Lean vs Six Sigma: differences, history and when to use each

Lean and Six Sigma are both improvement methods, and many companies run them together. They start from different questions, though, and use different tools.

Lean asks where time and effort are wasted. Six Sigma asks why results vary. This guide compares the two and explains when to use each.

Key takeaways

Lean targets waste and flow. Six Sigma targets variation and defects.

Lean comes from the Toyota Production System. Six Sigma was developed at Motorola in the 1980s.

Lean Six Sigma combines both, often using the DMAIC cycle with lean tools inside it.

Use lean first when work waits and piles up. Use Six Sigma first when output is unpredictable or defects are hard to explain.

What is lean?

Lean is a way of improving work by removing waste, so value flows to the customer without delay. It grew from the Toyota Production System.

Its tools are mostly visual and practical: value stream mapping, 5S, kanban, SMED and standard work. The lean methodology guide covers the principles.

What is Six Sigma?

Six Sigma is a data-driven method for reducing variation and defects. Motorola developed it in 1986. Engineer Bill Smith is usually credited with the approach, and chief executive Bob Galvin backed it across the company.

The name comes from statistics. A process running at six sigma, allowing for a 1.5 sigma shift, produces about 3.4 defects per million opportunities. The sigma level calculator converts your own defect counts.

Six Sigma projects follow DMAIC: define, measure, analyze, improve and control. Practitioners are trained to levels named after martial-arts belts, such as Green Belt and Black Belt.

The Six Sigma guide covers the method, the belts and the statistics in more depth.

A short history of lean and Six Sigma

Lean grew out of the Toyota Production System, built by Taiichi Ohno and others from the late 1940s onwards. The Toyota Production System page covers its two pillars, just-in-time and jidoka.

The word lean came later. John Krafcik used it in a 1988 article. Womack, Jones and Roos made it widely known in their 1990 book The Machine That Changed the World.

Six Sigma began at Motorola in 1986. General Electric adopted it in the mid-1990s under Jack Welch, which spread it across large companies.

The two streams met in the late 1990s and 2000s. Consultants and companies combined them as Lean Six Sigma, often with DMAIC as the common frame.

Lean vs Six Sigma in manufacturing

In manufacturing, the difference between lean and Six Sigma shows up in the problems each one picks. Lean looks at the whole flow of material, from the supplier to the customer.

A lean team might find that a part waits days between two processes that each take minutes. The fix is flow: smaller batches, kanban pull and quicker changeovers with SMED.

A Six Sigma team looks inside one process. It might find that a machining step drifts out of tolerance on some shifts. The fix comes from data: measurement checks, a designed experiment and statistical process control.

Both meet on the shop floor. A stable, flowing line makes variation easier to see, and a capable process makes flow easier to keep.

Small plants can get outside help with both. Clients of the Manufacturing Extension Partnership, a NIST program that supports small and medium-sized manufacturers, reported $15 billion in new and retained sales and $2.6 billion in cost savings in fiscal year 2024.

One problem, worked both ways

Here is an illustrative example with made-up numbers. A distributor's orders take 9 days from receipt to delivery, and 4% of them arrive wrong. Two teams look at the same process.

The lean team walks one order through the building and times it. The actual work adds up to about 3 hours. The rest is waiting: 2 days for a credit check, 3 days until the weekly planning run, and 2 days in a picking queue that is worked in large batches. It moves the credit check to the quote stage, plans daily and picks in small waves. Lead time falls to about 4 days. The error rate doesn't change, because the lean team never looked at why orders go wrong.

The Six Sigma team counts errors instead. Out of 1,000 orders, 40 are wrong. Each order has 5 places where it can go wrong, so that is 40 / 5,000 × 1,000,000 = 8,000 DPMO, about 3.9 sigma. A Pareto chart shows that 60% of errors are wrong delivery addresses. The data splits by channel: phone orders have a 6.5% error rate and web orders 1.2%. Address validation on the phone order screen brings the overall rate down to about 1.5%. Lead time doesn't change.

Each team fixed the problem its method is built to see. Run together as a Lean Six Sigma project, the distributor gets both results: about 4 days and about 1.5% wrong.

Lean vs Six Sigma compared

What is Lean Six Sigma?

Lean Six Sigma combines the two. Lean makes the process flow, and Six Sigma makes the flowing process consistent.

In practice, many teams use DMAIC as the project frame and pick lean or statistical tools inside each phase. A value stream map often appears in Measure, and SMED or poka-yoke in Improve.

Training bodies certify Lean Six Sigma belts. The Lean Six Sigma certification page compares providers, and the Lean Six Sigma tools hub groups the software.

Limits and criticisms of each

Both methods have well-known failure modes, and knowing them helps you avoid them.

Lean used for headcount cuts. If improvements lead to layoffs, people stop suggesting them, and the program stalls. Many lean leaders promise up front that no one will lose a job because of an improvement.

Lean as a tools program. Teams can copy 5S and kanban boards without the daily management that keeps them alive, and the gains fade within months.

Six Sigma bureaucracy. Large programs can turn into tollgate paperwork and savings reports, with more effort on documenting projects than on running them.

Belt factories. When training counts matter more than results, companies certify many belts who never run a real project. The Six Sigma belts guide covers how to judge a certificate.

Statistics for the sake of it. Not every problem needs a hypothesis test. A team that runs a designed experiment on a problem a five whys would solve has wasted its time.

Which should you use first?

Start with lean when work waits, inventory piles up or lead times are long. These problems are usually visible without statistics.

Start with Six Sigma when output varies and nobody can explain why. Examples are scrap that comes and goes, or a test that sometimes fails.

Many teams do both in turn. They stabilize and simplify the process with lean, then use Six Sigma on the defects that remain.

Whichever you start with, the management routine around it matters. A Census Bureau survey of about 35,000 US manufacturing plants, analyzed by Nicholas Bloom and colleagues in the American Economic Review in 2019, found that structured management practices, such as monitoring performance and setting targets, account for more than 20% of the variation in productivity between plants.

A quick checklist helps. Are items waiting more than they are being worked on? Start with lean. Do you have reliable data, and does the result vary in ways nobody can explain? Start with Six Sigma. Is the problem both slow and error-prone? Map the flow first, then run a DMAIC project on the worst defect.

For a wider comparison that includes kaizen, theory of constraints and other methods, see process improvement methodologies.

Match your situation to a starting point

Which tools each side reaches for

Which fits your organization's maturity?

How each one changes the culture

Software for each

Lean work is often paper-first, then moves to kaizen software, kanban software or value stream mapping software.

Six Sigma work needs statistics. Six Sigma software covers analysis tools, and SPC software monitors processes on the line.

Frequently asked questions