Identity graphs, and the companies that partner around them, are going to shape the next major innovations in digital marketing. And with any luck, they’ll prop up the depth and breadth of quality content available through every digital platform.
Why identity graphs are so compelling
We all know why we can’t rely on cookies. They’re too obscure, can’t reliably be linked back to an individual, and don’t span devices. Identity graphs (aka device graphs) sprung up as a way to try to tie individuals to the widely proliferating range of computers, tablets, phones, and other interactive screens in our lives. But they don’t go far enough to be truly stable, real-time identifiers.
A truly sophisticated identity graph is a constantly updated profile that includes both attributes that rarely change (like demographic data) along with more fluid facts and inclinations like purchasing intent and content preferences. It blends together identified and anonymous, online and offline, deterministic and probabilistic, and individual and household-level characteristics. And to be valuable, an identity graph must be readily accessible and updated as facts change – not stitched together in daily or weekly batches.
Trapped in Wanamaker’s world
With identity graphs, publishers and brands can get out of the business of pushing irrelevant ads and embrace people-based marketing. The current, dominant mode of digital advertising isn’t serving anyone well, apart from the two companies collecting most of the revenue. The digital advertising world we’re in today constantly reminds us of the old John Wanamaker observation that says half of the money spent on advertising is wasted – but it’s hard to know which half.
We should be so lucky as to live in Wanamaker’s world where only half of an ad budget is wasted. Between click fraud, closed media ecosystems that don’t provide sufficient performance information, and throwing irrelevant messages at consumers who don’t want to hear them, many advertisers are wasting far more than half of their budgets. In fact, ad fraud was reported to cost marketers $16.4 billion globally in 2017.
And that problem only gets worse as digital advertising is increasingly dominated by the walled gardens of Facebook and Google. Five years ago they collected 35% of digital ad spending. In 2017, they collected more than 73% in the United States, and are well on their way to increasing in 2018.
That’s bad news for marketers, who get little to no information back about the individuals they’re spending ad dollars to reach. And it’s bad news for publishers, who have no real option but to keep cutting costs as they lose digital ad spending to the walled gardens. What’s more, it’s bad news for consumers, who have fewer choices and lower-quality content.
Are identity graphs a better solution for everyone?
Identity graphs will give marketers what they want: a more accurate picture of every potential customer they want to reach. They’ll also benefit publishers, letting them command a higher price from advertisers to make well-qualified matches with their readers. And, perhaps most importantly, they can give power back to the consumer.
Will we see legislation similar to that which already governs promotional email and text messaging? The type of legislation that allows consumers at any point in time to say they don’t want to be approached anymore. That would be almost impossible to enforce without a stable identity graph.
Marketers can embrace identity graphs by embracing the core lesson of classic direct marketing: Put your message in front of people who fit your profile and want to hear from you.
And you don’t need to wait for the technology to mature before you start exploring. Today’s identity graphs are already solving tangible, bottom-line marketing challenges, like being able to distinguish the likelihood to buy between two prospects with identical cookies. They’re a way to connect on an individual level with the customers who want to hear from you, and a means of breaking out of the walled gardens that take your ad dollars but offer little data in return.