DMAIC vs DMADV: How to Choose the Right Six Sigma Method
DMAIC improves a process you already have. DMADV designs one you do not have yet. That is the short answer, but real projects are messier: a process can exist and still be impossible to fix, and a “new” project can turn out to be a copy of something that works. This guide compares the two methods side by side, gives you a five-question test, sorts twelve everyday situations, and shows what to do when a DMAIC project hits a wall.
The short answer
ASQ, the professional body for quality, describes DMAIC as a structured problem-solving approach for improving existing processes. Its glossary describes DMADV as a design methodology for new products or processes, used when existing processes cannot be improved enough to meet customer requirements.
- Use DMAIC when the process runs today, you can measure it, and the gap comes from errors, delays or variation inside the current design.
- Use DMADV when there is nothing to fix yet, or when even a perfectly run version of the current design cannot meet the requirement.
The full guides are DMAIC explained step by step and DMADV explained, with a worked example. This page is for the decision between them.
Side-by-side comparison
| DMAIC | DMADV | |
|---|---|---|
| Stands for | Define, Measure, Analyze, Improve, Control | Define, Measure, Analyze, Design, Verify |
| Starting point | A process or product that exists and falls short | A need with no good existing answer, or a process that must be rebuilt |
| Core question | Why does it fail, and how do we fix and hold the fix? | What must it do, and what design will do it? |
| What the data describes | How the current process performs | What customers and the business need |
| How success is judged | Performance compared with the baseline you measured | Test results compared with requirements you set in advance |
| Typical output | A better process plus a control plan | A verified design ready to release |
| Main risk | Polishing a design that cannot reach the target | Designing something nobody needed, or a test that proves nothing |
| Scale of change | Incremental | Large, often a new or rebuilt process |
The last two rows summarize our own reading of the two methods rather than a fixed rule. Projects vary.
Same letters, different work
Define, Measure and Analyze appear in both methods, which hides how different the work is. Only the last two phases have different names, but the first three are not interchangeable.
| Phase | In DMAIC | In DMADV |
|---|---|---|
| Define | Write a problem statement with a number, set a goal and scope | Describe an opportunity and the limits of the design: who it is for, budget, rules |
| Measure | Collect baseline data and check the measurement system | Collect customer needs and turn each into a measurable requirement (CTQ) |
| Analyze | Find the proven root causes of the gap | Compare design concepts against the requirements and pick one |
| Improve / Design | Test fixes to the current process | Build the detailed design, prototype and simulate |
| Control / Verify | Lock in the gain with a control plan and an owner | Test against the requirements, pilot, and decide: release, revise or loop back |
A five-question test
This is our own checklist, not an official standard. Answer in order and stop at the first clear result.
Which method? Five questions
1. Does the process or product run today? If no, choose DMADV. If yes, go on.
2. Can you measure how it performs? If not yet, you still lean toward DMAIC: the Define and Measure phases are where you build that measurement. Go on.
3. Does the gap come from errors, delays or variation inside the current design? Typical signs: some days or people do well and others do not; known steps get skipped; settings are wrong. If yes, choose DMAIC.
4. Have the requirements moved beyond what the current design can deliver? Examples: new volumes, a new rule, a new type of customer, a technology change. If yes, lean toward DMADV.
5. Would the best fix you can think of still miss the target? Do rough arithmetic. If even a good fix leaves you short, choose DMADV, or start DMAIC and use the Analyze phase to confirm the wall (see below).
If you answer yes to both 3 and 4, split the work: use DMAIC for the part that is merely broken and DMADV for the part that is outdated.
Twelve situations sorted
All of these are made up. Several look like one method but belong to the other.
| Situation | Method | Why |
|---|---|---|
| 18% of deliveries arrive late because the order cut-off is after the carrier pickup | DMAIC | The process exists, the cause sits inside it |
| Launch an app for booking appointments where bookings are now only by phone | DMADV | A new service, so there is no baseline to fix |
| Open a second branch that will use the same procedures as the first | Neither at first | Copy the proven process. Use DMAIC later if the new branch lags |
| 4% of invoices contain errors | DMAIC | Measurable defect in a running process |
| The old order system cannot support a new product line or a new reporting rule | DMADV | The requirement is beyond the current design |
| Phone wait times are long because staff schedules do not match call patterns | DMAIC | A fixable mismatch inside the current service |
| Create onboarding for a service line the company has never offered | DMADV | Nothing exists to improve |
| A stable production line makes parts with too much variation | DMAIC | Classic variation reduction |
| Add a genuinely new feature to an existing product | DMADV, scoped to the feature | Design the new part. Leave the rest alone |
| Nobody has measured a process and everyone has a different opinion about the problem | DMAIC | It exists, so define the problem and measure it first |
| Merge two billing processes after an acquisition | Often DMADV | The combined process does not exist yet. Some teams instead run DMAIC on the better of the two |
| Complaints about a service one month after launch | DMAIC, unless the cause is the design | It now runs and can be measured. If the root cause is in the design, loop back |
The pattern: the word “new” does not decide the method. What decides it is whether there is something to measure and whether the existing design can meet the requirement.
When DMAIC hits a wall
Sometimes you cannot tell at the start. This made-up example shows how to find out, and what to do.
An insurance claims team handles 1,000 claims a month. In the Measure phase it finds that 300 (30%) take more than 10 days. The goal is 10% (100 claims).
| Step | Result | Late claims |
|---|---|---|
| Baseline | 30% take more than 10 days | 300 |
| First DMAIC cycle: intake checklist and triage rules | Late rate falls to 18% | 180 |
| Pareto of the remaining 180 | 117 (65%) are waiting for paper forms to arrive by post | 117 of 180 |
| Best incremental fix: scan forms at the branch | Removes about 30 late claims | 150 (15%) |
| Goal | 10% | 100 |
Check the arithmetic: 117 ÷ 180 = 65%, and 180 − 30 = 150, which is 15% of 1,000. Even the best fix the team can think of leaves it five points short, because the root cause is the design of the intake itself: paper, posted to one office.
That is the signal to switch. The team closes the DMAIC project with the gains it has (30% down to 18%), records the finding, and opens a DMADV project to design an online submission service with requirements taken from customers.
A simple rule. After the Analyze phase, ask: if we removed every cause we found, would we hit the goal? If the honest answer is no, and the remaining causes are built into the design, propose DMADV. Do this before you spend money on the Improve phase, not after.
How the two work together
The methods are not rivals. Many organizations use them in sequence on the same process:
- DMADV designs the process or product and verifies it against requirements.
- The process goes live and an owner takes charge of it.
- DMAIC keeps improving it as real data comes in and small problems appear.
- If requirements change enough that the design reaches its limit, DMADV starts again.
You will often see DMADV described as one roadmap within a wider approach called Design for Six Sigma. Our design for Six Sigma article covers that bigger picture. Other design roadmaps have different letters, so if a colleague mentions another acronym, check which phases it uses before assuming it matches DMADV.
Mistakes when choosing
- Using DMADV because it sounds bigger. A broken invoice process does not need redesigning if a checklist fixes it. Choose the smallest method that can meet the goal.
- Using DMAIC to polish a design that is the cause. Teams can spend a year trimming a process whose structure caps the result. Do the Question 5 arithmetic early.
- Treating “new” as “design.” A new branch, new team or new location running a proven process needs training and standards, not a design project.
- Starting with the solution. Both methods fail if the answer is decided before Define. A DMADV project with only one concept is a decision dressed up as a method.
- Skipping the baseline in DMAIC. Without a measured starting point you cannot show improvement.
- Skipping pre-set pass criteria in DMADV. If you choose the success criteria after the test, Verify proves nothing.
- Ignoring the people. Both end in changed daily work. Involve the process owner from Define onward.
Learn more
Start with the two detailed guides above, then see Six Sigma vs. Lean Six Sigma, the belt levels and a short-cycle option in kaizen event steps. For training, our Six Sigma Green Belt course guide compares real prices.
Online providers such as Coursera and Udemy list Six Sigma courses. We may earn a commission if you sign up, at no cost to you. We are not affiliated with any university. Check each syllabus: many courses teach DMAIC in depth and mention design methods only briefly.
Frequently asked questions
What is the main difference between DMAIC and DMADV?
DMAIC improves an existing process. DMADV designs a new process or product, or rebuilds one from scratch. The first three phases share names, and the last two differ: Improve and Control versus Design and Verify.
When should I use DMADV instead of DMAIC?
When nothing exists to improve, or when the best possible fix to the current design would still miss the requirement.
Can a project start with DMAIC and switch to DMADV?
Yes. If the Analyze phase shows the root causes are built into the design and fixing them cannot reach the goal, close the DMAIC project with what it achieved and open a DMADV project.
Is DMADV the same as Design for Six Sigma?
The ASQ glossary says DMADV is also known as Design for Six Sigma. Other guides call it one roadmap within Design for Six Sigma. In everyday use the terms are often interchangeable.
Which one is easier to learn?
In our view, DMAIC is the better place to start. Its tools, such as measuring a baseline and finding root causes, are also useful in design work. Check your course syllabus to see which method it covers.
Do I need statistics for both?
Both use data, charts and tests. DMAIC leans on describing and comparing current performance. DMADV leans on setting requirements, comparing concepts and testing a design. How deep you go depends on the project and the tools it needs.
Browse certificate, course and degree programs side by side.