This promoted content is produced by a publishing partner of Open Mic. A paid-for membership product for partners of The Drum to self-publish their news, opinions and insights on thedrum.com - Find out more
How to target your audience without third-party cookie data
August 10, 2022
By John Tigg, GM International, Yieldmo
The non-addressable audience expands with the introduction of every new privacy law, iOS version change, and browser update designed to squash user tracking. Despite the latest delay, when Google finally pulls the plug on third-party cookies, audience addressability on the open web will plunge from around half to just 20%. Not only will this further limit the audiences that advertisers have become reliant on targeting, but it also means that prices for this shrinking pool of addressable inventory will skyrocket, leaving them with little option but to look elsewhere.
This is where the remaining 80% becomes a key focus for the future. However, it will require new and effective privacy-compliant solutions — not to mention a different mindset — to address the non-addressable.
Recognizing hard truths
As an industry, we must accept that these changes are on the way, and both the buy-side and sell-side need to embrace a new approach sooner rather than later. Perhaps first, we need to be honest with ourselves that we are dealing with the consequences of our own actions. Third-party cookies are being removed because the tracking has become too invasive, and people have begun to be not only irritated by digital ads but concerned about them. In short, digital ads have become creepy; whatever solutions are being developed to fill the cookie-shaped hole in the ecosystem, they must prioritise user privacy and the industry must respect this.
Thankfully, technology has evolved incredibly since the golden age of the cookie. Advances in artificial intelligence (AI) — machine learning in particular — have enabled the rapid processing of trillions of data points and for signals to be identified and acted upon in real-time. With the right models, we can take what we need to know about audiences without prying into what we don’t.
There is a suite of identity resolution and addressability solutions available today that utilize machine learning to provide the insights the programmatic marketplace needs to function efficiently, without treading on the toes of regulators and privacy-conscious consumers. Let’s look at one of the ways we can address and, most importantly, monetise today’s so-called ‘non-addressable’ audience.
The non-addressable audience is not so hard to address
There is plenty we can learn about the non-addressable audience without using any personally identifiable information if we turn our attention to behavior. Audiences today do not want to tell advertisers, or any service they use, more than they need to know. However, these same audiences still tell us a lot through their actions.
By monitoring the swipes, tilts, taps, scrolls, location, and timings of potential customers, it is possible to make a rapid series of highly educated predictions as to which customers are the most engaged (and therefore likely to make a purchase) versus those who are simply window-shopping to pass time on their way to work.
As internet users move through virtual spaces where ads are displayed — whether browsing, using apps, or watching TV — it is possible to identify, through their interactions, how much attention they are paying and optimize their ad experience to their attention level. We don’t need to know the details about who people are, only what they do.
Attention is, after all, one of the most valuable metrics in advertising, coveted long before computers were invented. If ad sellers can see which ad inventory is receiving the most attention, then that inventory is more accurately valued, meaning advertisers waste less spend on inventory that no one is looking at.
Right place, right time
This context-based targeting also opens opportunities to refine ad creative and how it is served. For example, if someone has a video stream (which they only ever listen to and not watch) open in the background of their work laptop, the only ads worth serving to them are audio. Likewise, if someone is watching videos on mute, the ads should be optimized for visual appeal.
Previously, the biggest barrier to identifying these context clues was that audiences must be monitored in real-time for the most valuable inventory to be immediately highlighted. Today, this barrier has been removed by machine learning, allowing otherwise incomprehensible volumes of real-time behavior data to be filtered, categorized, and analysed in seconds to show every potential customer the right ad at the right time — which sounds a lot like addressability.
By understanding and responding to audience behavior, non-addressable inventory — which is currently underutilized and undervalued — can become a new frontier for advertisers to lift their traffic and conversions and enhance programmatic inventory with a more dynamic, real-time perspective.
Contextual targeting is just one of the emerging addressability solutions that use the power of machine learning to unveil the once-hidden insights of the non-addressable audience. This new generation of privacy-compliant audience insight and smart inventory delivers on what both consumers and advertisers want while only using the data that is needed.