The History of Six Sigma

By Victor Hoang · Updated Oct 24, 2026

Here is the problem every era in this story tries to solve. A process can look fine on average and still ship defects. The spread around the average is what customers feel.

Six Sigma took shape at Motorola in the 1980s, which the company dates to 1986. It became a global management method after General Electric adopted it in 1995. In its 1988 annual report Motorola described the target as 3.4 parts per million, which it called virtual perfection. The throughline is a long climb from a statistical idea to a management system. First came a way to measure variation, then a chart to watch it. Then a company-wide target, then a belt-and-project machine run from the CEO's office.

This article follows that idea from an eighteenth-century curve to today's standards. For the method itself, start with what Six Sigma is, the DMAIC cycle and the belt system. This page tells where they came from. The wider story of how managers learned to run organizations is in the history of management.

Timeline at a Glance

The Variation Years (1733-1924)

Mathematicians measure spread; Shewhart draws a chart

Six Sigma starts as arithmetic. In November 1733, Abraham de Moivre circulated a Latin pamphlet that first published the normal approximation to the binomial distribution, the bell-shaped curve that later became the picture of "normal" variation. He was working on games of chance, not factories. His Doctrine of Chances had first appeared in 1718, and the approximation reached the 1756 edition.

The word "sigma" needed a measure to stand for. Karl Pearson supplied the name in a lecture on 31 January 1893, when he introduced "standard deviation" as a convenient substitute for the clumsier "root mean square error" and older terms like "mean error." That is the unit in Six Sigma's name: a standard deviation is one sigma.

The bridge from mathematics to the shop floor came at Western Electric. Walter Shewhart, born in Illinois in 1891, joined the company's Inspection Engineering Department at Hawthorne in 1918. In May 1924 he wrote a short memo, about a page long, with a schematic diagram of what became the control chart. MacTutor's biography says he created the first statistical control charts of manufacturing processes that year. A control chart plots a measurement over time against limits, so a manager can tell ordinary wobble from a real change.

Keep precursor and concept apart. De Moivre, Pearson and Shewhart built the tools of variation, and none of them proposed Six Sigma. What they supplied was the idea that quality is about spread, and that spread can be watched in real time.

What this means for teams today: If you only track averages, you are working with a pre-1924 picture. Chart the spread too, and look for the points that fall outside it.

The Shewhart and Deming Years (1925-1960s)

Control charts cross the Pacific

The second era is about spreading the tool. The Bell Telephone Laboratories opened in 1925, and Shewhart moved there, staying until he retired in 1956. In 1931 he published Economic Control of Quality of Manufactured Product, the book-length statement of his method.

The idea then travelled with W. Edwards Deming. The Library of Congress notes that after World War II Deming went to Japan as a statistician helping with a census, and that in 1950 he was invited by a Japanese scientists' and engineers' group to give a talk, the first of many. The following year the Union of Japanese Scientists and Engineers set up a prize in his name. Its history page records that the first Deming Prize ceremony was held on September 22, 1951, with Deming as honorary chairman of the committee. Our profile of Deming's leadership covers the man in more depth.

This era made statistical control a management topic rather than an inspector's trick. It also left a gap that Six Sigma later tried to fill. Control charts told you when a process had drifted, but they did not tell a company how to set a target, assign people to fix it, and measure the savings. The same era's rival tradition, Taylor's scientific management, had measured work but not variation.

What this means for teams today: A tool does not spread by being correct. It spreads because someone teaches it to senior people, as Deming did in Japan in 1950.

The Motorola Push (1979-1987)

A radio maker sets targets that look impossible

The pivot comes in a company that made radios, pagers and chips. Motorola's 1988 annual report says the quality process it was recognized for was one the company initiated with renewed energy in 1979. Sales were growing, but key distribution executives told it that customers expected higher quality. By 1981 the company had set a goal that "appeared to be impossible": a tenfold improvement in quality in five years.

The 1986 annual report describes the same program as a formal effort to increase quality tenfold over five years, with many manufacturing operations meeting and exceeding the goal early. The same report says the next focus was zero defects. The 1988 report adds that in 1986 a more formal program of customer visits began, and that Motorolans met the tenfold goal only to find it had not been ambitious enough.

Motorola's own timeline says it invented the Six Sigma quality improvement process in 1986, giving the company a common worldwide language for measuring quality. In January 1987 the company wrote that language into a goal, in the 1988 report's words: 10 times better by 1989, 100-fold by 1991, and Six Sigma capability by 1992. The 1987 report put it as a 99.9997% perfect product or service.

The people behind the method are described differently depending on the source. John Ramberg, writing in Quality Digest in 2000, credits Motorola reliability engineer Bill Smith with originating Six Sigma. Ramberg says Smith found that field failure rates exceeded what final testing predicted, that Motorola's earlier design margin was 4 sigma, and that his analysis led to a 1.5 sigma allowance for drift in the process mean. A letter from the former president of the Six Sigma Academy adds that Mikel Harry, a senior staff engineer in Motorola's Government Electronics Group, began experimenting with statistical problem solving in the late 1970s. The author of that letter is an interested party, so treat the credit allocation as one account. Bob Galvin, the chairman who set the targets and accepted the later award, is the one name every source agrees on.

The 1.5 sigma allowance is why Six Sigma means 3.4 defects per million and not about two per billion. ASQ's Six Sigma Forum definition, quoted in a 2006 Quality Digest column, says allowing for the shift gives 3.4 defects per million opportunities, while ignoring it gives 2 defects per billion. That distinction would later become the most argued detail in the method.

What this means for teams today: Motorola's lesson is the target. A goal that looks impossible forces a new method, and the first goal here, tenfold in five years, was met well enough to need a harder one.

The Baldrige Moment (1988-1994)

A national award turns a company program into a movement

On November 14, 1988, President Reagan presented the first Malcolm Baldrige National Quality Award to Motorola, with Galvin accepting for the corporation. The 1988 annual report records Reagan's line: "Customer satisfaction through better quality is the goal." Motorola noted that it was the only major manufacturer to win company-wide. NIST's history of the program lists Galvin alongside the other 1988 winners, and quotes David Garvin's 1991 Harvard Business Review judgment that the award had become "the most important catalyst for transforming American business."

What Motorola did with the attention is documented in its own reports. The 1988 report says more than $45 million a year went into training, with 40 percent of it devoted to quality. The 1989 report says the company created a university structure for employee training, committed to one week of training per year for every employee, and kept moving toward Six Sigma quality by 1992.

The institution-building that followed, including a Six Sigma research institute headed by Harry, is described in the Six Sigma Academy letter cited above.

This era solved the credibility problem, because a well-known award validated the work. It created a new one: Six Sigma was still a Motorola method, tied to electronics and to one chairman. It needed a CEO who could carry it to a different kind of business. The same quality-first movement produced its sibling in total quality management.

What this means for teams today: External recognition buys attention, not results. Motorola paired its award with training budgets and a fixed target, and that pairing did the work.

The Welch Years (1995-2001)

GE turns Six Sigma into belts, projects and a boardroom number

The method's second life began in a conglomerate. GE's 1998 annual report letter says that after observing the transformational effects of Six Sigma at the few companies pursuing it, GE "plunged into Six Sigma with a Company-consuming vengeance" just over three years earlier, and that Six Sigma "came along in 1995." The letter makes the case in numbers. GE had invested more than a billion dollars. It reported more than three quarters of a billion dollars in savings beyond that investment in 1998, with a billion and a half in sight for 1999. Operating margin, which stood in the 13.6% range in 1995, reached 16.7% at the end of 1998, and working capital turns rose from 5.8 to 9.2. These are GE's own figures about itself, from a letter written to shareholders.

The letter also records the machinery. About 5,000 full-time Master Black Belts and Black Belts drove scores of thousands of quality projects, and virtually every professional at the company was a Green Belt who had completed a project. It also announced a next step: every new GE product and service would be Designed for Six Sigma. That is the origin of the belt ladder described in our guide to Six Sigma belts, and of the project-by-project rhythm of DMAIC.

In 2000 Mikel Harry and Richard Schroeder published Six Sigma: The Breakthrough Management Strategy, which was marketed as the program heralded by GE, Motorola and AlliedSignal. The same year Ramberg asked in Quality Digest whether it was "Fad or Fundamental?" He reported the critics' view that Six Sigma was a marketing ploy with little that was new, and answered that it reuses statistical tools from earlier quality movements but adds leadership, infrastructure and a project cycle.

The era overlapped with a rival fashion. Business process reengineering promised radical redesign in the early 1990s, while Six Sigma promised disciplined, measured improvement. Which one a firm picked said a lot about its appetite for risk.

What this means for teams today: GE's contribution was organization, not mathematics. Put project leaders in named roles, give them a number to move, and report it to the top.

The Lean Merger (2002-present)

Variation and waste get fixed in one program

By the early 2000s companies were running Six Sigma and lean production side by side, and the two had overlapping tools and rival cultures. Michael George's Lean Six Sigma, published by McGraw-Hill in 2002, proposed combining them in a single program to cut cycle times, lower cost and reduce process variation. The reasoning was complementary: Six Sigma attacks variation, lean attacks waste and delay, and many problems have both. Our article on Lean Six Sigma explains how the combined version works.

The method also got a formal standard. ISO 13053-1, published in September 2011, describes the DMAIC methodology in five phases (define, measure, analyse, improve, control), recommends how projects should be managed, and covers roles and training. It states that it applies to service and transactional processes as well as manufacturing. By the time the standard arrived, Six Sigma had moved well beyond the factory floor that produced it.

The open arguments are the old ones. The 1.5 sigma shift remains debated, and Ramberg's question about whether the method is fundamental or fashion has never been fully settled. What is settled is the vocabulary. Terms like DMAIC, defects per million and Black Belt came out of this history and still appear in quality, operations and service teams. It also sits beside later ways of tracking performance, such as the balanced scorecard and modern performance management.

What this means for teams today: Use the variation tools and the waste tools together. If your problem is inconsistency, measure the spread. If it is delay, map the flow.

Key People

Key Facts

Key Facts: Six Sigma History

• Motorola's own timeline says the company invented the Six Sigma quality improvement process in 1986.
• Motorola's 1988 report defines Six Sigma as 3.4 parts per million and says it was the first Baldrige winner among major manufacturers company-wide.
• Allowing for a 1.5 sigma shift gives 3.4 defects per million; ignoring it gives 2 per billion.
• GE reported savings of more than three quarters of a billion dollars beyond its investment in 1998.
• GE had about 5,000 full-time Master Black Belts and Black Belts by the end of 1998.
• ISO 13053-1 on DMAIC was published in September 2011.

The Pattern Behind the History

We call this the Measure, Target, Organize pattern. It is our reading of the timeline, not a research result.

1. Measure. De Moivre, Pearson and Shewhart gave managers a way to see variation before anyone tried to manage it.
2. Teach. Deming carried the tools to senior people in Japan, and Motorola later trained its whole workforce.
3. Target. Motorola turned the measure into a dated goal: tenfold in five years, then Six Sigma by 1992.
4. Organize. GE added the belts, the projects and the savings report that let a conglomerate run the method.
5. Combine. Later practitioners merged it with lean when the two traditions kept meeting in the same plants.

What It Teaches Teams Today

• Measure the spread, not just the average. That is the oldest idea in this story.
• Set a target that forces a new method. Motorola's first goal was ambitious enough to need a harder successor.
• Teach the method at the top. Deming's 1950 lectures and Welch's 1998 letter both worked through senior leaders.
• Name the roles and report the money. GE's belts and savings table are what made the program run at scale.
• Read the criticism. The 1.5 sigma shift and the fad question are open arguments you should understand before quoting a sigma level.

For the wider story, read the history of management and the management theory timeline.

Six Sigma began as a way to see variation and ended as a way to run a company. The teams that still get value from it tend to be the ones that keep the original question in view: how much do our results spread, and what are we doing about it?