Lean Six Sigma: DMAIC Principles and How to Put Them Into Action
“Give me a D! Give me an M! Give me an A-I-C! What’s it spell? DMAIC!” You probably won’t see any cheerleaders in your office cheering on your staff’s efforts at continuous improvement—but maybe you should. Those actions are critical to the success of your Lean Six Sigma efforts, and they warrant steady encouragement.
DMAIC stands for Define, Measure, Analyze, Improve, and Control. All five of these elements are important to maintaining the quality and continuous improvement efforts that lie at the heart of Lean Six Sigma philosophy. Unfortunately, it’s easy to lose focus and shortcut one or more of the elements of the DMAIC continuous improvement process. When that happens, the process fails in the same way that one broken connection can take out a complex electrical circuit.
Let’s take a look at each element in a bit more detail.
The definition stage starts off by determining your customer’s needs and goals, along with what it’s going to take for you to fulfill them. These critical characteristics are called critical-to-quality (CTQ) and can range from part dimensions and tolerances to packaging and delivery concerns. Usually a voice-of-the-customer (VOC) survey or similar method is used to clarify all the necessary CTQs and a formal acknowledgement is created for both sides to review and agree upon.
However, the definition stage should constantly be evolving. As you get more feedback from your customer in terms of needs and requirements or from your processes in terms of capability and defect prevention, the new information must be processed, understood, and fed back through the DMAIC cycle.
Too often an initial DMAIC cycle is completed with a great sigh of relief, and critical new information is ignored over time. The process slowly breaks down. People then question the principles of DMAIC instead of realizing that they aren’t being applied correctly. It’s important as a manager to stress the “continuous” in continuous improvement. Demand regular feedback to verify that DMAIC is operating in a loop as it should.
Determine the proper measurements to make based on the CTQs and how often those measurements need to be made. Make sure that your measurement techniques and results are documented by work instructions and that the measurement tools have sufficient accuracy and precision to perform the task.
Outline the process, then identify and assess the most likely modes of failure through a Failure Mode Effects Analysis (FMEA). Set up initial control plans based on the best information to date.
Keep your measurements focused. Measuring extraneous things just to have the data in your back pocket is not terribly useful. It builds up huge amounts of unnecessary data, wastes time and resources, and can lead to data processing overload, spurious data connections, and the infamous “paralysis by analysis.”
If measurements aren’t giving you the information you need, go back to the definition step and determine what you really do need to measure and how often you need to measure it. Skip straight to analysis with the wrong set of measurements to identify and solve the problem, and you’re likely to also jump to a wrong conclusion.
Once data is gathered, you can begin analyzing the data for areas of potential improvement. Be careful to present all data in a factual and non-misleading format. The scales and type of graph chosen can introduce a subtle bias.
Process control and capability information can be combined with a fishbone diagram to determine the most likely root causes of variations or defects. You may find causes unrelated to your original assumptions during the FMEA phase. This approach leads to data-driven fact-based conclusions instead of hunches and guesswork.
You should exit this phase with clear and prioritized ideas on the areas for improvement based on providing the greatest value to the customer, as well as the most likely root causes of the underlying problems. Of course, anything that results in failure to meet customer specifications must be dealt with immediately.
Now that the root causes have been identified, it’s time to deal with them. The proper technique to use depends on the type of problem, but the general procedure is “plan-do-check-act” (PDCA). Outline a focused plan to address the identified problem, perform the necessary testing, check the results, and act on those results—either moving on to the control stage to implement that change or back through the improvement cycle until the root cause is verified.
Unless there is a logical reason to do so, resist massive testing to solve multiple problems at once. That doesn’t mean a comprehensive design of experiments (DOE) approach is not useful, but it should be focused on the most important issues. “Mission creep” makes it more likely that the results will be confounded and it will take longer to find the solution.
A well-constructed control plan formally implements the solutions from the previous step and verifies that the results are sustainable. After a suitable period of time with positive feedback, that problem is considered solved and feedback is provided for the next DMAIC loop. The results should be periodically audited to verify the solution over the long term.
After each step, it’s wise to hold a simple review to verify that the important tasks have been completed and that everyone agrees on what has been learned and what the next actions should be.
Keep the Cycle Going
Momentum and focus are important components of a continuous improvement process, and we at Smartrend take pride in our implementation of the DMAIC process to produce high quality parts to our customer’s specifications. Let us put Lean Six Sigma principles to work in fulfilling your contract manufacturing needs.
Meanwhile, if you need cheerleaders to keep the momentum going and focus on your Lean Six Sigma efforts, by all means use them. Do whatever it takes.