When Dentsu Aegis Network acquired performance agency Merkle last year, it gained deep expertise in data-driven marketing as well as a data hub for its agency brands.
That hub is called M1, Merkle’s data and media planning tool launched last year. M1 can pull in an advertiser’s data and match it to personally identifiable information (PII) on more than 280 million known user IDs from Merkle’s days as a direct marketing agency to create a privacy-compliant universal ID.
M1 can then sync that ID to a publisher’s audience data to find exact impression matches. Merkle has signed on AOL as a partner for this two-way ID sync, and will bring on CBS, ESPN, Weather, Pandora, Comcast and more publishers in the coming months.
“We’re using a secure, standard key that publishers link to their known audience and we link to our advertisers’ audience,” said John Lee, chief product and data officer at Merkle. “M1 can match directly across everything AOL [for example] has on owned-and-operated as well as programmatic inventory.”
Merkle has only used its ID sync with a dozen or so clients, but early results are strong. Performance varies from around 20% to the “very high double-digit range,” Lee said.
Using ID sync, one financial services client reduced cost per lift on mobile by over 80% and on desktop by over 25%. Another client saw 60% unique reach when running its segments against M1’s ID sync publishers versus the open exchange.
“We’re reaching people we set out to reach and avoiding issues like finding a person with six cookies,” Lee said.
Lee spoke with AdExchanger.
AdExchanger: How does Merkle keep PII data compliant for media?
JOHN LEE: We’re the matchmaker. M1 leaves off by telling us that our advertiser’s audience matches to a set of IDs on one of our publisher partners. Those audiences get syndicated and pushed into the activation platform.
We manage a very tight vault that has PII associated with it. That gets translated into anonymous IDs to then trade with a demand-side platform and allows us to trade hashes with [walled gardens like] Facebook or AOL. We can trade on a known user level without actually transmitting PII.
How do you work with other agencies part of Dentsu Aegis Network?
We can work with any client across the network to build and manage a first-party CRM test infrastructure. We can also act as a systems integrator for clients that are looking to build an in-house infrastructure.
We also act as an advanced analytics agency in complement to our sister agencies. We can develop very sophisticated segmentation, predictive analytics and attribution analytics for clients.
As part of a big agency network, how do you justify helping clients bring data and media in-house?
Our business model is to help clients reach their business goals, regardless of which model.
If Disney wants to bring programmatic capabilities in-house, which happened, that’s a decision that aligns with their strategy. If we can help them be most successful, which we did, we’re happy to do that. That’s a meaningful engagement for us.
Do you have ongoing relationships with clients after helping them launch in-house desks?
We played a hand in helping two Fortune 100 brands stand up in-house media trading functions. Both relationships go back at least two to three years.
We’re still engaged with those clients, whether that’s from an M1 platform, analytics or ongoing technology advisory perspective, or some combination.
Holding companies are positioning around owned data and technology. How does Merkle play into that trend?
You don’t need to look further than Omnicom, WPP and others to see more aggressive investment and positioning around data and technology as an owned asset. Merkle was Dentsu’s manner of addressing those assets.
Does owning technology create a conflict of interest for clients?
I would agree if we told [clients] they can’t have their own relationship with X, Y and Z tech providers. Clients are free to bring the best of whatever technology and data assets they have, even from some of our direct competitors, which is the case in many instances. Our data platform is interoperable with any of them.
For the Dentsu Aegis Network agencies, there is no mandate or requirement for any client to use our platform.
Can holding companies act as data co-ops as second-party data plays a bigger role in marketing?
An agency network [needs] the apparatus to take in and manage data as a safe harbor from distinct corporate entities and combine that data within a very tight set of parameters.
Very little of it happens right now but there is a good opportunity to do it. If we look at Merkle and its traditional data cohorts like Acxiom, Epsilon and Experian, it happens all the time. Agencies aren’t in that position technologically.
Does Merkle have a data co-op?
We’ve done bespoke matchmaking where noncompetitive entities can bring together data where they can be helpful to each other and there’s no conflict. We don’t personally happen to manage a co-op.
This interview has been edited and condensed.
Update: This story previously said Merkle has more than 1 million known user IDs. Merkle has more than 280 million known user IDs.
