Lean Six Sigma

Lean Six Sigma is an improvement methodology that blends the best of Lean and Six Sigma is a disciplined approach to improvement and problem solving. Six Sigma was originally designed to achieve near perfection in product and service quality through variation, error and defect reduction – near perfection defined as less than 3.2 errors per million opportunities.

Motorola Engineer Bill Smith originally developed Six Sigma in the 1980’s  to simplify Continuous Improvement methodologies of the day. Unfortunately, and for a variety of reasons, Six Sigma itself has grown into horribly complex monster. Lean Six Sigma attempts to address this by taking the best of Lean and Six Sigma improvement methodologies and blending them into a consistent, and simplified, improvement framework.

Six Sigma and Lean Six Sigma both make use of the DMAIC problem solving model, a variation of the PDSA model developed by Walter Shewart and Edwards Deming. The PDSA Cycle is in fact, at the core of the DMAIC model (as illustrated in the diagram).

DMAIC is usually placed within an organization context through the addition of Recognize and Sustain phases of the model.

Recognize represents the efforts of the organization to identify the problems that exist within the organization and prioritize them so as to set an agenda for change. DMAIC is then applied to address the problem assigned through the Recognize phase.

Sustain embeds the solution developed through DMAIC to other areas of the organization — in essence ensuring that organizational learning and adoption is an explicit function of the model.

In between the Recognize and Sustain components of the model of course, is DMAIC itself. It defines a problem solving approach of:

Define: the performance problem to be solved. Generally the approach is to boil the improvement effort down to a single customer driven performance metric.

Measure: gather hard data relating to the problem. Focus is palced on potential causal factors driving performance.

Analyze: the data gathered through the measurement phase.The emphasis here is on analytic methods that attempt to confirm causal effects.

Improve: implementing evidenced-based solutions that have a hig probability of effectively addressing the performance problem.

Control: adjusting and finally standardizing the solution so that we don’t loose the gains made. Essentially hard wiring the solution into the way things are done.