For years ad networks, particularly behavioral targeting networks have sold “audiences” rather than “inventory.” Dissatisfied with those offerings, Interpublic Group-owned Mediabrands has launched its own platform for segmenting audiences and targeting them across digital media. The agency holding company is among a new breed of large agencies building out technology offerings.
People guiding the new firm, dubbed Cadreon, believe they have a leg up on ad networks or exchanges, in part because they’re able to apply advertiser client data to the ad targeting process, something missing in common ad network buys. “The biggest [difference] would be our use of data to craft audiences around exactly what our clients know their audiences to be,” said Michael Brunick, VP, media technology director at Mediabrands agency UM. He argues that the typical ad network “shoehorns” advertiser target audiences into predetermined segments.
The firm launched Cadreon yesterday following about nine months of testing involving 55 advertiser campaigns. The company name is a play on the word “cadre” to express the focus on segmenting groups of people.
As digital media have made agencies more accountable to their advertiser clients, large agencies and agency holding companies firms like WPP and Mediabrands have chosen to build or buy their own data-driven platforms rather than rely on outside firms. Large agencies and agency holding companies are beginning to look more and more like technology firms, competing with the vendors they once partnered with. WPP, for example, acquired ad management provider 24/7 Real Media and invested $25 million this year in analytics and optimization firm Omniture.
“It’s kind of like an arms race to see who’s going to do it at scale first,” said Brendan Moorcroft, UM’s SVP and director, global digital strategy, regarding the recent wave of tech developments among agencies. Indeed, Moorcroft expects Cadreon will eventually sell access to its platform to other agencies and directly to advertisers that don’t want to work with agencies.
For now, however, Cadreon will serve other Mediabrands agencies including Initiative, UM, and MAGNA. The new platform was developed in conjunction with Microsoft — a client — and various technology partners. As opposed to buying inventory and segmenting it for advertisers afterwards, the firm buys it specifically for each advertiser campaign in order to target audience profiles carved from demographic and psychographic data, online behavioral information, and other proprietary CRM data from advertisers.
The company thinks its ability to apply that proprietary data to segment audiences is unique because most advertisers are not comfortable supplying such data to networks; also the process is intended to be more efficient because the platform works across chosen networks and exchanges. Typically media buyers would have to determine audience segments on a network-by-network basis.
Through the system, the company buys inventory from brand name publishers, smaller niche sites, and transparent networks and ad exchanges. “We don’t want to just buy inventory upfront and segment it out,” said Moorcroft. Another plus for advertisers is that the platform is designed with their needs in mind, rather than from a publisher-centric inventory management perspective. “Our algorithms are the inverse [of most networks and exchanges],” said Moorcroft. “They optimize to the performance that we set at the beginning of the campaign.”
Cadreon currently enables targeting for online display, rich media, and video advertising. In the next month, mobile and digital out-of-home advertising will be available. The firm also plans for digital television buys later this year and is testing a dynamic ad creation tool.
Despite the fact that it faces growing competition from Facebook, Instagram and Snapchat, Google-owned YouTube is still one of the most popular ... read more
Amazon prides itself on being the most “customer-centric” company in the world, but according to investigative journalism non-profit ProPublica, Amazon’s algorithms are often anything but ... read more