Poindexter Nabs Amazon's Chief Scientist

Advertiser-side behavioral targeting provider Poindexter Systems named Dr. Andreas Weigend as its chief scientist today. Weigend previously held the same position with Amazon.com, where he developed predictive models for recommendations and cross-selling.

Weigend, an expert in data mining and consumer behavior, will guide Poindexter’s product development and strategy. The company’s second-generation ad-serving model incorporates clustering and predictive modeling to identify audience behavior to deliver “the right message, to the right customer, at the right time,” said Rich Person, chairman and CEO.

“Poindexter Systems and Andreas share a common vision — to make advertising so relevant to each consumer’s needs that it is a welcomed resource,” Person said. “We want to get to a point where people appreciate being presented with relevant information.”

Publisher-side behavioral targeting, such services from Revenue Science or Tacoda Systems, is a selling tool, making less desirable ad inventory more valuable by identifying user characteristics that are attractive to certain advertisers. Poindexter’s advertiser-side behavioral targeting is an analytical tool. It allows an advertiser to analyze user behavior, build an ideal customer profile, and serve relevant ads to users who match that profile.

“It takes the ‘plumbing’ component of first-generation ad-serving platforms and adds intelligence to serve more relevant ads,” Weigend said.

Weigend likens the service to search engine marketing, where ads are served based on the keywords a user searches. To that explicit information, Poindexter adds implicit information based on data models, including information about where the user is coming from, and where he works, and uses it to predict online behavior.

“What we saw in search last year is what we’ll see in behavioral this year,” Weigend said. “Poindexter has the potential to change the world of advertising by making it highly relevant to consumers instead of wasteful or painful.”

At Amazon, Weigend ran hundreds of simultaneous experiments to gather and analyze information to predict user behavior. He’ll use the same approach at Poindexter, working to perfect the balance of combining past behavior with the user’s current situation.

“Having an agile way of doing testing will affect where the market is going,” he said.

Weigend has published over a hundred scientific papers and co-authored six books. He currently teaches graduate courses on data mining and technology innovation at Stanford University, and was a professor at New York University’s Stern School of Business and the University of Colorado at Boulder.

“A university is a great place for ideas, but it’s hard for bringing ideas to fruition,” Weigend said. “Companies like Amazon and Google are integrating ideas from academia, and that’s how they constantly innovate to stay ahead.”