In The Media
Probing the Limits: Statistical Quality Techniques
Source: Quality Magazine, February
IImagine using statistical quality techniques in a marketing
department. Sound unlikely? It doesn't have to be. In fact,
a new marketing company is using Taguchi methods of designed
experiments to improve the effectiveness of its clients' advertising
campaigns. This is an example of how the quality profession
can be revitalized by deploying quality techniques in non-traditional
The company by James Kowalick, an engineer who previously
used Taguchi methods to design engines. Kowalick, and his
partner, Mario Fantoni, founded their company to deploy designed
experiment techniques to advertising campaigns.
The purpose of a designed experiment is to identify the variables
in a process that have the main effect on the process outcome.
Understanding the critical variables is valuable in running
an efficient process.
A designed experiment leverages statistics to test many variables,
and combinations of variables, with just a few tests. Tremendous
insight is gained with relatively little effort and time.
I've seen the power of a designed experiment firsthand. When
I worked as a food-packaging engineer, there was a debate
on the packaging floor about how to control the bag-sealing
process. We argued about the variables that affected the strength
of the bag seal - was it seal time, seal temperature, seal
pressure, seal bar width, seal bar texture or a combination
of variables? Each operator set up the sealing machine based
on his own theory about which variable was the most important
in creating a strong seal.
To resolve the debate, we used Taguchi's methods to plan a
simple designed experiment, and a few tests, and a few hours
later, the debate was over. Seal pressure was clearly the
one and only variable that affected seal strength. We put
tighter controls on seal pressure and less effort and money
in controlling the other variables. In a short amount of time,
efficiencies and quality improved and the debate ended.
I think of that seal strength argument and designed experiment
any time I'm in a meeting where people are debating a topic
using nothing but guesses and speculation as arguments. It
is mind-boggling to think about all the waste that occurs
in organizations caused by "winging it" instead
of doing some simple Taguchi-like testing to gain insight
on how the process really works. This is especially true for
areas of the company where quality professionals rarely reside,
such as marketing and sales.
I applaud Kowalick and Fantoni for starting their company
and recognizing advertising as a process that can be greatly
improved. Like any designed experiment, their goal is to identify
key variables that affect the response rate for the advertising
process. Typical print advertising variables include text
content, color, artwork and the use of humor. A designed experiment
is conducted by testing a few different combinations of these
variables in small mailings. By looking at the response rate
for each combination, they learn what variables are key in
getting a high response rate, and then they use those key
variables to create the final print advertisement for the
mass mailing. The company reports greater than 1,000% improvement
in response rate using this method, and I don't doubt it.
I've worked with advertising firms and I've seen my share
of marketing "quackery" from them. First, these
advertising executives try to get you to believe that they
understand your market (which they don't), then they sell
you on an advertising campaign by using a flashy presentation
and an impressive demonstration of self-confidence in what
they think would be a good ad. It's all about selling their
ad concept to the company, not about developing an effective
message to appeal to the consumer. In short, they know nearly
nothing about what the customer wants and what type of advertisement
will appeal to them. It is hard to imagine how much money
is wasted each year on clueless advertising that fails to
send an effective message to consumers.
The quality profession is going through some tough times now,
but companies like Kowalick's show the potential that quality
techniques have to yield tremendous paybacks. It is encouraging
to see companies like this one start up and demonstrate how
quality techniques can be deployed in nontraditional ways
to yield bottom-line results. They are a great example for
the rest of us to follow by showing how changes in the profession
can put quality professionals back into positions of strategic
importance within companies.