Database marketers: time to flex your analytical muscle
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 their results.
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his firm's technology's superiority with a mathematics-based solution, pointing to all other methods as inferior and outdated.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 your company.
David King is CEO of Fulcrum. Reach him at email@example.com.