DMAIC (pronounced “DUH-MAY-ICK”) is the problem solving process central to the Six Sigma improvement methodology. Although it is generally integral to Six Sigma it can be used as a separate framework for problem solving. This article falls within a series of Problem Solving posts and is intended to cover only DMAIC itself, a later post will discuss Six Sigma and hopefully provide a balanced view of that method.
DMAIC is inspired by the earlier improvement method of Plan Do Study Act or the ‘Shewhart cycle’, however while PDSA is more theorethical focused DMAIC is very much practical. The acronym stands for Define, Measure, Analyse, Improve and Control – the 5 phases of the process which should be followed in order to arrive at an improved and controlled process.
Clarify the Problem, set Goals, create a Team.
- What are the specifics of the problem that needs to be solved?
- What is the existing process that needs to be improved?
- What is the scope, what needs to be solved and what doesn’t?
- Who makes up the team that will work on solving the problem?
- How will the improvement be measured?
- What are the targets that you want to achieve?
Define what needs to be measured, collect the measurements, test your measurements
- Define the inputs and outputs of the process (Xs and Ys)
- Create a ‘Measurement Plan’
- Test the Plan
- Collect the Data
- Perform a Gauge Study to test your measurement system
Review your measurements, identify causes of problems, identify the root causes
- Identify the gaps between current output (measurement) and the Goal
- Identify how the inputs (Xs) affect the outputs (Ys) – correlation and causation
- Identify problems preventing you from reaching the Goal
- Prioritize the problems that need to be fixed
- Identify the root causes of those problems
- Confirm the root causes using statistical analysis
Create solutions to fix the root causes of problems and prevent them from re-occurring
- Create solutions (This can vary greatly depending on industry and problem)
- Test the solution, FMEA (Failure Mode and Effects Analysis) is common
- Create an Implementation Plan
- Follow the Plan (sounds obvious right? but… )
Once a process has been improved it must be controlled to maintain improvement
- Implement monitoring systems, statistics based
- Updates process
- Deploy training on the new system
- Update all relevant documents
- DMAIC has a strong Project focus essential to ensure that Goals are defined, Stakeholders are involved and Risks are accounted for
- General approach has allowed it to be applied to a variety of industries and opportunities
- Strongly defined tools and methods, compared to Plan Do Study Act
- Heavy use of Statistics, identifies hard to spot problems and allows verification of causes and solutions
- Although heavily inspired by PDSA, DMAIC lacks the cyclical nature which leads to a common view of improvement as one-off projects
- General approach while widely applicable means that it is not as powerful as very industry/scenario specific methods
- Heavy use of numbers, DMAIC is great for big problems but for small business the full suite of tools & statistics can be prohibitive
- Emphasis is on refinement and reducing variation, great for improving an existing process but doesn’t enable creation of new processes
- DMAIC encourages you to divine Root Cause and empowers you to confirm cause through statistic methods but other methods are more powerful at identifying potential causes (e.g. Kepner Tregoe)