Demystifying Agency Data Platforms

Advertisers want to market using the type of rich data provided by walled gardens – but that’s not possible on the open web. So the major holding companies have each developed data platforms to meet their clients’ needs.

The question is: What’s the difference between their offerings?

Publicis Groupe’s Spine, IPG’s AMP, WPP’s mPlatform, Dentsu Aegis Network’s M1 and Omnicom’s Annalect are all centralized platforms designed to resolve identities and organize a client’s first-, second- and third-party data around a known consumer ID.

Their methodologies may differ, but their value prop is the same: better management of reach and frequency and a more consistent view of the consumer across platforms.

Clients, however, can look in the not-too-distant past to weigh the risks of investing media dollars in an agency-owned technology – and how that can affect the client-agency relationship.

“Anywhere the agency is taking an upfront cost to acquire data and selling those services to their customers, it has stopped being an agent,” said Chris Kane, CEO at ad tech consultancy Jounce Media. “You’ve broken out of the role of being an agent to your customer and you’re trying to sell them a product.”

Identity Resolution: How It Works

The identity resolution process begins with an anonymized ID on a known consumer.

Because most agencies don’t have the skills to handle personally identifiable information (PII), holding companies buy anonymized, PII-based consumer IDs from data providers like Acxiom, Experian and Neustar. Agencies then onboard the IDs onto their platforms using a partner like LiveRamp. (Merkle, because of its heritage as a database marketing company, can process and onboard PII directly.)

The agency platform then appends to that ID whatever first-, second- and third-party data it has on the corresponding consumer. That can be CRM data, cookies from media events or third-party data from partnerships with MasterCard, Polk, IRI or another provider. Before activating the IDs, the agency platforms perform match tests with demand-side platforms to ensure their audience segments are aligned.

Assembling data around a known individual’s behavior makes it easier to manage reach and frequency across platforms, allows for better targeting and helps the holding companies define audiences consistently across their network of agencies.

That last aspect is important because multiple divisions in an agency work with these audience segments.

“Planning defines an audience and buyers redefine those audiences,” said Arun Kumar, chief data and marketing technology officer at IPG Mediabrands. “By bringing the audience piece up front, we define those segments and consistently use them across every channel.”

So what’s the difference?

Because these platforms aim to solve the same problem for clients, agencies can differentiate with the type of data they bring into their platform and the way they analyze and deploy it.

GroupM leans heavily into WPP’s Kantar and KBM for data that isn’t available off the shelf. Annalect pushes partners for respondent-level data so that it can segment the data itself, rather than buying the “pre-segmented and aggregate versions of comScore every media planner has,” said Erin Matts, North American CEO at Annalect.

“If you’re getting something off the shelf, you’re also getting the exact same data set your competitors might be getting,” she said.

Skill sets also matter. Because Merkle onboards data itself, clients can be more certain that its IDs are purely based on PII, said John Lee, president of M1 at Merkle. Merkle can also use PII to perform direct person-level matches with publishers in its Publisher Addressable Marketplace.

“[Other agencies] can inject PII through onboarding, but they’re not storing everything back to the person level,” Lee said. “In many cases, they’re limited to the data that a tag is firing on an ad. They’re dependent on contextual data.”

Integrations with other platforms in the agency’s stack can also be differentiators. IPG connects AMP into its Unity platform, which can port PII-based audiences into walled gardens such as Facebook, Google and Amazon for a consistent definition across platforms.

And access to data across markets is important, as it can vary significantly. Most holding companies are doing identity resolution in the US and UK, where third-party data is available from an array of providers. GroupM, however, has a leg up internationally because mPlatform, which is built off Xaxis’ legacy Turbine DMP, is already scaled in 49 markets including China, said Colin Barlow, global chief operating officer at GroupM.

“Once you get out of the US and the UK, the ability to purchase valuable data sets gets very thin very quickly,” he said. “We can react quickly to changes in legislation and the needs of our clients.”

And, as always for an agency, neutrality is a key. Without financial incentives regarding data, clients can feel safe the platform being used solely for their benefit, Matts said.

“We’re not making money in the supply,” she said. “We just don’t have those relationships.”

What’s The Catch?

While these platforms can help marketers manage their media spend more effectively, clients should be wary about how much control they give their agencies over their data. Clients should also be aware of any friction that could occur should they want to remove their data off the agency’s platform, said Jounce’s Kane.

“If your planning process gets intertwined with [an agency] system, even if you own the technology, the process they’ve built becomes very sticky,” Kane said. “That’s another hook now that lives with the agency.”

Clients who opt in to these platforms can protect their data by only sharing with their agencies access to what’s necessary for a campaign. Most GroupM clients, for example, tend to share “more selective data sets for specific reasons,” Barlow said.

“The client isn’t going to hand over the keys to all their data, and neither should they,” he said. “They’ll give us all the things necessary to make their media dollars effective.”

Clients should also be aware of the conflict of interest these platforms create in the agency-client relationship, Kane said. Marketers should demand the same level of transparency around how their agencies source the data, how much they pay for it and any markups involved the same way they do for their programmatic media.

Holding companies price these platforms most often on a percentage of media spend. In some cases, clients who want to use the platform only for segmentation can be charged for the services of operating the platform.

Agencies, however, are anticipating enhanced scrutiny over data ownership and contractual agreements as clients continue to pursue more control over their marking spend.

Annalect, for example, is starting to allow clients to own their own cloud-based instance of its identity resolution platform, which the agency would operate as a managed service. Merkle offers a similar capability where “every client has their own unique instance of M1,” Lee said.

“Clients increasingly want to own their entire environment,” Matts said. “We’re responding to that increasing desire to develop capabilities in-house.”

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!

1 Comment

  1. Good read. We see a similar trend, developed custom audiences for Brands that can be activated directly to any chosen platform. Will agencies bolster their service arm to fight tech partners? is this the arms race agains “digital” consultancies like Deloitte D and Accenture I ?