Six Sigma

What is Six Sigma?

Six Sigma is a data-driven process improvement methodology that statistically measures variation and defects in a process, proves their causes and reduces them structurally, aiming for no more than 3.4 defects per million opportunities. The method was introduced at Motorola in 1986 and became widely known when General Electric built its business strategy around it from 1995 onwards. Six Sigma works in projects that follow the fixed DMAIC cycle, run by trained specialists (the “belts”) under one strict rule: every improvement must be proven with data. It is applied in manufacturing, healthcare, logistics, finance and IT, and on the factory floor it connects naturally to data from the MES, SCADA and the historian.

🕰️ Where does Six Sigma come from?

Six Sigma builds on the statistical quality control of Walter Shewhart and W. Edwards Deming. Motorola engineer Bill Smith linked repairs made during production to later failures in the field, and translated that into a single measurable target per opportunity for a defect.

📊 What does the sigma level mean?

The Greek letter σ (sigma) denotes the standard deviation: the spread of a process around its mean. A process reaches “six sigma” when six standard deviations fit between the process mean and the nearest specification limit. The smaller the spread, the more sigmas fit and the fewer products fall outside tolerance.

The well-known figure of 3.4 defects per million opportunities (DPMO) follows from a convention Motorola adopted: over the long term a process mean drifts by about 1.5σ due to wear, raw material changes and adjustments. A process that achieves 6σ in the short term therefore performs like 4.5σ in the long term, which gives 3.4 DPMO. Without that shift, a perfectly centred 6σ process would produce only about 0.002 defects per million.

You calculate DPMO as: number of defects ÷ (number of units × opportunities per unit) × 1,000,000. An example: a packaging line checks five characteristics per box (fill weight, seal, label, batch number and best-before date). For 20,000 boxes that is 100,000 opportunities; 500 defects found give 5,000 DPMO, roughly 4.1σ. In SPC terms, a 6σ process corresponds to a short-term Cp and Cpk of 2.0; with the 1.5σ shift, the long-term Cpk drops to 1.5.

🔧 How does the DMAIC cycle work?

For existing processes Six Sigma uses DMAIC, a five-phase cycle inspired by Deming’s PDCA cycle. Each phase usually ends with a tollgate review in which the sponsor decides whether the project may continue.

The Analyze phase is in effect a structured Root Cause Analysis. For new products or processes, organisations use DMADV (Define, Measure, Analyze, Design, Verify), also known as Design for Six Sigma (DFSS): there you design quality in from the start instead of repairing it afterwards.

🎯 What does Six Sigma deliver?

• Fewer defects - less scrap, rework and complaints
• Less variation - predictable output, even with changing raw materials or shifts
• Measurable performance - every project starts and ends with a baseline and a KPI
• Lower cost of poor quality - GE reported savings of 350 million dollars in 1998
• Data instead of assumptions - decisions rest on statistical evidence
• Cross-functional collaboration - production, quality, maintenance and IT work on one problem

Outside the factory the goals are just as concrete: fewer medication errors or better use of operating theatre capacity in healthcare, invoices processed without errors or delay in finance, higher uptime and shorter response times in IT, and predictable lead times and error-free deliveries in logistics.

🧠 Which roles does a Six Sigma organisation have?

Besides the DMAIC phases, ISO 13053-1 also describes the roles, expertise and training of the people involved. In practice a karate-style ranking is used:

🔄 What is the difference between Six Sigma, Lean and TPM?

Lean Six Sigma combines the speed of Lean with the measurement discipline of Six Sigma and is widely used in the Netherlands and across Europe, in factories as well as in hospitals and public bodies. Within a quality management system based on ISO 9001, Six Sigma is one way to give concrete shape to the requirement for continual improvement.

🏭 How do you apply Six Sigma in an IT/OT environment?

In process industries and discrete manufacturing, the data Six Sigma needs is often already available in the OT layer. A practical approach:

1. Pick a problem with financial impact - for example overfilling on a filling line or scrap at an extruder.
2. Unlock the process data - collect measurements from sensors and PLCs via SCADA and the historian, and link them to batch and order data from the MES following the ISA-95 model.
3. Test the measurement system - a Gage R&R study and up-to-date calibration prevent you from mistaking instrument noise for process variation. Also watch data integrity: timestamps, units and missing values.
4. Analyse with statistics - regression or machine learning on historical data shows which process parameters drive the variation.
5. Improve in a controlled way - change setpoints and recipes through change management, so that control remains traceable and safe.
6. Sustain automatically - let the MES or SCADA update control charts in real time and raise an alarm on a trend, instead of relying on manual sampling.

This makes Six Sigma part of Industry 4.0: the Control phase shifts from paper control plans to real-time monitoring, and predictive maintenance stops wear from driving variation back up. In IT service management based on ITIL, the same approach is used to structurally reduce recurring incidents and long resolution times.

⚖️ What are the limitations of Six Sigma?

• The 1.5σ shift is a convention - statistician Donald Wheeler calls it arbitrary; the sigma level is a benchmark, not an exact probability.
• Little that is new - quality pioneer Joseph Juran saw Six Sigma mainly as a repackaging of existing quality methods.
• Bureaucracy - belt programmes with heavy certification and reporting can crowd out the actual improvement work.
• Not suited to innovation - reducing variation fits poorly with research and experimentation, where variation is exactly what you need.
• Dependent on good data - without reliable measurement data, the statistics create false certainty.

❓ Frequently asked questions

What does 3.4 defects per million opportunities mean in Six Sigma?

Six Sigma aims for no more than 3.4 defects per million opportunities (DPMO), which equals a yield of 99.99966% per opportunity. The figure assumes that the process mean drifts by 1.5 sigma over the long term. It counts opportunities for a defect, not products: a product with ten characteristics counts as ten opportunities.

What is the difference between DMAIC and DMADV?

DMAIC (Define, Measure, Analyze, Improve, Control) improves an existing process that is not performing well enough. DMADV (Define, Measure, Analyze, Design, Verify), also called Design for Six Sigma, is used when a process or product has to be designed from scratch. Both follow the same data-driven logic.

Is Six Sigma a standard you can be certified against?

Six Sigma is defined in ISO 13053-1 and ISO 13053-2 from 2011 as a methodology with its associated tools and techniques. In the Netherlands, certification bodies such as Kiwa certify organisations and training providers against ISO 13053. Personal Green Belt and Black Belt certificates are issued by training providers and professional bodies, not by ISO itself.

What is the difference between a Green Belt and a Black Belt?

A Green Belt leads smaller Six Sigma projects alongside their own job and masters the basic statistics. A Black Belt is largely dedicated to improvement work, leads complex cross-departmental projects and coaches Green Belts. A Master Black Belt in turn trains and supports the Black Belts.

What is Lean Six Sigma?

Lean Six Sigma combines the waste-reduction and flow focus of Lean with the statistical variation reduction of Six Sigma. The term became popular after Michael George’s 2002 book on the subject. Many training providers now offer Lean Six Sigma rather than Six Sigma on its own.

How does OT data help a Six Sigma project?

OT data from sensors, PLCs, SCADA and the historian provides continuous, timestamped measurements instead of manual samples. This lets a Six Sigma team establish a reliable baseline faster and prove causes statistically. In the Control phase, the MES or SCADA then monitors the improved process in real time with control charts.

📌 In summary

Six Sigma is a statistically grounded improvement methodology that uses the DMAIC cycle to reduce process variation and defects, aiming for 3.4 defects per million opportunities. The method originated at Motorola in 1986, has been defined in ISO 13053 since 2011 and is often combined with Lean. In an IT/OT environment, data from SCADA, the historian and the MES provides the measurement baseline, and real-time monitoring with control charts sustains the gains.