Thoughts on Turtledove
Things are getting exciting. We are holding our breath to see what solution will eventually replace 3rd party cookies. This is almost as exciting when waiting for the final GDPR docs; we all know it is going to hurt our industry but how, how much, and exactly when. Apple has paved the way with ITP and IDFA, now Google is also making big changes to the world’s most popular browser Chrome and its Privacy Sandbox. For advertisers and traditional data vendors, this nightmare is called Turtledove.
What is Turtledove?
Turtledove will basically be an API (application programming interface) which tries to enable behavioural targeting while limiting the amount of user data that is shared with advertisers. What Turtledove is proposing is to completely re-architect the current system so that user behaviors are not stored in a remote database, they are stored in your browser and they never leave your browser. In other words, Turtledove brings the auction and final decisioning out of the cloud and into the browser. Kind of funny though, since Google — that company that lamented the terrible things the header was doing to latency while proclaiming the brilliance of server connections — now decided to leverage edge computing.
There is no official doom day date yet but it will be released sometime within the next 24 months. Note that this applies to Chrome, it is too early to say if this move has any affect on Safari and the rest.
User benefits
Questions also remain about how users would really benefit from Turtledove. Well, you could argue that it:
Eliminates third-party cookies and all cross-site tracking.
Reduces browser signals dramatically making “inferred IDs” (aka, device fingerprinting, cache inspection, link decoration etc. ) unreliable and unusable.
Enables sufficiently large user groups ascertaining anonymity of any one person/device within that group, but sufficiently small to enable relatively fine-grained reach.
Utilizes these user groups instead of device-based or cookie-based identifiers.
Potential pitfalls
At this stage, some of the pitfalls are:
Turtledove would disable many common web advertiser tactics such as frequency capping, A/B testing, attribution and reporting. For these to work, Turtledove relies on the Aggregate Reporting API. Yet another early sandbox proposal to be defined.
Turtledove could risk handing Google yet another advantage over its ad tech competitors. Only this time they are being more tactical in their approach to their antitrust practices.
An average citizen is still going to feel like being chased.
The browser turns from a relatively passive piece into an active player in the bidding. This would strain the bandwidth and computational power consumed by ad tech, particularly on mobile devices. Servers can process huge amounts of data, but data on a phone is limited by Wi-Fi, cell service and available device storage.
Complex auction logic does not seem possible. Even not-so-complex segmentation, such as merging two interest groups to create a more nuanced audience segment.
What about Safari users today, ITP and alternative solutions?
Basically this boils down to ITP (Safari, Firefox), which in turn means no cookies, my friends. And because the adtech world is built on 3rd party audiences, ITP poses an impossible challenge for marketers. Only contextual targeting seems to provide the working solution. Which in turn, can be considered a rather one-sided approach if we merely target based on the page content. But it has proven to be effective, so it is a viable solution that is able to address all browsers.
On top of old-school contextual targeting, advances in machine-learning and AI have opened up a whole new world of opportunities for intelligent targeting. Personally I am very excited about how well a system can profile a user without any PII and prior information with just a few signals at the time of the data request. It feels like magic but is actually based on years of AI and realtime scalability tuning. When compared to 3rd party audiences, this is revolutionary. Realtime moments can be seen as more relevant and efficient ways to profile and target. And the best part — it is a sustainable model and it is here now.
Final words
Targeting without personal data is set to become the new normal, and to me this now seems entirely appropriate. Advertising theories and methods have evolved continually over the years and I see this shift in ad tech, seismic as it is for many, as simply the latest chapter for this smart and agile industry. Moving away from cookies is going to be uncomfortable for many and unanswered questions do still remain. It will require loads of brainstorming, and brave minds to have the courage to try out new approaches in the new world of adtech.
The relief is that there already are future-proof solutions that indeed can replace and sometimes even more efficiently, the old world targeting cookie-based models. So heads up, this is going to be a rewarding ride!
Rami Alanko
Beemray
