Lean Six Sigma: what it is, when you use it and how

Lean Six Sigma combines two improvement methods: Lean makes processes faster by removing waste, Six Sigma makes them more predictable by reducing variation and errors. You use the combination when a process repeats often and is both slow and unreliable. Lean alone if you can see where it gets stuck; Six Sigma alone if nobody knows why the outcome keeps shifting.

This page covers what both methods involve, how they work together, which route to choose for which problem, what a project looks like week by week, what it delivers, and when you should stay away from it altogether.

What is Lean?

Lean is a way of working that looks at the flow of work through a process and strips out everything the customer isn't waiting for and won't pay for. It's rooted in the Toyota Production System; Toyota describes it as "just-in-time" plus stopping immediately at a deviation. Womack and Jones summed it up in Lean Thinking with five principles: define value from the customer's perspective, map the value stream, let the work flow, work on pull instead of push, and keep improving.

Lean counts in time: lead time, waiting time, number of handoffs. Its tools are visual and simple — value stream mapping, 5S, kanban, kaizen. See also the full page on Lean.

What is Six Sigma?

Six Sigma is a data-driven method for reducing variation in a process so the outcome stays within the customer's requirements. The name refers to the statistical ambition: six standard deviations between the process average and the nearest specification limit, which over the long term works out to 3.4 defects per million opportunities. Six Sigma originated at Motorola and became widely known through General Electric; the method is described by ASQ and standardised in ISO 13053.

Six Sigma counts in spread: standard deviation, DPMO, process capability. Its tools are statistical — measurement system analysis, control charts, hypothesis tests, regression. See also the full page on Six Sigma.

Lean versus Six Sigma in one table

Why the combination works

On their own, both methods have a blind spot. Lean makes a process faster, but a faster process with a lot of variation mostly produces faster mistakes. Six Sigma makes a process more reliable, but a stable process full of unnecessary steps stays costly. Together they cover the two questions present in almost every process: are we doing unnecessary things, and are we doing the necessary things consistently.

The practical order is almost always the same: Lean first, then Six Sigma. Remove the unnecessary steps, waiting times and duplicate checks. Whatever still fluctuates after that is the real variation problem — and it's usually much smaller and easier to measure than it looked at the start.

Which route do you choose?

Two questions are enough to decide. Is this about waste or about variation? And do you already know the cause?

• Cause visible, problem is slowness → Lean. Start with a value stream on one page and cut.
• Cause unknown, problem is scatter → Six Sigma via DMAIC. Measure first, act after.
• Both → Lean Six Sigma: a Lean round as the first two weeks of the DMAIC project.
• No repeating process → neither. Use ordinary project planning instead, for example with a Gantt chart.

How a Lean Six Sigma project runs

Six Sigma works to a fixed rhythm: DMAIC — Define, Measure, Analyze, Improve, Control. Every phase has a gate question you need to be able to answer with evidence before moving on.

A realistic schedule for one well-scoped process:

1. Week 1 — Define. Problem in numbers, customer requirement (CTQ), scope and sponsor. One process, one outcome measure.
2. Weeks 2–3 — Lean round and Measure. Map the value stream, time net processing time versus lead time, remove obvious waste, record the baseline.
3. Weeks 4–7 — Analyze. Search for causes and test them against data. This is usually where half the assumptions fall away.
4. Weeks 8–12 — Improve. Test solutions small, measure, then roll out.
5. Week 13+ — Control. Document the standard, set up monitoring in the line, hand over. Without this phase, an improvement typically slips back within a few months.

Pick the outcome measure before you start and stick with it; how to do that is covered in Choosing KPIs.

What "six sigma" means in numbers

The key lesson: don't automatically aim for 6σ. The step from 3σ to 4σ is cheap and delivers a lot; the step to 6σ isn't worthwhile for most commercial processes. Choose the level that matches the cost of a single error. A wrongly sent newsletter is not the same as a wrongly dosed medicine.

The eight wastes: where the time really goes

Transport, inventory, motion, waiting, overproduction, overprocessing, defects and unused talent (TIMWOODS). In office and service processes, waiting is almost always the biggest of the eight.

Belts and roles

Six Sigma works with training levels, typically in three tiers:

• Yellow belt — knows the concepts, takes part in an improvement team. A few days of training.
• Green belt — leads improvement projects within their own work. Usually 5 to 10 training days plus a real project.
• Black belt — leads more complex projects, masters the statistics, coaches green belts. Often full-time or largely dedicated.

Important: certification isn't standardised. No single body issues "the" title, and quality varies widely between providers. Always ask about the actual project behind a certificate — it says more than the paper does.

Where it's used outside the factory

The method originated in manufacturing but is now used widely: hospitals (outpatient lead times, medication errors), banks and insurers (acceptance, claims), government (applications and decisions), logistics, and also marketing and sales. Examples from our own field where this thinking applies directly:

• Lead follow-up: the time between a request and the first response is almost always mostly waiting time. That's a Lean problem. See online leads.
• Content production: too many review rounds and rework. That's overprocessing and defects.
• Website conversion: results that fluctuate week to week, where without measurement you can't tell if anything worked. That's variation — exactly why you test A/B results for significance and work more broadly on conversion.

When to stay away from it

Being honest about the limits, because this is where most projects fail:

• The process barely repeats. Without volume there's no data and no learning curve.
• There's no usable measurement, and there won't be one. Six Sigma without reliable numbers is an opinion with charts.
• The cause is already known and the solution is ready. Just do it. A DMAIC project to prove what everyone already knows costs months for nothing.
• The real problem is strategy, positioning or demand. Improving a process only makes the wrong offer more efficiently wrong. Start with a SWOT analysis and the strategy instead.
• There's no owner in the line. Without someone to maintain the new standard, the gains disappear during Control.

Common mistakes

• Measuring as an end in itself. Months of data collection without a single improvement. Put a deadline on Measure.
• Confusing tooling with method. A dashboard isn't an improvement; a control chart only matters once someone acts on it.
• Not defining the customer requirement. Without a CTQ, every "defect" is a matter of opinion.
• Only tackling what's easy to measure. What's easy to measure isn't always what hurts.
• Skipping Control. The most common mistake, and the costliest: the process slips back and the next improvement proposal gets no credit.

How to approach it in practice

A starting approach you can run yourself, without training:

1. Pick one process that happens often and that people are frustrated by.
2. Write the steps down on a single page — exactly how it works now, not how it's meant to work.
3. Time the net processing time and the total lead time for one job. The gap is your starting win.
4. Mark each step as value, control, or waiting.
5. Cut or combine whatever nobody can explain. Do this with the people doing the work.
6. Track the same number for four weeks, weekly, and make it visible.
7. Whatever still fluctuates after that: that's where you start analysing.

Sources

• ASQ — Six Sigma
• ASQ — Lean
• ASQ — DMAIC
• ISO 13053-1: Quantitative methods in process improvement — Six Sigma
• Toyota — Toyota Production System
• Lean Enterprise Institute — Value stream mapping

We're not Lean Six Sigma consultants; we use this thinking in digital, marketing and AI processes — lead follow-up times, content production, and the measurability of campaigns. Want to talk it through? Book a consultation.