Epsilon Wins DMA Analytic Challenge

DALLAS, TX — Marketing Services Leader Epsilon announced today that the firm won the 2009 DMA Analytic Challenge. This year’s challenge, sponsored by CAC Group and the Direct Marketing Association (DMA) Analytics Council, was to apply primary research segments to a customer database.

According to Bryan Kennedy, President and CEO, Epsilon, "Best-in-class analytic discipline and expertise have been a cornerstone of Epsilon’s offering for decades. As new marketing channels emerge and the complexity of marketing multiplies, the importance of sound, scientific analytic techniques for understanding and predicting consumer behavior is paramount. This win validates the leadership position Epsilon maintains and reflects the challenges our team solves for clients every day.”

Competing against 13 other teams, Epsilon’s team of six used software tools, variable creation, and advanced modeling techniques to outperform their competition. The judges cited Epsilon’s team approach to discussion and debate, which spurred and refined new ideas and decisions. As the top-performer, Epsilon’s team ranked highest on both model-building and validation among the three finalists.

“I am proud of our team and their accomplishment. The team developed a great solution for a thorny problem -- one that many marketers face. They leveraged our core assets for this challenge including great tools and techniques, tremendous depth and breadth of capabilities, and a team-based approach for developing the best solution,” said Michael Penney, Executive Vice President of Strategic and Analytic Consulting at Epsilon.

Brad Rukstales, President of CAC Group and moderator of the Analytic Challenge said “Epsilon had a unique analytic approach to managing the challenges of this year’s competition, and it led to the winning entry. We appreciate the time and effort that goes into the challenge every year and congratulate Epsilon on their ingenuity and creativity.”

The Analytic Challenge presented a case study client with five well-established, primary research segments. The task was to apply the segment definitions into a direct marketing strategy. Challenge participants were provided with data to build a classification algorithm for assigning the segments to the customer database. There were four steps of analysis: Sample Selection, Variable Creation, Modeling and Assignment. The participants were evaluated based on the percentage of observations classified into the correct segment among a blind sample of respondents.