The Tools and Methods of Six Sigma
March 31, 2004
Six Sigma is a buzzword that’s getting attention in all types of industries. The success of GE, Motorola, and others has prompted many an executive to declare that he/she wants “Six Sigma.” But what does that mean? What are the methods and tools of Six Sigma?
Generally, Six Sigma is concerned with reducing variability.
The goal of Six Sigma is a six-sigma level of quality or 3.4
defects per million opportunities. To accomplish this, projects are
identified and prioritized.
Then, project teams attack the problems using the DMAIC method:
Define, Measure, Analyze, Improve, Control.
Let’s talk about each phase and the tools associated with
each.
Phase 1: Define
In this step, the problem is defined.
This includes identifying the CTQ’s, the critical-to-quality
characteristics as defined by the customer and identifying the key
factors affecting the CTQ’s.
Phase 2: Measure
The measure phase consists of two major
steps. Step 1 is to identify
the internal processes that influence the CTQ measurements.
Step 2 is to conduct a measurement systems analysis, which
includes gauge studies and to perform a thorough process capability
study.
Phase 3: Analyze
In the analyze phase, the project team
determines which inputs are affecting the outputs.
This is done using the following three steps:
Step 1: Develop and state a hypothesis
about the cause or causes.
Step 2: Test the hypothesis through data
analysis
Step 3: If your hypothesis is correct, add
your cause to a list of significant causes affecting the process.
Phase 4: Improve
The improve phase involves quantifying the
effects of the key variables on the process and developing an
improvement plan that modifies the key factors to achieve the
desired process improvement.
Phase 5: Control
In the control phase, you put in place a
system to maintain the changes/improvements you developed in the
Improve phase. This phase
involves documenting and monitoring processes.
The Tools
There are statistical methods as well as
qualitative tools that are used for process improvement using Six
Sigma. This article will
explain some of the key tools in brief.
1. Process Mapping or Value Stream Mapping
One of the most important steps in process
improvement is to map out the process in question.
Process mapping is one way to accomplish this.
For the lean enterprise, Value Stream Mapping adds the
information flow dimension
to this tool. In general,
this involves mapping your product and information flow throughout
any given process. This can
be done in manufacturing as well as administrative functions.
2. Basic
Statistics
These
basic statistics tell the analyst the central tendency and the
dispersion of the data. They
consist of the following for a given set of data:
Mean:
Sometimes called the average or arithmetic mean, the
mean is the sum of a series of values divided by the number of
values.
Median:
The midpoint in a series of data points.
Mode:
The value that occurs most often in a set of values.
Range:
The range measures a spread of data points.
Very simply, it is equal to the largest value in a set of
data minus the smallest value.
Variance:
The variance is the average squared deviation of each number in a
data set from the mean.
Standard Deviation:
The most common measure of spread, the standard deviation is the
average distance from the mean.
It is equal to the square root of the variance.
Correlation:
The correlation is the degree to which two variables are related,
measured in terms of correlation coefficient.
3. Basic
Graphical Analysis
Histogram:
The histogram is a vertical bar graph representing the distribution
of a given data set. It is
also known as a frequency distribution bar graph.
Scatter Plot:
The scatter plot is a graph in which individual data points are
plotted in two dimensions.
Scatter plots graphically show correlation.
4. Cause
and Effect Diagram (Fishbone Diagram)
The
Cause and Effect Diagram or Fishbone Diagram is a
graphical tool for identifying the relationship between a problem
and its potential causes. One
of the most effective ways of constructing such a diagram is to
brainstorm potential causes in a team environment.
5. XY
Matrix
The
X-Y Matrix is used to link customer CTQ’s to process inputs.
It is simply a group of rows and columns, with one set of
increments marked along the X axis and another set of increments
marked along the Y axis.
6. Gauge
R&R (Repeatability and Reproducibility)
A
Gauge R&R ensures that you’re measuring what you think you’re
measuring. IT looks at your
units of measure and number of variables, calibrates the measurement
gauge, and then randomly selects samples to measure against
different operators.
7. Pareto
Chart
A
Pareto Chart graphically depicts the relative importance of
causes, defects, and other aspects of a process. It is used to
identify factors in a process that have the greatest cumulative
effect.
8.
Multivariate Analysis
Multivariate Analysis
offers a way to reduce possible causes of variation in a process to
a family of related causes.
Multivariate charts present an analysis of three main sources of
variation: intra-piece (within a single piece or batch), inter-piece
(piece-to-piece or batch-to-batch), and temporal (time related).
9.
Hypothesis Testing
Hypothesis Testing
defines the problem, statistically tests data assumptions, selects
samples, and determines whether or not the probability of a defect
is caused by random chance or has a more tangible cause hidden in
the process.
10.
Failure Mode and Effects Analysis (FMEA)
FMEA
is a disciplined methodology that identifies potential failures and
puts in place a plan to prevent them from happening.
There are two types of FMEA’s that can be performed: Product
or Design FMEA (DFMEA) and process FMEA (PFMEA).
DFMEA relates to product failures and PFMEA relates to
process failures.
11.
Design of Experiments (DOE)
DOE
is a powerful tool that systematically identifies and quantifies the
effects of two or more factors on the outcome of a process by
experimenting with many factors and variables simultaneously.
12.
Control Chart
The
Control chart is the fundamental tool of Statistical Process
Control (SPC). It determines
whether a process is operating consistently (in control) or if a
special cause has occurred to change the process mean or variance.
Click here to subscribe to our free e-newsletter Learning to Lean and receive three articles like this one each month.
About the Author
Darren Dolcemascolo is an internationally recognized lecturer, author, and consultant. As Sr. Partner and co-founder of EMS Consulting Group, he specializes in productivity and quality improvement through lean manufacturing. Mr. Dolcemascolo has written the book Improving the Extended Value Stream: Lean for the Entire Supply Chain, published by Productivity Press in 2006. He has also been published in several manufacturing publications and has spoken at such venues as the Lean Management Solutions Conference, Outsourcing World Summit, Biophex, APICS, and ASQ. He has a BS in Industrial Engineering from Columbia University and an MBA with Graduate Honors from San Diego State University.
EMS Consulting Group helps companies implement lean strategies through lean training and lean consulting services. To learn more, read our lean manufacturing case studies or lean manufacturing articles.