DEFINITION AND INTRODUCTION TO DMAIC
DMAIC is a data-driven improvement cycle or framework that breaks down problem solving into five steps as shown in the below diagram. Each step has key activities and tools/templates used to set up and execute a problem-solving activity. This framework can be considered as a stand-alone method but comes from the Six Sigma methodology/initiative and DMAIC is used as the framework for GO Productivity’s Lean Six Sigma Greenbelt course.
DEFINE the problem with a clear problem statement and resulting project charter and other project management plans. DMAIC is data-driven, so the first phase must also involve data collection in the form of collecting the customer feedback or Voice of the Customer. Also you may consider doing a high-level Value Stream Map to know how this problem fits holistically into the whole organization.
MEASURE the process / problem first by conducting a process mapping exercise and collecting baseline data. This will be the ‘before picture’ that most of us forget to take before we rush into a project.
ANALYZE the data to reveal the true causes for the problem that was defined. There are various tools available including Failure Mode and Effects Analysis, Design of Experiments and other graphical and statistical tools.
IMPROVE the process / problem by following the results of the analysis to address and eliminate the root causes. Additional tools available include conducting Kaizen events.
CONTROL the refreshed process and keep collecting data to validate the expected benefits. Additional tools for process control apply including Quality Control Plans, Statistical Process Control, 5S and Mistake-Proofing or Poka-Yoke.
All of these tools are covered in detail in our Lean Six Sigma training and are applied on real-life projects for the Greenbelt and Blackbelt levels to ensure learning through applied knowledge and experience.
Voice of the Customer – feedback directly from the customer on what is most important and what is ‘nice-to-have’ or ‘delighters’ for the product or service.
Failure Mode and Effects Analysis (FMEA) – a thorough step-by-step approach to collecting all potential risks and ways that a process or product could fail. Failure modes are what happens to the product or process that is considered a fail and the effects are the resulting consequences of that failure. This usually populates a large spreadsheet and can then be used to prioritize and analyze to create plans for preventing failures.
Design of Experiments – using statistics to design controlled experiments targeted at determining root causes when there are multiple potential variables which would be too exhaustive to test each individually.
Kaizen Events – Kaizen is typically translated to ‘continuous improvement’ from Japanese. The Kaizen event is a micro process improvement project happening typically within seconds, minutes or hours. A team comes together at the place where the work is done and makes a small change and immediately validates if the change improves the process. These small, gradual changes are the reason for calling it a Kaizen event or continuous improvement event.
Quality Control Plan – documentation on quality specifications, standards, and practices for maintaining quality within expected limits for a product or service.
Statistical Process Control – using graphical and statistical tools to monitor and control process performance variables. For example, using historical performance data to determine acceptable limits on a machine and then creating control charts to visualize when it goes above or below acceptable limits.
5S – a 5 step process for refreshing a process or group of processes promoting cleanliness and orderliness: Sort, Set in Oder, Shine, Standardize and Sustain. The 5 S’s are based on Japanese words that also starting with S: Seiri, Seiton, Seiso, Seiketsu and Shitsuke.
Mistake-Proofing or Poka-Yoke – involves implementing a device or process that prevents an error from occurring and makes the error immediately obvious if it does happen.