Knowing When To Shift Gears
September 7, 2004 - by Robert E.
Stevens,
GENESIS
II (The Second Beginning) E-Mail: views@aol.com
Market research and racing in the Indianapolis 500 are similar in many
respects.
Namely it is all about knowing when to shift gears. Can you imagine
driving
in the 500 with only one gear? You up shift on the straight away and
down
shift on the curves. You would never be competitive with only one gear.
The
same holds true for competing in the world of consumer products. If you
only
have one speed or only one way of doing research, you will never be
doing
your employer justice. I have never seen a real artist with only one
brush
or a master carpenter with only one tool in his/her tool box. It is the
same
for a master of market research. There will be many tools, each with a
distinct
purpose.
In the 500, the driver knows when to down-shift, when to accelerate,
and
when to make a pit stop. The driver also knows when to practice extreme
caution.
In market research, there are times for speed and a time for cautious
deliberation.
There are times for casual research, the learning and understanding
phase
of idea cultivation. There are other times in product development where
the
research is more like a Demolition Derby. Instead of a lot of wrecked
cars,
you have a lot of failed or not exactly the right kinds of
ideas/products. It is
the winner of the demolition derby that goes on to compete in the big
race --
the race for market share.
Unfortunately while the cars in the 500 have only one track to follow,
our
products not only have roads with turns but also forks where there are
options. We must be prepared
at times to stop and back up and take an alternate course. While the
race
car does not have a reverse, our product must have a reverse gear, and
we
must be prepared to use it. It is all about the "Fine Tuning" of our
entry
into the race for market leadership. The masters of race cars will have
tools
for acceleration, speed, suspension, steering, and in-race repair. We
must
have tools for identifying opportunities, turning opportunities into
visions,
visions into products, and products into brands. Our tools must be able
to
evaluate appearance, image, communication as well as performance.
If it were easy, there would be no challenge. And if it were easy, who
would
want the job? OK, I can think of some who would prefer the easy road,
and
I imagine you also have a few in mind. They are the ones with no tool
box,
not even a tool belt. They just carry their hammer wherever they go
regardless
of what task they are approaching.
It's a
question that comes up at least once a year.
The question has to do with Confidence Intervals. You know the "It is
accurate
to +/- 4 points." Most people who hear or see this number think that
the
"population statistic," the true answer, is within these limits.
Actually
the confidence interval is really a measure of the reproducibility of
the
test results. That is, if you would do the same test under exactly the
same
conditions, it would be expected that the results would be within the
limits
19 out of 20 times. Note, it is under the exact same condition
which means
that each test contains the same errors and biases. These are tests,
and tests
involve estimates. Your estimate, as it relates to the true statistic,
is
only as good as the test you conducted. There are two important
characteristics
at play here, reliability and validity.
The textbooks explain confidence limits as follows. "A sample estimate
of
a population value will almost never be exactly equal to the population
value.
However, most estimates obtained from probability samples are accurate
in
the sense that there is no bias present. In other words, there is no
tendency
for the estimate consistently to understate or overstate the quantity
being
estimated."
Sponsor: Sorensen Associates Inc
Portland, OR: 800.542.0123
Minneapolis, MN: 888.616.0123
the
in-store
research company™
-- Dedicated to the relentless pursuit of WHY?
![[Back]](../grey-left.gif)
![[Views home]](file:///C:/Personal%20Drive%20Mirror/POPSG/public_html/views/grey-up.gif)
![[Next]](../grey-right.gif)