Since the General Data Protection Regulation (GDPR) came into effect in the UK last year, an air of ‘business as usual’ seems to have prevailed among the advertising industry, with many deeply unsure how to target digital audiences without using personal data.
But in recent weeks, the Information Commissioner’s Office (ICO) has cast a spotlight on ad tech and how it handles personal information which is making ‘keep calm and carry on’ seem like a less viable option. So my advice to media buyers right now is: start preparing for a world without cookies.
In its report on June 30, the ICO raised serious concerns about compliance in the cookie-based system of real-time bidding (RTB) and it has given the industry notice to improve. If that wasn’t challenging enough, there’s the additional unfolding reality that browsers including Mozilla and Safari are no longer prepared to keep supporting tracking cookies, with Chrome widely expected to follow suit.
So even if the ICO satisfies itself that the ad tech industry is using cookies responsibly, steps are being taken by the dominant players to stop them.
Cookieless targeting – is that even possible?
The programmatic ecosystem has been fueled by personal data and cookies for more than two decades and for many newcomers to the industry, it will be hard to imagine doing business any other way.
But as the cookie crumbles, machine learning is offering a whole new realm of targeting opportunities which are compliant, cookieless and delivering high-impact campaigns at scale. What is more, they are actually providing better performance results for brands, better returns for publishers and better experiences for the end-user.
The solution has actually been right before our eyes all along, as, for years, while traditional media buyers have been chasing audiences based on historical activity or running against pre-defined contextually relevant pages, significant insights have been left on the table.
Contextual re-engineered
These audience insights derive from real-time content consumption, learned by analysing the content on the page when ads are being served, and are proven to teach brands a huge amount about what drives new audiences to want to hear from them. What is more, by fusing this contextual information with powerful AI algorithms, these insights can be used to intelligently scale campaigns to find new and relevant audiences at the optimal moment.
We call it ‘mindset targeting’; others such as the New York Times call it ‘motivation targeting’ – both are similar iterations of this type of agile, adaptive, expansive contextual technology which stands ready to replace the cookie. This style of targeting doesn’t rely on personal data; it maintains privacy, is brand safe and improves the exchange between brand and user by communicating in the right places and at the right moments in time.
The Washington Post has just announced its own version, Zeus – the latest AI-driven contextual programme from a major US publisher, whose arrival is all the more exciting in a country which doesn’t yet have severe data laws. So you could deduce that the brilliant minds at these top titles see this as the best way forward anyway, irrespective of compliance.
Content, not cookie, is king
In fact, these premium publishers stand to gain most as the pendulum swings back towards contextual signals, as algorithms such as ours often place the highest value on rich long-form content. And research* shows that quality content which keeps people on a page for longer is the natural environment for high-quality ad engagement. This is the founding principle of publisher collectives such as the Ozone Project which aims to put power back into the hands of publishers, thereby, as commercial director Danny Spears has said: “reestablishing their sovereignty”♱.
So in time, the new wave of contextual targeting technologies will hopefully address a clutch of industry issues: data compliance, browser functionality and brand safety; as well as finally driving a U-turn in revenues back to a time when content, not cookie, was king.