Lean Six Sigma analytics

Lean Six Sigma analytics combines the search for waste with statistical analysis of variation and defects. It supports improvement in manufacturing and services when a process, customer requirement and decision can be defined clearly.

When to use it

Use the approach when quality, cycle time, cost or consistency matters and repeated process data are available. Maintain routine monitoring, but launch projects only where the expected value justifies the analytical effort.

It can answer:

• Which outcomes matter to customers?

• Is performance stable, improving or deteriorating?

• Which factors plausibly cause the change?

• Where does delay, rework or excess effort occur?

• Which improvement supports a strategic objective?

Origins

Lean draws from the Toyota Production System and later codification of value, flow, pull and waste reduction. Six Sigma was developed at Motorola, associated particularly with engineer Bill Smith, and later spread through organisations including General Electric. Lean Six Sigma joined complementary traditions: one emphasises flow and customer value; the other variation, capability and disciplined problem solving.

What it is

Lean commonly examines seven forms of waste:

1. Transportation that adds risk or time without customer value.

1. Unnecessary motion by people or equipment.

1. Excess inventory or work in progress.

1. Waiting for work, information, equipment or approval.

1. Overproduction ahead of actual demand.

1. Processing beyond what the outcome requires.

1. Defects, correction and avoidable failure.

Six Sigma uses statistical ideas to describe process capability and reduce variation. Under a conventional long-term assumption, “Six Sigma” performance is often stated as 3.4 defects per million opportunities. Repeating 3.4 as a slogan does not make the denominator, defect definition or shift assumption appropriate; each organisation must define a meaningful opportunity and acceptable risk. The target of 3.4 may be unnecessary for one process and dangerously weak for another.

Why it matters

Waste consumes time and resources; variation makes outcomes unreliable. Both affect customers, employees and economics. Improvement should consider safety, accessibility and resilience as well as speed and cost.

How to use it

Select a strategically relevant problem and define the customer requirement. Use DMAIC:

• Define the problem, boundary, customer, outcome and business case.

• Measure the current process with an operational definition and trustworthy sampling plan.

• Analyse flow, variation and plausible causes; distinguish correlation from causation.

• Improve through designed tests or pilots that address verified causes.

• Control with standard work, ownership, monitoring and a response plan.

A football analogy sometimes describes a goalkeeper facing 50 shots in each of 50 games and conceding at a Six Sigma rate only once in 147 years. The image is memorable but statistically crude: opportunities are not independent or identical. Use process-specific data instead.

Practical example

A manufacturer receives an order for 1,000 jeans: 400 size-30, 400 size-32 and 200 size-34. The customer accepts size-32 waists from 31.8 to 32.2 inches, a 0.2-inch total tolerance around the nominal size. A premium customer may require 100 per cent of accepted units to remain within a tighter 0.1-inch condition.

Define the critical requirement, validate the measurement system and study the distribution by machine, shift, material and operator. Improve a confirmed cause, then control the process. Do not inspect quality into the product after production or pressure workers to conceal outliers.

Top practical tip

Choose projects from strategic and customer priorities, validate the measurement system and require evidence that the proposed change addresses a cause.

Top pitfall

Do not optimise a small process metric while damaging the wider system. Local savings, certification activity and statistical sophistication are not substitutes for customer and enterprise outcomes.

Further reading

For more on Lean Six Sigma analytics see for example:

• Pande, P.S., Neuman, R.P. and Cavanagh, R.R. (2000) The Six Sigma Way, New York: McGraw-Hill

• www.sixsigmaonline.org

• www.6sigma.us

• http://www.tutorialspoint.com/six_sigma/six_sigma_measure_phase.htm

• http://www.ehow.com/about_7296799_difference-six-sigma-six-sigma.html

• http://www.processexcellencenetwork.com/lean-six-sigma-businesstransformation/articles/case-study-using-six-sigma-to-reduce-excess-invent

• http://www.qualitydigest.com/inside/quality-insider-column/threetypes-lean-six-sigma-projects.html

• http://www.miconleansixsigma.com/six-sigma-tutorial.html