Poindexter Adds Publisher Inventory Optimization

Poindexter Systems, whose technology has traditionally helped advertisers optimize online campaigns, has unveiled a new release that helps publishers optimize their inventory using an auction-pricing model.

Poindexter’s core product, the Progressive Optimization Engine (POE), previously provided information on audience segments to help advertisers optimize performance. With the new release, POE 3.0, the company is providing a tool with which publishers can segment their audiences.

The company beta-tested the new approach with a publisher, a portal site, for six months. “They had more than double the performance,” claimed Joseph Zawadzki, founder and chairman of the board of Poindexter. Zawadzki would not further identify the publisher.

The new capability pits Poindexter against behavioral marketing firms Revenue Science and Tacoda Systems, which work exclusively with publishers to optimize their inventory.

With Poindexter’s system, advertisers can bid for specific customers on a publisher’s site, such as 18-to-23-year-olds living in the southern U.S. logged on at a specific time.

“One advertiser, say a credit card marketer, says, ‘I want people who look like this, people at this daypart and this behavior, and I’m willing to pay $20 for this person.’ The other bidder, maybe an airline, is only willing to pay $5 for this person. So the credit card marketer wins,” said Zawadzki.

Poindexter will charge for the new optimization tool on a performance basis, Zawadzki said. “There is a performance element that scales with how we deliver value. If we deliver 10 percent lift we make X, and if we deliver more lift we make more,” Zawadzki said. He declined to give more specific numbers. (Competitor Revenue Science uses a similar pricing model.)

Poindexter will run a control group not using the tool to gauge how many impressions the publisher would have gotten without optimization and compare the two numbers, he said. “The amount [we charge] is based on a percentage of the gains we deliver.”

Joe Germscheid, a media buyer at Zentropy Partners, said, “Poindexter is going to go after the market. It’s a good thing.”

Germscheid said his company has used Poindexter’s product for a couple of years “and put together some good custom programs, so I know they’re totally capable of doing that.”

According to Mark Redetzke, a partner at online retailer Second Act and a former media director, “this squarely places the burden of fact on agencies and the advertisers to understand what their target consumers are worth.” Redetzke said it’s “unlikely” most agencies understand what reaching their customers is worth.

“The advertisers and agencies such as Digitas and Zentropy who have worked with Poindexter know better what their customers are worth,” Redetzke said.

For others with less experience, it will be a trial and error process, he said.

“There’s always room for intelligent products in the marketplace,” Redetzke said. “The data-driven approach behind the Poindexters and the Tacodas is just the beginning of the way people will market in the future.”

Indeed, behavioral targeting is heating up. Interactive marketing and technology company 24/7 Real Media launched a behavioral targeting product in March. Tacoda Systems and Revenue Science have rapidly been winning converts, in part as a way to handle problems with inventory shortages.

Poindexter’s new release has a number of other new capabilities, including multiple competing models for slicing and dicing data. New proprietary algorithms have also been added.

The new release also beefs up Poindexter’s ability to incorporate and optimize data from a multiplicity of channels, including graphical ads, search, email, “almost any digital marketing vehicle,” Zawadzki said. This means Poindexter can optimize a wider variety of campaigns in different channels.

Related reading