“The strategy we designed with Xaxis was to develop the pieces of the ad tech stack that gave the most competitive advantage to our clients,” said Xaxis CEO Brian Lesser. “That’s the DMP without a doubt. That’s not to say other pieces of the ad tech stack aren’t important, because they are, but data is where you get leverage.”
Prior to Turbine, Xaxis used 24/7 Media’s Zeus, a “first-iteration” DMP (WPP acquired 24/7 in 2007 and merged it with Xaxis in December) built with IBM’s data products subsidiary, Netezza. Lesser said Zeus “is still an important part of our business” – Xaxis continues to use Zeus for analytics, but not to build segments for targeting, which is now Turbine’s purview.
Consequently, Turbine’s value proposition revolves around its ability to build audience segments quickly, and to make immediate alterations to those segments as necessary. The advent of large, unstructured data sets (or big data, to use the buzzword), made the upgrade necessary.
“We looked at the problem of disparate data and increasing amounts of data sources and said we needed a different kind of solution,” Lesser said.
One of the benefits clients get when using an agency-owned DMP is greater reach – a benefit that overrides the risk of having data housed on a shared software platform. Many agency DMPs have the ability to partition data from different clients to keep those valuable assets separate.
“If you’re a single advertiser running your own DMP solution, you only can reach the people you’ve seen in the past,” said Scott Hagedorn, CEO of Annalect (the data hub fielded by WPP rival Omnicom, which standardized on Neustar Aggregate Knowledge’s DMP). “The ability for us to unite our clients and supply them with 100% reach of the Internet everyday, which is what we have the ability now to do, is more interesting than the concern with commingling cookie data.”
For Xaxis’ Turbine, that greater reach comes from its hook-in with publishers, through which it can anonymously pool data, find similarities and patterns in consumer activity and use that to inform when marketing messages should be sent. However, Susan Bidel, senior analyst at Forrester Research, questioned this feature.
“To the best of my knowledge, Xaxis doesn't have a first-party (publisher) relationship with anyone,” she said. Lesser declined to state how many publishers have latched into Turbine.
Bidel also wondered about conflicts with other platforms already in use by the publisher community.
“Since the various publishers they intend to work with are, by definition, operating with different DMPs, there will be challenges to normalizing that data,” she said.
Lesser said this isn’t an issue.
“In fact, we encourage publishers to license DMP technology or build their own technology so they have the best understanding possible of their own customers,” he said. “That helps us in our endeavor to better understand where those customers will be online and what type of advertising will serve them.”
Still, Xaxis is shouldering a considerable risk building and maintaining a DMP in-house. Following the initial multimillion-dollar cost of building it comes the substantial cost of maintaining the technology and ensuring it’s up to date and on par with other DMPs in the market.
But Lesser emphasized Xaxis doesn’t intend for Turbine to compete with the venture-backed DMPs, since he realizes that while Xaxis fields about 300 engineers, it can’t develop every part of the stack and keep up with the aggressive pace of investment in ad tech. But Lesser said he’s comfortable with Xaxis’ continued ability to invest in the product.
And, in building a DMP, Xaxis can control its fate.
“There’s a risk that any company we partner with could be acquired or become obsolete or be eclipsed by something new,” Lesser said. “We hold ourselves to a high standard in terms of what we do develop.”
That Xaxis built its DMP in-house contrasts with the philosophy of rival trading desks. For instance, Omnicom’s Accuen uses Turn’s DMP (as well as its demand-side platform and analytics platform) while Annalect, as mentioned, uses Neustar Aggregate Knowledge’s DMP.
For a time, Hagedorn said Annalect had considered building its DMP in-house. “Relative to building our own and maintaining it internally, in the era of consumer privacy, we wanted the DMP technology and the partner we were going with to be auditable by a client and not have it as an internal black box,” he said.
Intentionally or not, this speaks to one of the criticisms commonly levied against Xaxis. Some large brands have declined to participate with Xaxis because of its black-box model. Kimberly-Clark and Unilever, two of WPP’s biggest clients, run their programmatic campaigns through WPP-owned Mindshare, independent of Xaxis’ involvement.
To this point, Lesser noted that the 2,700 clients working with Xaxis know exactly the nature of its business model. “Many companies in the industry like to criticize us about transparency when in fact we’re very transparent in saying that because we make investments in data technology, inventory and people, we’ve developed a specific business model to reflect that,” he added.
Nevertheless, Turbine is only available to Xaxis clients.
“GroupM has other relationships with DMP providers and other solutions for clients that are not clients of Xaxis,” Lesser said, adding that Xaxis works closely with all the GroupM agencies “so they have the resources they need to build their own programmatic programs.”
Email This Post