely to chance. Or each variable may be strongly affected by one or more other outside (or confounding) variables that haven’t been identified yet.

However, a causal connection probably does exist if you can establish all three of the following conditions:

  • A reasonable explanation exists for cause and effect.

  • The connection happens under different environmental conditions.

  • You’ve ruled out potential confounding variables.

One way to determine these conditions is through a designed experiment where you expose groups strongly similar to one another in terms of the most important variables to different conditions and then analyze them to see whether the variable of interest performs differently. One or more control groups are also held constant and not subjected to treatment(s).