What is „Six Sigma“?
Six Sigma is a structured, data-oriented method to reduce defects, waste and quality issues of any kind in production, services, management and further business activities. Six Sigma is a trademark of Motorola, that shaped the term in the eighties. The method is based on a combination of established techniques of quality assurance, basic and advanced methods of data analysis and systematic training of the staff involved with the Six Sigma related processes.
Why is Six Sigma popular?
Six Sigma has shown that it not only increases the quality, but also that it can lead to tremendous cost reductions. Some spectacular success stories from huge companies have been published. Jack Welch (the former CEO of General Electric, one of the world’s largest manufacturing companies) has said: “Six Sigma is the most important initiative GE has ever undertaken…it is part of the genetic code of our future leadership.” Welch attributes Six Sigma for billions of dollars in cost reductions.
Many other companies claim huge cost reductions after introducing Six Sigma at their production sites, as well. Motorola (one of the leading members of the group of companies that developed the Six-Sigma-approach) for example claims to have saved 11 billion dollars after introducing Six Sigma twelve years ago. Allied Signals mentions cost reductions of about a billion dollars over a couple of years.
Statistical Background
The term six sigma represents the statistical goal to reduce the number of defects to a negligible amount, corresponding to the six sigma value of a (corrected; see below) normal distribution curve: Six Sigma tries to push failures and quality issues to the outer bounds of the distribution to reduce the problems to rare outliers of an otherwise faultless process.
To reach this “Six-Sigma-Goal” a process cannot have more than 3.4 defects per 1 million possibilities, where defect is defined as any form of undesired outcome for the process under inspection. Please be aware that 3.4 failures per million in fact represents a z-value of 4.5 of the normal distribution instead of 6 as the method allows for a dynamic shift of 1.5 sigma (defined by Motorola as long term dynamic mean variation).
Therefor one of the basic Six Sigma related tools is the Six-Sigma-Calculator to calculate the number of failures for one, two, .., six sigma Processes. Naturally, there are much more advanced techniques to be applied based on the processes over the different levels of a Six-Sigma-project.
How to Apply Six Sigma?
The strength of Six Sigma lies in the empirical, data-driven approach (and the use of quantitative measures of performance) to reach the goal of process improvements and reductions in variation. In “Six-Sigma-Quality-Improvement-Projects” the work is organized following the Six-Sigma-DMAIC model:
Define:
In the definition phase the goal and scope of a project are defined, issues are collected that need to be tackled to reach a higher (better) sigma level.
Measure:
In this phase data is gathered about the current situation to form a baseline of the process and to identify problems.
Analyse:
Identification of the causes of quality issues and confirmation of these using data analysis.
Improve:
Implementation of solutions that have been developed based on the knowledge gained in the Analysis-phase.
Control:
This phase ensures that the improvements implemented in the previous phase stay valid and active
Each of these steps makes use of specific analytical (quantitative) methods that are part of the overall spectrum of methods suggested for Six Sigma.
Further information regarding Six Sigma can be found in two books that discuss the Six-Sigma-methodology and its application: Six Sigma: The Breakthrough Management Strategy (2000) by M. J. Harry and P. Schroeder and The Six Sigma Handbook (2001) by T. Pyzdek.
Statistica
Statistica supports the data collection and analysis on every level of a Six-Sigma-project and can hence serve as the analytical basis of Six-Sigma-initiatives and implementations for companies of any size. The software has the following functionalities:
- Comprehensive set of Six-Sigma-tools like the Six-Sigma-calculator, Six-Sigma-reports with integrated graphics and cause-and-effect-diagrams (Ishikawa)
- A Six-Sigma-menu that is organized according to the “Six Sigma DMAIC”-approach. Additional Six-Sigma-related tools can be embedded in the UI
- Calculation of the process capability respecting the time-dependent distribution models according to DIN 55319 / ISO 21747
- High quality and flexible graphing capabilities. These capabilities are also accessible via programming (in Statistica Visual Basic) which offers a nearly unlimited potential to customize and build tailor-made visualizations.
Advanced Methods
Statistica supports advanced methods like Machine Learning and Data Mining algorithms to be applied as alternative approaches in Six-Sigma-projects. These can be the foundation of an innovation advantage over to the competition.
Additionally, the workspaces provide a visual interface to design and automate analyses. Using workspaces, it is possible to share the work between all involved team members easily. For example, a Black-Belt might design analyses as a workspace to be later used by Green-Belts for application in the field.
Further advantages are:
- Implementations of Machine Learning and Data Mining methods for classification and prediction including artificial neural networks, decision trees, support vector machines and others.
- Algorithms specifically suited to work with large data sets
- Extraction of process related information from historical data using Feature Selection as alternative to expensive experimentation.
- Workspaces to automate analyses and data preparation (feature engineering, cleaning, filtering, shaping etc.)
Organization-Wide
With Statistica Server it is possible to design company-wide applications for quality assurance and improvements using Six Sigma. The customization capabilities allow to shape Statistica into a tool that feels and acts like it was specifically designed for your needs. The platform provides:
- Real-time monitoring and alarming for the production site, analytical tools for the engineer and reporting options for the management.
- Centralized setup and management of database queries and analysis templates for shared use of data sources and specific applications.
- User-specific interfaces for all involved parties: from simple UIs for the worker to higher methods for the Green-Belts to highly complex environments for analyses and data mining for master Black-Belts.
- The platform is not only a comprehensive environment for complex Six-Sigma-analyses and data mining, but it also serves well as a training environment.
- Server-based monitoring of processes and quality improvements for an automated control phase with flexible alarms.
- Scale-able, customizable, and easy to integrate into existing database and ERP-systems.
- Extract, Transform and Load (ETL) data from heterogenous data sources.
- Scheduling can be used to produce reports with up-to-date data.
- Execution of resource intensive process on the server.
- Accessible via web browser and can be tailored to specific user needs.
- The integrated customization options allow for tailor-made web-based information portals to allow access to current quality analyses of the Six-Sigma-projects.
- Simultaneous monitoring of thousands of quality measures on one or more servers and easy management from a client PC.
- Centralized installation and management of your Six-Sigma-software.
StatSoft as Your Partner
StatSoft is your reliable partner when it comes to the software Statistica. We assist you with the selection, configuration and installation and will also provide you with the needed software- and methodological-know-how. In analytical projects we can aid with consulting and execution. Together we can generate more insights from your data and deliver a sustainable value.