The I in DMAIC is about redesigning the process to achieve improved capability. Prior to the Improve phase being completed, the process must be run with its updated configuration and settings, and the new short-term sigma level must be confirmed.

Common tools in the Improve phase

Some of the more popular tools in the Improve phase include regression, hypothesis testing, designed experiments, analysis of variance (ANOVA), and simulation.

New operating tolerances must be established and confirmed for the “vital few” causes identified in the Analyze phase.

The difference between the short-term and long-term sigma levels

A project has achieved Six Sigma capability when short-term data shows six standard deviations between the process average and the nearest specification limit.

The specification limit is the point at which the product or service will no longer meet customer expectations.  For a product, the specification limit is a measurable product feature, such as interior noise in an automobile.  For a service or transactional process, the specification limit might be time based, such as the amount of time a caller waits to speak with a live person.

The long-term sigma value is estimated by subtracting 1.5 from the short-term sigma value. This subtraction takes place due to the 1.5-sigma mean shift that is assumed to take place over the long run. Thus, a short-term Six Sigma process will have a long term sigma value of 4.5, and a theoretical long-term defect rate of 3.4 PPM.