This is how the cookie finally crumbles. Part II: Or “All Your Base Might Belong To Google”
Last week, I reviewed Google’s journey from the time it announced getting rid of the 3rd party cookie in early 2020 up through the 4th of January, two weeks ago when Google started its long-promised sunsetting it. This first phase began with shutting down cookies across about 1% of users, or approximately 30 million Chromers.
Over the last few years, a variety of alternatives have been proposed that might replace the cookie with something less invasive but just as valuable. No solution is specifically endorsed, but Google has gone ahead with the plan.
While the broader implication of Google's decision is a significant shift in digital advertising practices, challenging the industry to innovate in targeting and tracking while respecting user privacy, Google’s move can be interpreted much more broadly than concern with privacy or an attempt to not run afoul of privacy ordinances around the globe.
I suggested in closing that maybe Google has concluded that it can use the massive amounts of data it has collected every day for years, plus oodles of data collected every day from here to eternity, to identify audiences – separating humans from bots -- and use it all to have the most complete targeting apparatus known to advertising-kind throughout the media-verse.
To recap a few observations:
With original Search and YouTube, Google owns the world’s search market, online video market, and online music market.
Google has user data on nearly every kind of online behavior, and the starting point for that behavior is mostly search and YouTube.
DV 360 supports 30-40 % of the ad serving market, representing over 80% of display and between 15-30% of video. As more videos go programmatic, that share will grow.
Google has often been the invisible, and sometimes not so invisible hand, of privacy policy development. GDPR and CCPA were shaped in part by Google.
Google’s mobile operating system hovers just over 70% globally and 57% in the US.
Google’s Smart and connected TV operating system is officially between 10 and 15%, but unofficially over 20% and growing.
Chrome has a hair under 65% global penetration, with just over 50% in North America.
All media is moving to be connected to an IP, which means it’s connected to a machine-readable ecosystem.
The endpoints of the ecosystem are devices.
The devices all have IDs.
The devices are running Google’s OS.
Google has optics on the entire media universe, including the universe’s most extensive ID graph, that reads across every major media.
Many have fallen at Google’s ramparts trying to guess what it's up to when it decides to do something; I contend that Google is letting go of the cookie because it can.
Google is positioned to hold an enormous part of the media universe, including TV, that should be the envy of any media owner.
As TV becomes more CTV and buying gets committed on a trade desk or DSP, a lot of advertisers are going to throw up their hands and say, “fuck it, I’ll put more of my schedules here since I already do so much with Google Ads, anyway, and I’ll at least be able to get a universal view of audience and activity in one place across multiple platforms in GA360 for a pretty big chunk of my audience, and it’s just too much to think about dealing with the moving parts outside of the Google walled garden.” I’ve seen clients go 80%+ all-in on Google because of the unity of the buying platform and analytics, even if the advertiser understands intellectually that Google grades its own homework and doesn’t let you study with books from any library but its own. Companies are and will continue to attempt to provide optics on the media universe outside of Google, for example, Lotame’s Panorama ID or Trade Desk’s UID2, and these will preserve and aid optics on the open web… as much as there will still be.
Another thing to consider: after decades of collecting, cleansing, normalizing, structuring, observing, and analyzing data, Google has concluded that microspecific audience data isn’t particularly meaningful for yielding success. Deterministic person-level targeting isn’t necessary. When an advertiser wants to move the needle on their business, it needs reach. Google now has enough eyes on everything to conclude that granular targeting specificity isn’t the best way to achieve it. When it’s all about reach, reach means crowds, and crowds are more about fluid dynamics than they are about a particular collision of atoms.
Complex probabilistic modeling using the sample sizes Google has access to can’t be worse than deterministic decisions based on data signals that are 50% fraudulent. For all the criticisms leveled at Nielsen for its antiquated methods, statistical modeling based on samples is, for the most part, sound. If you don’t think so, the next time you see the doctor for a blood test, tell them they must take all of it so you can be sure of the results.
This explains the possible move to affinity collectives, e.g., TopicsAPI. Compelling as it has been for over two decades, one-to-one and personalization don’t produce results at a scale that outstrip more general segment targeting. Knowing that someone is a club-footed home-office radiologist who likes left-handed pitchers for the El Paso Diablos doesn’t get that customer to like your brand of toilet paper better or buy more of it. People aren’t looking for a deeper relationship with their mayonnaise. And if you’re McDonald’s, you want to reach people with mouths; if you are Crest, you want people with teeth in those mouths. Anything else is just a matter of potentially particularizing your creative, but not who’s interested in the product.
Where does this take us? If I knew, I’d be a rich man. Maybe Google seeks to become the dominant media channel for advertisers. Maybe Google will start trading data with the big traditional media companies for access to content. Maybe Google creates an alternative audience-based currency, competing with iSpot.tv or VideoAmp to beat Nielsen at its own game.
Whatever it is, you can be sure there’s a reconfiguration of stars and planets coming that we’ll be drawing new constellations.
And now, for your enjoyment, a throwback to one of the internet’s first memes. 11940(https://www.newgrounds.com/portal/view/11940)
