From Monet to Money – How Statistics Can Impact Your Bottom-Line
Introduction:
From Monet to Money?How Statistics Can Impact Your Bottom-Line. Discover
how viewing impressionist paintings is similar to the big picture of statistical
usage to improve training effectiveness and bottom-line performance.
From
Monet to Money
How Statistics Can Impact Your Bottom-Line
by Dr. Paul Mullenix & Saleha Yusof,
SystatS Consulting Sdn. Bhd.
What does appreciating great works of art, such as Monets? water lily series
have to do with how statistics can improve your business performance? In both
cases you have to take a few steps back to see the big picture (Fig.s 1&2).
Fig. 1. The Water Lily Pond
(1920), Monet
Many companies train in SPC or other statistical methods without a good
understanding of how it all fits together. Consequently, they may not be getting
the best value for their training.
This big picture is organized around three primary activities driving business
processes, viz. characterization, improvement and control. These activities can
be applied to any value-chain function from R&D to Marketing to Sales to
Manufacturing to Customer Service, to name a few.
Fig. 2. Dollar Sign?The
Statistical Big Picture
The fundamental requirement of any business process is the ability to measure
the process. The validity of any measurement process requires a careful
characterization using a statistical measurement system analysis either tailored
to continuous or attribute data. This may recommend further controls for the
measurement process.
When data is taken on many metrics and consigned month after month, year after
year, to large databases, we end up with what statisticians call a data cemetery
(Fig. 3). This is where data goes to die?data goes in, but information never
comes out.
Fig. 3. Data Cemetery
To extract information from these data cemeteries, statisticians use
characterization methods such as Data Mining, also called Exploratory Data
Analysis or a Multi-Vari Study.
If sources of instability are found from the Data Mining, then root cause search
techniques such as 8D problem solving are used to improve the stability. With a
stable process, capability can be assed using statistical Capability Analysis.
If the capability analysis indicates improvement is necessary, then active
process improvement proceeds using Design of Experiments (DOE), Response Surface
Methods (RSM) for manufacturing cases (Fig. 4) or Business Process
Re-engineering (BPR) for service and other business processes.
Fig. 4. Example Experimental
Design
Armed with a complete characterization from either Data Mining or DOE/BPR, the
key factors needing control are now evident or may be derived from a Sensitivity
Analysis. Using Statistical Process Control (SPC), Engineering Process Control (EPC),
poke-yoke solutions or other methods, controls can be tailored to the process.
The steps we have outlined are also included in other unifying statistical
methods such as Six Sigma. These steps result in a statistically characterized
process which impacts the bottom-line by identifying opportunities for improved
cost, performance or quality. With improvement tools, these opportunities can be
turned into reality. With controls in place, these realities can be sustained.
So the next time you go to an art gallery, remember to take a few steps back?and
the next time you consider how to impact your bottom-line, take a few steps back
to consider the big picture of statistical methods for the big dollars.
Copyright 2005 by SystatS Consulting Sdn. Bhd.