A Complete Guide to the DMAIC Improvement Cycle in Six Sigma
Summary: The DMAIC improvement cycle is a core tool within the Six Sigma methodology for improving business processes. It is an in-depth, data-driven cycle that offers a good framework for addressing even the most complex business processes. In this guide, I will provide a definition for each of the five components of the DMAIC continuous improvement cycle as well as a step-by-step checklist for each phase so you can make sure you are not missing any important steps in the DMAIC project process.
If you would like the checklist I lay out below in spreadsheet format, I have a free DMAIC Checklist template you can use within Google Sheets, so make sure to check that out.
Before We Get Started… Can You Use DMAIC Even If Your Company Does Not Follow Six Sigma?
The DMAIC process is often described as a key tool within the Six Sigma methodology, however I believe that the principles of DMAIC can be beneficial for any company, department, or project team who is looking for a framework to follow to successfully implement a process improvement practice. There are no parts within the DMAIC methodology that require a Lean Six Sigma or Six Sigma project integration, and you certainly don’t need a Six Sigma Black Belt or Green Belt certification to use this improvement process. So by all means, use this guide whether or not you are part of a Six Sigma practice.
The Five Stages of the DMAIC Improvement Cycle
The DMAIC framework includes five stages in the existing process improvement cycle. They are:
Define
Measure
Analyze
Improve
Control
These five steps provide a systematic framework for moving Six Sigma projects from initial stages all the way to completion. The DMAIC model is useful for new processes and existing processes.
Next, I am going to provide a definition for each step and a checklist of items that you need to consider before advancing onto the later stages in DMAIC. It is important that these stages are done in order and are fully completed before moving forward to ensure a successful process improvement project.
The Define Phase
The first step in the DMAIC improvement cycle is DEFINE. This is where you define the business process that maybe be a problem and in need of improvement. The end result of this phase should ultimately provide you with the answer the question: “What is the real problem that must be addressed?”
Often the answer is not obvious because the problem that is expressed may only be a symptom of an underlying and more complex problem. Therefore, the Define phase may take some time to fully complete. To ensure complete understanding of the Define stage, consider the following checklist.
The DEFINE Checklist
If you can respond “Yes” to each statement below, you’re off to a great start with your project, and are ready to move to the “Measure” phase of DMAIC.
For this project we have...
Confirmed that our project is a worthwhile improvement priority for our organization and is supported by business leaders.
Been given (or written) a brief Project Rationale explaining the potential impact of our project on customers, financials, and its relationship on the company's business strategies.
Composed and agreed to a 2-3 sentence description of the problem as we see it, the Problem Statement, focusing on symptoms only (not causes or solutions).
Prepared a Goal Statement defining the results we are seeking from our project, with a measurable target. No solutions are proposed in the Project Goals Statement.
Prepared other key elements of a DMAIC team charter, including a list of constraints and assumptions, a review of players and roles, a preliminary plan and schedule, and a process scope.
Reviewed your project charter with your sponsor for this project and confirmed their support.
Identified the primary customer and key requirements of the process being improved and created a SIPOC diagram of the areas of concern.
Prepared a detailed process map of areas of the process where we expect to focus our initial measurement.
If you have successfully completed all of the above components of the Define stage, it is time to move on to Measure!
The Measure Phase
The second phase of the DMAIC improvement cycle, Measure, requires that measurements be made to better understand the current state of the problem you defined in the first stage. Measurements require data. Sometimes the data is readily available through baseline measurements you already have, but other times it may have to be collected through observation, accessed by retrieving data stored in a database or system, or by administering questionnaires.
The checklist below will help guide you through the process of the Measure stage.
The MEASURE Checklist
If you can respond yes to each statement below, you're doing well with measurement, and are ready to move on to the Analysis phase of DMAIC.
For this project we have…
Determined what we want to learn about our problem and process and where in the process we can go to get the answer.
Identified the types of measures we want to collect and have a balance between effectiveness/efficiency and input/process/output.
Developed clear, unambiguous operational definitions of the things or attributes we want to measure.
Tested our operational definitions with others to ensure their clarity and consistent interpretation.
Made a clear, reasonable choice between gathering new data or taking advantage of existing data collected in the organization.
Clarified the stratification factors we need to identify to facilitate analysis of our data.
Developed and tested data collection forms or check sheets which are easy to use and provide consistent, complete data.
Identified an appropriate sample size, subgroup quantity and sampling frequency to ensure valid representation of the process we are measuring.
Prepared and tested our measurement system, including training of collectors and assessment of data collection stability.
Used data to prepare baseline process performance measures, including proportion defective and yield.
If you have successfully completed all of the above components of the Measure stage, it is time to move on to Analysis!
The Analyze Phase
The third step focuses on analyzing the data to better understand the defined problem, to further reveal its root cause, and point to a solution. Analysis may involve the use of spreadsheets, process maps, charts, or statistical analysis. For example, process maps may narrow the problem to one specific step in a process. Statistical analysis may be useful in comparing process solutions to determine if one is better than the other. Analysis may also lead to solutions by pointing to inefficiencies in a process or even identifying unnecessary steps.
The checklist below will help guide you through the process of the Analyze stage.
The ANALYZE Checklist
If you can respond yes to statements #5 or #7 below, and if you have done many of the tasks described in the other statements, chances are good you are ready to begin developing solutions in the Improve phase of DMAIC.
For this project we have…
Examined our process and identified potential bottlenecks, disconnects and redundancies that could contribute to the problem on which we are focusing.
Conducted a value and cycle time analysis, locating areas where time and resources are devoted to tasks not critical to the customer.
Analyzed data about the process and its performance to help stratify the problem, understand reasons for variation in the process, and identified potential root causes.
Evaluated whether our project should focus on process design or redesign, as opposed to process improvement, and confirmed our decision with the project sponsor.
For Process Design/Redesign:
Ensured that we understand the key workings of the process so we can begin creating a new process to meet the needs of the customer effectively and efficiently.
For Process Improvement:
Developed root cause hypotheses to examine the problem we're solving.
Investigated and verified our root cause hypotheses, so that we are confident we have uncovered one or more "vital few" root causes that created our problem.
How you complete the Analyze phase checklist will depend on your unique project, but if you are comfortable with the answers you provided above, then you are ready for the Improve phase!
The Improve Phase
With analysis complete, the next step is to actually take the steps to improve the process. You made it! Depending on the nature of the problem, this can be as simple as eliminating steps, purchasing a new tool, or as complex as redesigning almost every step in a process. It is important to note that this can be a difficult phase because behavioral issues may interfere. Often those attempting to change or improve a process must contend with resistance to change where those involved in the old process resist changes to a process that has become routine and familiar.
The checklist below will help guide you through the completion of this phase.
The IMPROVE Checklist
If you can respond yes to each statement below, you've achieved success with your improvement, and are ready to plan to the Control of your process or solution.
For this project we have…
Created a list of innovative ideas for potential solutions.
Used the narrowing and screening techniques to further develop and qualify potential solutions.
Created a solution statement for at least two possible proposed improvements.
Made a final choice of our solution based on success criteria.
Verified our solution with our sponsor and received buy-in and the go-ahead.
Developed a plan for piloting and testing the solution, including a pilot strategy, action plan, results assessment, schedule, etc.
Evaluated pilot results and confirmed that we can achieve the results defined in our Goal Statement.
Identified and implemented refinements to the solution based on lessons from the pilot.
Created and put in place a plan to expand the solution with refinements to a complete implementation.
Considered potential problems and unintended consequences of the solution and developed preventive and contingent actions to address them.
By completing the Improve checklist, you owe you and your team a big applause. This truly was the main phase for why you kicked this project off in the first place. Now it is time to set up the components of a control plan and track your new process improvement.
The Control Phase
One of the fundamental challenges faced by every organization is to establish monitoring and control systems to ensure that process improvements remain in place and conform to expected levels of performance. In this final stage, actual process results are compared with expected results and when actual results fail to conform to expectations remedial action is taken.
The checklist below will guide you through the process of creating and monitoring the Control phase of DMAIC.
The CONTROL Checklist
If you can respond yes to each statement below, you've completed all key steps in your dimak project and are ready to celebrate and maintain your improvement.
For this project we have…
Compiled results data confirming that our Improvement has achieved the goal defined in our DMAIC team charter.
Selected ongoing measures to monitor performance of the process and continued effectiveness of our solution.
Determined key control charts and graphs for a process scorecard on this process.
Prepared all essential documentation of the revised process, including key procedures and process maps.
Identified an owner of the process who will take over responsibility for our solution and for managing continuing operations.
Developed, with the process owner, Process Management charts detailing requirements, measures and responses to problems in the process.
Prepared a storyboard documenting the team's work and data collected during the project.
Forwarded other issues / opportunities which we were not able to address to Senior Management.
Celebrated the hard work and successful efforts of our team.
Congratulations! You have successfully completed all the phases of the DMAIC improvement cycle. Now let’s move on to some other important topics relating to DMAIC.
Measurement Is Critical Across the Entire DMAIC Cycle
Measurement, the second step in the DMAIC framework process, is critical not only during the initial stages when a problem is first defined but also later, when measurements must be taken during the Control stage to ensure that an operational process is performing as expected. Clearly, quite a bit depends upon measurement and the data collection plan. It is central to data collection during the Define stage and it is central to the Control stage. What we would like to believe is that we have confidence in our ability to measure what we observe or that the data we collect is accurate. Yet this is not necessarily true. If it isn't, we have a real problem, the entire Six Sigma process in which we engage may not produce the desired results.
Our ability to measure is actually limited in several ways. First, it is limited by the accuracy of the tools we use to take the measurement. Second, it is limited by the human errors we make while taking measurements.
Limitations in DMAIC Tools for Measurement
The tools we use to measure an outcome may be less than perfect. For example, a questionnaire to determine customer satisfaction may not accurately detect proper levels of satisfaction. In addition to imperfect tools, there is human error. Two individuals monitoring the same process may reach different conclusions. For example, two marketing managers reviewing the same resume may reach very different conclusions, one may recommend that the individual be hired, the other may conclude that the person is unqualified for the job.
The Gage R&R Study
Since measurement problems are not uncommon, and since the Define and Control stages rely on measurements, a formal analysis to uncover the extent of the measurement problem is often conducted. It is called a Gage R&R study.
A Gage R&R study focuses on two dimensions of the measurement process. First, it focuses on whether an individual collecting the data or performing the tests is consistent across many measurement tools. Second, it focuses on whether the variation between different people involved in data collection is consistent.
The first focus, consistency associated with a single individual, is called repeatability. The second focus, consistency across individuals performing the same task, is called reproducibility. This is the basis for the term Gage R&R, gage repeatability and reproducibility.
Precision and Accuracy
Measurement systems must be both accurate and precise. It is important to distinguish between these two words.
Precision refers to the closeness of two or more measurements to each other. If they are very close then it might be thought that the data are truly representative of the process or parts being measured. Yet this measurement by itself may not be sufficient.
Accuracy is the proximity of measurement results to the true value. It means that the data are indeed a true representation of what it is we are trying to measure.
Measurements in a data set can be precise and not accurate, they can be accurate and not precise, and they can be both accurate and precise.