Poindexter Upgrades Ad Optimization Technology
The new feature in the New York ad technology company's Progressive Optimization Engine 3.0 lets publishers run an auction system for determining which advertiser will yield the highest effective CPM. POE 3.0 analyzes which ad placement will perform better by using predictive and clustering modeling to determine the best audience for the ad.
The system allows for more variable pricing, which Poindexter founder and chief technology officer Joe Zawadzki said helps publishers and advertisers.
"You can't just have a single CPM -- you'll under- and overcharge," he said.
The auction pricing is the publisher version of Poindexter's PassBack technology, which lets advertisers gauge ads that are underperforming and return that inventory to publishers.
Unlike companies in the behavioral targeting space, such as Tacoda Systems and Revenue Science, Poindexter works with both publishers and advertisers.
"The lack of info is hurting everyone," Zawadzki said. "We're trying to help both sides."
POE 3.0 also incorporates six clustering models, including Bayesian Belief Networks, and lets them compete for the most effective in predicting effective ad placements.
"We're going to have a market-driven approach to how the optimization is applied," Zawadzki said.