PDCA: Predict-Do-Check-Act?

Let me start out by saying that, contrary to the title, I am not really advocating changing the name of the PDCA cycle to “Predict-Do-Check-Act”. What I am really pushing for is that part of the planning cycle includes a prediction of what will happen when you make a change.

If you recall back to high school, one of the steps of the scientific method is to form a hypothesis. And what is PDCA if not a specialized version of scientific, analytical thinking?

An action plan requires two main assumptions. The first is an assumption about the root cause. As data intensive and structured as you root cause analysis may have been, in the end, it is an educated guess. And the action item on your plan is yet another SWAG. So, there are two degrees of uncertainty in an improvement plan.

What happens, though, is that teams will make a change, and might get, say, a 3.6% improvement as a payoff for their efforts. They are happy with that effort, and move on. So what’s the problem? The problem is that any change has an opportunity to affect the output. Most planners will be cautious about making changes that will obviously hurt an operation, so generally, action plans have a better chance of bringing about positive results than negative ones.

Now you can look at that with the attitude that you made an improvement, so why complain. Or, you can look at it with the attitude that you want to get the biggest return possible out of your improvement efforts. Further, that 3.6% improvement might have improved something closely related to the issue you were trying to address, and your original problem might still be present.

Think of the power of compounding. When you put money into an investment, the returns create additional future returns for you. The same happens with continuous improvement. Compare a series of 3 improvements that yield 3.6% improvement with 3 that yield 4.6%. The first operation will see an 11.2% gain. The latter will have a 14.5% improvement. And that is just for three little projects. In the real world, the gains built upon gains can be much greater.

So how does predicting outcomes help? Well, it is a way for you to confirm that your analysis is correct. If you anticipate a 5.7% improvement, and end up with a 3.4% or a 9.5% gain, you made a mistake somewhere. It is important to know that so you can improve your problem solving. The check step needs something specific to check against, not just whether the change made things better.

Try this next time you make an improvement: Predict the impact your changes will have, and then confirm how close you are to your estimates. If you are off, take action to correct your improvement process. Yep. That’s right. Predictions let you use PDCA to improve how effective you are at using PDCA.