Lean Six Sigma Guide: DMAIC, Tools and How to Improve a Process
September 30, 2026 · Professional Development · 8 min read
Lean Six Sigma works best when the problem is real, measurable and worth solving
The method is often reduced to belts, jargon and a five-letter acronym. The useful version is much simpler: define a business problem precisely, measure how the process behaves now, identify the causes that actually drive poor performance, test improvements, and put controls in place so the gains do not disappear after the project team moves on.
ASQ describes DMAIC as a structured problem-solving approach for improving existing processes that fail to meet performance standards or customer expectations. NIST’s Manufacturing Extension Partnership similarly frames lean and continuous improvement around disciplined methods that improve flow, reduce defects, shorten lead time and strengthen productivity.
Source: ASQ, DMAIC Process
Source: NIST MEP, Lean and Process Improvement
Lean, Six Sigma and Lean Six Sigma: what each contributes
| Approach | Primary emphasis | Typical question | Useful when |
|---|---|---|---|
| Lean | Flow, waste, customer value and efficient work | Which steps add no value or create delay? | The process has waiting, handoffs, queues, excess movement, rework or unnecessary complexity. |
| Six Sigma | Variation, defects, measurement and process capability | Why does the output vary, and which inputs drive that variation? | Performance is inconsistent, defects recur or the process needs stronger data analysis. |
| Lean Six Sigma | Combines flow improvement with disciplined analysis of variation and causes | How do we make the process faster, more reliable and easier to control? | The problem involves both inefficiency and inconsistent outcomes. |
The DMAIC roadmap, without the jargon
Define: agree on the problem before solving it
Write a problem statement that describes the current condition, the affected customer or process, the business impact and the scope. Avoid solutions disguised as problems. “We need automation” is a proposed solution. “Invoice approval takes a median of nine days and creates late-payment penalties” is a problem worth investigating.
Measure: establish how the process actually performs
Map the real process, define the metric, check whether the measurement is trustworthy and establish a baseline. ASQ explicitly places measurement-system validation and baseline performance inside this phase because unreliable data produces unreliable conclusions.
Analyze: find causes, not convenient explanations
Use process data, Pareto analysis, cause-and-effect thinking, root cause analysis, FMEA or statistical tools as appropriate. The aim is to identify the inputs and conditions that materially drive the problem.
Improve: test changes against the causes
Generate solutions, prioritise them, pilot where possible and measure whether the process responds as expected. A change that feels sensible but does not move the target measure is not yet an improvement.
Control: make the better process the normal process
Create a control plan, standard work, monitoring, ownership and a clear reaction plan for drift. ASQ lists control plans, statistical process control, 5S and mistake-proofing among the tools that can support this phase.
A worked DMAIC example: late customer quotations
Problem: commercial quotations are taking too long
Define: A B2B services team finds that many quotations miss its internal turnaround target. Customers are following up before proposals are ready, and sales staff are escalating manually.
Measure: The team maps the process from enquiry to approved quote and timestamps each handoff. It separates simple quotations from complex ones instead of mixing them into one average.
Analyze: The longest delays are not in price calculation. They occur when proposals require clarification from operations and when approvers receive incomplete information. A Pareto view shows a small number of missing inputs account for a large share of rework loops.
Improve: The team introduces a standard intake form, clearer thresholds for when operations review is required, and a pre-approved pricing band for routine work. It pilots the changes with one sales team.
Control: The process owner tracks turnaround, rework rate and exceptions weekly. If performance drifts, the team reviews the reason rather than simply reminding people to “work faster.”
Tools to use in each phase
| Phase | Useful tools | What they help you see |
|---|---|---|
| Define | Project charter, VOC, SIPOC, stakeholder map | Problem boundary, customer needs, scope and ownership |
| Measure | Process map, value stream map, data collection plan, capability analysis | Current flow, baseline, bottlenecks and measurement quality |
| Analyze | Pareto chart, 5 Whys, fishbone, FMEA, scatter plots, hypothesis testing where appropriate | Patterns, likely causes, failure modes and relationships |
| Improve | Kaizen, prioritisation matrix, pilots, mistake-proofing, design of experiments where justified | Which changes actually address validated causes |
| Control | Control plan, standard work, visual management, control charts, reaction plan | Whether gains hold and when intervention is needed |
When DMAIC is the wrong tool
A structured method is valuable, but not every improvement problem needs the full machinery.
If a broken form field causes every request to fail, fix the field. Do not launch a three-month analysis project.
ASQ distinguishes DMAIC for improving an existing process from DMADV approaches used when designing a new product, service or process.
Start by establishing a usable operational definition and data collection method before drawing statistical conclusions.
DMAIC can improve a process, but it cannot decide whether the organisation should exit a market, change its business model or redefine customer value.
How Lean changes what you notice
Lean asks teams to look at work from the customer’s point of view and identify effort that consumes time or capacity without creating the intended value. NIST MEP highlights tools such as value stream mapping, 5S, quick changeover and cellular flow as practical methods for making work more visible and reducing friction.
Process walk checklist
- Where does work wait before the next step?
- Where is information entered more than once?
- Which approvals exist because of an old problem that may no longer exist?
- Where does work move backward for correction?
- Which handoffs cause lost context or delay?
- Where do employees create personal workarounds because the standard process is difficult?
- Which steps matter to the customer, regulator or business outcome?
- Which steps consume effort without changing the outcome?
The measurement traps that derail projects
How to choose a first Lean Six Sigma project
A strong first project is important enough to matter, narrow enough to finish and measurable enough to learn from.
| Good candidate | Warning sign |
|---|---|
| Recurring defect, delay, cost or customer issue | One-off incident with no repeatable process |
| Clear process owner and cross-functional participation | No one has authority to implement changes |
| Data can be collected within the project window | Success depends on a metric nobody can define |
| Scope can be contained to a value stream or process segment | Project charter is effectively “fix the whole company” |
| Potential business impact is meaningful | Team picked the project only because data are convenient |
Build a project charter before collecting data
A Lean Six Sigma project becomes much easier to control when the team agrees on the boundary before analysis begins. A useful charter is short enough to read in a meeting and specific enough to prevent the project from absorbing every adjacent problem.
Lean Six Sigma project-charter template
- Problem: what is happening now, where, how often and with what consequence?
- Goal: what measurable condition should improve, by how much and by when?
- Scope: where does the process start and stop, and what is explicitly out of scope?
- Customer: who experiences the effect and what requirement matters to them?
- Business impact: cost, delay, quality, risk, capacity or customer consequence.
- Owner: who has authority to change the process?
- Team: which functions and process experts need to participate?
- Primary measure: the metric that will determine whether the project succeeded.
A control plan should tell people what to do when performance drifts
| Control element | Example |
|---|---|
| Metric | Quotation turnaround time |
| Target | 90% completed within two business days |
| Frequency | Reviewed weekly |
| Owner | Commercial operations manager |
| Trigger | Performance below target for two consecutive weeks |
| Reaction | Review exception log, identify dominant cause, assign corrective action and verify recovery |
The control phase is where an improvement becomes an operating system rather than a project result. If nobody owns the measure, there is no agreed trigger for action, or the team stops looking at the process after the project closes, the old behaviour can return without anyone noticing.
Learn the method by applying it to real work
MATSH’s Lean Six Sigma and Process Improvement Course develops the full DMAIC logic, process analysis, improvement tools and control thinking needed to move from improvement ideas to evidence-based operational change.
Sources
Need In-House Training?
We run all our courses as private programmes for organisations across the GCC and Africa.
Request In-House →