The “fairytale” of cross-device identification: Why third-party data isn’t enough

Now wake up. Look around. Is that how it really works? We didn’t think so.

But this enticing fairytale is being shared profusely amongst both marketers and publishers, propagating the notion that many industry players have what they call “people-based marketing” figured out. At the core of this misconception is a simple myth: that third-party data is comprehensive and representative of real people.

Multiple channels, multiple complications

Media consumption behavior has changed. Today, 25 percent of people switch between three or more devices daily. The ability to deliver relevant, targeted ads to them across those touch points just got three or more times harder (Source: GFK). Another wrench tossed into the works? A sizable 40 percent of people begin browsing on one device only to finish their transaction on another, making purchase attribution and campaign performance a nightmare to measure (Source: GFK).

In light of this, evaluating a publisher has become more than determining the quality of the inventory. Advertisers now want to know what data sources their partners have access to, and it’s the quality of that data and the ability to connect the dots across all inventory and make the right ad decision that affects their final judgment. Because of this, publishers now face the challenge of piecing their inventory together, applying data from third parties where they can, and matching it as best they can to their advertisers’ targeting needs at a very high cost.

If you give a user a cookie…

But this “accuracy” is only a mirage. It’s easy for data companies to patch together what they call a “man,” “woman,” or some other desired segment based on nothing more than inferred information and sell it to the highest bidder. When advertisers double-check this targeting via measurement companies, their campaigns fall short, largely due to the modern approach (or lack thereof) to cross-screen attribution.

Today, most identity matching relies on cookies, despite research that shows cookie-based campaigns to be only 60 percent accurate (Nielsen). Identifying a person across multiple devices and attributing in-store purchases to online brand campaigns is just not possible for the marketers who default to this method.

Why? Third-party data partners use cookies to capture information about a “person” from a single encounter rather than assessing their collective behavior over time. They base their identification on inferred attributes (from age to gender and beyond) rather than declared registration or profile data. Worst of all, these assumptions are made based on a small data set. Scale is key here: If the sample is too small, using it to target a larger audience of online users is inaccurate and unreliable. With an error rate of 40 percent (Nielsen), a publisher cannot afford to lose revenue or pass on the cost to advertisers.

From proxies to people

Accurate targeting isn’t just a way to sate a buyer’s appetite for a particular audience: It also has a huge effect on user experience. No publisher wants to show a completely irrelevant ad to their viewers or show an ad that the viewer has already seen or that’s for a product they’ve already purchased offline. Inaccuracy disrupts the user experience that publishers take so much care to craft. Not to mention, it opens up the publisher to the possibility of getting stiffed for an ad because it didn’t hit the promised target audience.

For campaigns to reach unique users across devices, the technology platform facilitating them has to take a true people-based approach by combining three things: authentic information about real people (accuracy), insights about their purchase path over time (persistence), and a clear understanding of a large portion of people who are active online today (scale). This enables publishers to make the most informed decision for their business, their advertisers and their viewers. Only by making these principles the bedrock of their targeting can they steer marketers away from proxies (like first- or third-party cookies that don’t extend across devices and browsers) and back to real people.