Statistical modeling is one of the foundations of successful
direct and database marketing. But, while business managers recognize
analytics’ value, they often either underestimate or overrate its potential,
requiring database marketers to constantly strive to demonstrate the value of
models and analysis — particularly in a trying economy. Following are five
effective ways to increase analytics’ recognition and influence.
First, models are not the answer. Many business users
commonly believe that having a model will solve core marketing problems. Make
it clear that having a model to predict customer attrition will not reduce
attrition, just predict which customers are most at risk of leaving. The steps
that address the root causes of attrition are more relevant marketing, fixing
customer service problems and developing more competitive products, for example.
A good database marketer will also contribute to more than the analysis by
helping the organization formulate better marketing strategies and measuring
Second, seek out simplicity. The best model provides the
greatest business value with the least complexity. However, this does not mean
using the simplest statistical techniques. Many simple techniques, such as
cross-sell models, require considerable effort to use or implement. The usual
practice — building a propensity model for each product category using
relatively basic statistical methods — can be complex and expensive. Using more
sophisticated means, it is possible to build a single (and more useful) model
to simplify implementing greatly.
Third, get a bigger tool box. Lately, database marketers
seem to labor under the old saying, “If all you have is a hammer, everything
looks like a nail.” Many organizations are stuck using preferred techniques
such as RFM and logistic regression to solve every problem. There are many new techniques being developed
both in academia and industry all the time. Upgrade your skills and your
company’s capabilities by being open to new methods and actively seeking out
the experts knowledgeable about them.
Fourth, grow by outsourcing. If you don’t have adequate
knowledge or resources on staff, consider outsourcing. But beware the vendor
who insists that you just need the equivalent of a new hammer, like a representative
who proclaimed his firm’s technology’s superiority with a mathematics-based
solution, pointing to all other methods as inferior and outdated. It may have
been a perfectly good application, but good analytical minds are creative about
the solutions to problems, and someone who comes to the table with one — and
only one — answer is unlikely to provide a long-term answer.
Finally, build a stronger team. Get people that work well
with others in the organization and that can translate the technical details
into language that business users and managers can understand. Make sure that
people on your team develop the strongest possible technical proficiency. The
stereotype of the brainy analyst with low social skills unfortunately
encourages organizations to emphasize technical skills, but effective
communication and business knowledge can be taught and developed, too. Build
your analytics muscle by getting analysts fully engaged in solving the
business’ problems at all levels, making it more relevant and important within
King is CEO of Fulcrum. Reach him at [email protected]fulcrum.com.