TC50: DataXu optimizes ad campaigns in real-time

Online advertising took a beating in 2008, but the market continues to rebound. Advertising startup DataXu — pronounced “data zoo” — displayed a new technology at the TechCrunch50 event that optimizes online ad campaigns in real-time across several exchanges including Google and Yahoo’s marketplaces. DataXu uses machine learning techniques to analyze vast amounts of data, including clickthrough rates and purchasing patterns.

Backed by Flybridge Capital and Atlas Venture, the Boston-based startup says its technology enables companies to serve the best ad at the right time for the right consumer. DataXu’s platform launch comes as Google is revamping its ad exchange to let people bid on individual impressions including display ads and other rich media formats. Microsoft and Yahoo are also stepping up to offer similar real-time bidding.

So assume a company buys impressions across many networks for many locations during different hours of the day. DataXu will parse through that data and find out very specific ways to optimize the campaign — for example, ads for a new restaurant in Chicago might have the highest clickthrough rates at 4 p.m. in the Chicago area with a particular slogan. DataXu will also feed back information on whether Google’s exchange is more effective than AdMeld’s for example.

The company says its algorithms are rooted in technology that guided NASA’s Mars mission plans. If a company’s ad campaign improves, DataXu gets a share of that increased revenue. It also works on a cost-per-impression basis, or on the volume of ads that are channeled through the platform and shown to consumers.