DMAIC is a five-step method for fixing a process that already exists but misses its targets: Define the problem, Measure how the process performs now, Analyze the data to find root causes, Improve by testing and putting in solutions, and Control the process so the gains last. ASQ describes it as the method for improving existing processes that fall short of performance standards or customer expectations. Below, you will find what happens in each phase, the tools that go with it, a practical example, and the mistakes that sink most projects.

TL;DR

• DMAIC stands for Define, Measure, Analyze, Improve, Control. Use it when a process exists, it is underperforming, and you can measure it.
• The order matters. Each phase uses the data and decisions from the one before it, so skipping ahead usually means solving the wrong problem.
• Define sets the goal and scope, Measure sets the baseline, Analyze finds root causes, Improve tests and puts in fixes, Control locks the gains in.
• Control is where most projects quietly fail. A control plan, standard operating procedures and ongoing monitoring keep the process from drifting back.
• If you plan to lead DMAIC projects professionally, a structured certification gives you the vocabulary and tools in one place.

Table of Contents

• What DMAIC is and when to use it
• The five DMAIC steps at a glance
• Step 1: Define
• Step 2: Measure
• Step 3: Analyze
• Step 4: Improve
• Step 5: Control
• Why the order matters
• A practical example: reducing scrap in exhaust tubing
• Common mistakes in DMAIC projects
• Our perspective: what we see when people learn DMAIC
• Build the skill with a certification
• FAQ
• Sources

What DMAIC is and when to use it

DMAIC (pronounced “duh-MAY-ick”) is the core problem-solving cycle of Six Sigma and Lean Six Sigma. It turns a vague complaint such as “our orders ship late” into a measured problem, a proven cause and a fix that stays fixed.

Use DMAIC when:

• The process already exists and runs repeatedly (order fulfillment, claims processing, patient discharge, machining a part).
• Performance is below a standard or below what customers expect.
• You can collect data about the process, even if you have to build the measurement first.
• The cause of the problem is not obvious. If the cause is known and the fix is simple, just fix it.

DMAIC is not for building something that does not exist yet. Six Sigma practitioners use a different roadmap, usually called DMADV, for designing a new process or product. DMAIC is also not limited to Six Sigma: the framework works as a standalone improvement procedure and is used well beyond manufacturing, in healthcare, finance, government and retail.

The US Environmental Protection Agency notes that DMAIC resembles the Plan-Do-Check-Act cycle, working as a structured method that guides how statistical tools are applied. If you already know PDCA, DMAIC is the same logic with more rigor on measurement and root cause.

The five DMAIC steps at a glance

The tool lists come from ASQ’s DMAIC overview. Nobody uses every tool on every project. Pick the ones the data calls for.

Step 1: Define

Define answers a deceptively simple question: what exactly are we trying to fix? In this phase the team sets project goals, identifies the customers and their requirements, and writes a project charter that fixes the scope and focus.

A good charter contains:

• Problem statement. One or two sentences, with a number. “Invoices take too long” is weak. “Only 71% of invoices are paid within 30 days, against a target of 90%” is a problem a team can work on.
• Goal statement. A measurable target and a date.
• Scope. Where the process starts, where it ends, and what is off limits.
• Team and sponsor. Who does the work and who clears obstacles.
• Customer requirements. What the customer actually needs, gathered through voice-of-the-customer methods like interviews and surveys, then translated into measurable terms, often called critical to quality (CTQ) requirements.

Most teams also draw a high-level process map, such as a SIPOC (suppliers, inputs, process, outputs, customers), so everybody agrees on what the process is before anyone measures it.

The human side matters here too. You need to interview customers and process users, and you need to present the case to a sponsor. Practicing how you would frame a project to a hiring panel, using project manager interview questions, is a useful way to sharpen the skill of stating a goal, scope and stakeholders clearly.

For a deeper walk-through, see our guide to the Six Sigma DMAIC Define stage.

Exit check: Could a stranger read your charter and tell you what success looks like, by how much and by when? If not, you are not done defining.

Step 2: Measure

Measure establishes the baseline. You document how the process actually runs, not how the procedure manual says it runs, and collect data on how well it performs.

Typical activities:

1. Map the process in detail. Walk the process (the Lean term is “go to the gemba”) and record each step, handoff and delay.
2. Choose what to measure. Pick output measures tied to the customer requirement (defect rate, cycle time, cost) and process measures that might drive them.
3. Check the measurement system. If two people measure the same item and get different answers, your data cannot be trusted. Fix that before collecting more.
4. Collect data. Use a plan that says what, how much, from where and by whom.
5. Calculate baseline performance. Capability analysis shows how well the process meets the customer’s specification. Pareto charts show which categories of defect account for most of the problem.

This phase is where many teams discover that nobody has been tracking the thing they were asked to improve. That is a finding, not a failure. Building a reliable measurement is often half the value of the project.

See the Six Sigma DMAIC Measure stage for more detail.

Exit check: You have a baseline number you trust, and you know where in the process the defects or delays cluster.

Step 3: Analyze

Analyze turns data into explanations. The team identifies the critical inputs and root causes that drive the problem, and separates real causes from plausible-sounding ones.

Two moves usually happen in sequence:

• Generate hypotheses. Fishbone diagrams, 5 Whys and process-map review produce a list of possible causes. FMEA (failure mode and effects analysis) ranks potential failures by severity, likelihood and detectability.
• Test them with data. Multivariate charts, regression, hypothesis tests and stratified Pareto charts show which suspected causes actually relate to the output. A cause that cannot be shown in the data stays a theory.

A common example: a call center believes long hold times come from understaffing. Analysis of the call data shows that the longest holds cluster in the 20 minutes after a billing email goes out, and that most of those calls concern one confusing line on the bill. The cause is the bill, not the headcount. Adding agents would have cost money and fixed nothing.

Our article on the Six Sigma DMAIC Analyze stage covers the statistical tools in more depth.

Exit check: You can state the vital few causes, show the data behind each, and explain how fixing them should move the baseline number.

Step 4: Improve

Improve is where solutions are generated, evaluated and put in place. The goal is to optimize the process by acting on the root causes you proved in Analyze, not on symptoms.

Good Improve work follows a pattern:

1. Brainstorm solutions with the people who run the process. They usually know what will and will not work.
2. Evaluate them against impact, cost, effort, risk and fit with customer requirements. A simple solution-selection matrix works well.
3. Test before rolling out. Run a pilot on one shift, one site or one product line. Designed experiments (DOE) let you test several factors at once and find the settings that work best.
4. Compare results with the baseline. Use the same measurement you built in Measure. If the pilot does not move the number, go back to Analyze.
5. Implement and train. Update the process, tell the people affected, and train them on the new way of working.

Kaizen events, short focused workshops where a cross-functional team redesigns part of a process, are a common way to carry out the changes in Improve.

Read more in the Six Sigma DMAIC Improve stage guide.

Exit check: Pilot data shows a measured improvement against the baseline, and the team has a plan to scale it.

Step 5: Control

Control keeps the gain. Without it, processes drift back to old habits within months, and the project becomes a slide in someone’s presentation rather than a lasting result.

The control toolkit includes:

• Control plan. A document saying what is monitored, how often, by whom, what the acceptable range is, and what to do when a measure goes out of range (the reaction plan).
• Standard operating procedures. The improved method written down and used for training, so it does not live only in one person’s head.
• Statistical process control (SPC). Control charts that separate normal variation from signals that something has changed.
• Mistake-proofing (poka-yoke). Design changes that make the error hard or impossible to make, such as a form that will not submit with a missing field.
• 5S and visual management. An organized, standardized workplace that makes deviations easy to see.

Control also includes a formal handoff. The project team transfers ownership to the process owner, who is accountable for the results from then on. Document the savings or improvement, share lessons learned, and close the project.

The Six Sigma DMAIC Control stage page goes through control plans and charts step by step.

Exit check: The process owner, not the project leader, is watching the metric, and there is a written plan for what happens if it slips.

Why the order matters

DMAIC follows a strict sequence because each phase builds on the data and foundation of the one before it. ASQ lays out the phases in this order for that reason.

Here is what happens when teams break the sequence:

• Skipping Define gives you a project nobody agrees on and a scope that grows every week.
• Skipping Measure means you cannot prove improvement, because you never knew where you started.
• Jumping from Measure to Improve produces solutions in search of a problem. They are often expensive and sometimes harmful.
• Skipping Control lets the gains erode and makes the next project start from scratch.

That said, DMAIC is not a straight line in practice. Analysis often sends a team back to Measure for more data, and a failed pilot sends it back to Analyze. Going back is part of the method.

A practical example: reducing scrap in exhaust tubing

A well-documented case comes from automotive supplier Tenneco. In a published white paper, teams used DMAIC to reduce scrap, rework and costs in exhaust pipe manufacturing, with one effort focused on 2.5-inch steel exhaust tubing. The Define phase began with detailed process maps, which gave the team a shared picture of how the tubing moved through the plant before anyone started changing things.

That starting point is the lesson. Mapping the real process first exposes where material is lost and where handoffs add variation, which then guides what to measure, what to analyze and what to fix. You can walk through more projects in our DMAIC case studies with phase-by-phase results.

A simpler example you can adapt

Imagine a clinic where 22% of patients wait more than 30 minutes past their appointment time. These numbers are illustrative, not from a real project.

The example shows the method clearly: the problem was not staffing speed, it was how time was booked.

Common mistakes in DMAIC projects

• A fuzzy problem statement. If the goal cannot be measured, the project cannot be judged.
• Too big a scope. Teams that try to fix “customer service” stall. Teams that fix “billing-dispute resolution time” finish.
• Trusting bad data. A measurement system check is cheap. Rebuilding conclusions from unreliable numbers is not.
• Jumping to solutions. A favorite fix proposed in week one often gets in the way of the analysis.
• Leaving out the people who run the process. They know where the workarounds are, and they will decide whether the new method sticks.
• Treating Control as paperwork. Without an owner and a reaction plan, the gain fades.

Our perspective: what we see when people learn DMAIC

At the Management and Strategy Institute we have certified more than 300,000 professionals, and the same pattern shows up again and again. People learn the five letters in an afternoon. The hard part is the discipline between the letters: refusing to propose a fix before Analyze is done, and refusing to close a project before Control is real.

We also see DMAIC travel well beyond factories. Learners come to us from healthcare, retail, government and office operations, and the method works in all of them because it asks the same questions of any process: what is the problem, what does the data say, what is the cause, what fixed it, and how do we know it will stay fixed.

Our advice to anyone starting out is to practice on a small, real process you can measure in a couple of weeks. A project you have actually run teaches more than a stack of templates, and it gives you something concrete to talk about when you apply for improvement roles.

— David Lovell

Build the skill with a certification

If you want a structured path from the five steps to running real projects, our Six Sigma certification package bundles the study materials and certification exams in one all-inclusive program. It is self-paced and online, so you can study from home, and a 30-day satisfaction guarantee backs it. You can compare belt levels and related programs on our Six Sigma certification package deals page, or start with the landing page to see everything we offer. For the vocabulary behind the method, our Six Sigma definitions page is a handy reference.

FAQ

What does DMAIC stand for?

DMAIC stands for Define, Measure, Analyze, Improve and Control. It is the five-phase problem-solving method at the heart of Six Sigma and Lean Six Sigma, used to improve existing processes that miss their targets.

Is DMAIC only used in Six Sigma?

No. It is a core Six Sigma tool, but it also works as a standalone improvement framework, as Wikipedia’s overview of DMAIC notes. Many organizations use it alongside Lean and in settings far from manufacturing, such as hospitals and service desks.

What is the most important DMAIC phase?

None of them can be skipped, because each one builds on the last. In practice, many projects go wrong in Define, with a vague scope, or in Control, where the improvement is never locked in. Giving both phases real attention prevents the most common failures.

How long does a DMAIC project take?

It depends on scope and how easy the data is to collect. Many small projects run a few months, and larger ones take longer. Set the timeline in the project charter, and keep the scope tight enough that the team can finish.

What is the difference between DMAIC and DMADV?

DMAIC improves a process that already exists. DMADV (Define, Measure, Analyze, Design, Verify) is used when you are designing a new process or product from scratch, or when an existing one needs a complete redesign rather than incremental fixes.

Sources

• ASQ: DMAIC Process: Define, Measure, Analyze, Improve, Control
• US EPA: Environmental Professional’s Guide to Lean and Six Sigma, Chapter 3
• Tenneco: Using Six Sigma to Reduce Scrap, Rework, and Costs
• Wikipedia: DMAIC

Recommended

• Six Sigma DMAIC – Define Stage
• Six Sigma DMAIC – Measure Stage
• Six Sigma DMAIC – Analyze Stage
• DMAIC Case Studies: Real-World Examples with Phase-by-Phase Results