Google Cookie Apocalypse Now Redux

It’s not often I get three people reaching out to me the same day to get my POV on a webinar but that happened last week after Ari Paparo’s quite informative “Cookie Apocalypse” briefing on the Google Chrome announcement.

Why were people reaching out to me? The timing of Google’s announcement killing third-party data is a week ahead of the 10-year anniversary of my founding Yieldbot. I started Yieldbot to take on the third-party cookie economy and be privacy centric adtech using first-party data. Unfortunately for Yieldbot’s stock-holders the cookie apocalypse happened 2.5 years too late for my adtech righteousness.

As I noted in this launch interview from March, 2010:

Screen Shot 2020-02-11 at 12.19.47 PM.pngHere is what I blogged about back in 2010:

My Thoughts on Online Privacy

Links vs. Cookies” A Tale of Two Economies

2010 was actually a watershed year in digital privacy. Congress was investigating Facebook for sending IDs via RapLeaf into the nascent VC fueled open-web/programmatic ecosystem that was also given its first Lumascape in 2010.

Since Yieldbot I continued down the first-party data path and it led me full-circle to brands and their martech and the rise of Customer Data Platforms to manage identity. Third-party cookies were just a hack for identity. Brands are holding records. In 2020, same as 2010, I hold the belief that the ID future is something that should be owned, controlled, managed and complied by brands directly as a first-party with their customers.

I’ve also been fortunate to work with a very large data collector in the EU on go-to-market strategy in a GDPR world. In fact, with all the changes in cookies and regulations it occurred to me, after being “3 years behind” for so long the EU in one fell swoop of regulation and cookie changes has moved ahead of U.S. in adtech. Good for them!

So with all that context, I listened to Ari help the industry dissect what is going on with Google and the adtech ecosystem at large. Here are my thoughts for my friends who asked me (and the one who told me to make it into a blog post) and anyone else.

Cookie Apocalypse: Briefing with Ari Paparo, Feb 5, 2020 (US)

Screen Shot What We Know.png

I will start by emphasizing Ari’s first point that the industry stood idle and in denial while ITP (Intelligent Tracking Protection) was first introduced by Webkit (Safari) 2.5 years ago and obscured THE MOST VALUABLE WEB CONSUMERS. If we really want to go back in time Firefox (when people actually used Firefox) starting blocking third-party cookies 7 years ago. This should tell you all you need to know about the industry’s preparedness and ability to take this problem on now.

The key part of Google’s announcement is the idea of a Google Sandbox. This is in my opinion a shorthand/non-explicit way of saying Google Cloud Platform. Sandbox = database to bring different datasets and APIs together. In that context there are a couple important ideas worth expounding on.

First, in a previous post on Google Tag Manager , I talked about the battle for Google’s future market share being in cloud-infra and the gateway drug to GCP for brands being AdWords & DV 360. The sandbox idea fits neatly into that thesis.

Google Cloud Platform also sets the stage for a better Google Attribution offering since I can see Brands wanting to join their CRM and Customer Data Records with Google Behavioral Data in the “sandbox.” Let’s recall Google bought Adometry during the “Spring 2014 season of attribution vendor acquisitions.” (AOL acquired Convertro for $100M a few weeks later) Of course, Adometry (founded interestingly as a fraud-protection service) did not help Google win trust in attribution. Trust being the key word. But a sandbox with real data visibility can do this.

With brands owned data being joined in a GCP “sandbox” Google could present an attribution product that would not only be trusted, it would be untouched in accuracy. Further, this data asset when joined with Google data on a “first party-to-first-party” basis would drive many billions more in marketing and ad spend in Google’s direction because let’s face it, the market is less than perfect (though some may argue that is why Google prints money).

One other point of view on “View through and multitouch would not work at all…” They could work on the brand side. Brands have the ability to instrument through modern first-party customer data tools not only multi-touch attribution but cross-channel attribution as well. Products like Snowplow and Segment and to varying degrees of technical prowess the CDPs  can provide a lens on media attribution and help create ML based models for media mix.

Screen Shot Don't.png

Agree with Ari that MAIDs will follow path of “privacy” and that will be another even more titanic shake up. MAIDS are now the currency of choice in user-level tracking. It’s actually smart market strategy for Google and Apple to begin with third-party data as arguably this is a softer bomb drop on the industry that will have no option but to eventually accept surrender.

Masking UA (user-agent) data like IP address is already in working groups. While I can get my head around MAIDs obfuscating the UA string is something that on the surface seems incompressible to someone like myself who is still guided by the Cluetrain Manifesto and the Declaration of Independence of Cyberspace.

I’ll add one thing on cross-domain first-party cookies. This can be architected in compliance with current ITP standards already and I don’t think that is going to change with Google SameSite as long as the cross-site data collection and processing sits across a single domain owner. Cross-domain can be instrumented with a 2-year persistent cookie window. This is an exact project I recently completed for a multi-brand client and it takes the same form with any network of sites be that e-commerce or publisher.

The cross-site challenge I see is that this type of data collection and processing must be built. The minute you have a vendor involved there is a data mediator and I don’t believe that would be compliant. It’s possible you could stream the data from a vendor tag in the CNAME to a multi-brand owner endpoint but vendors typically are not built for that “hack.”

Screen Shot context.png

Contextual – Advertisers think in audiences. Been that way since day 1. Will not end.

Publishers – Publishers don’t have scale with segments. I saw this firsthand at Yieldbot. Not to mention their audiences and page-views are shrinking as attention turns elsewhere and zero-click searches grow. Advertisers want control over audience creation (that is their value) Ari mentioned the issue around self-declaration. But does the industry really need new middlemen? I think not. Bottom line with audience extension being dead publisher segments are dead too.

Hashing PII – In my opinion phone numbers or email hashes are not going to work in a stricter GDPR centric world. There has been research to support this opinion already.

One note worth mentioning here. People are acting like data sharing is the purview of the agency or buyer. What is clear with GDPR is that it is foremost a Brand responsibility. Importantly, many brands for their own protection/consumer trust will want to go “above and beyond” GDPR measures. I haven’t heard anyone in adtech speak about going above and beyond. Adtech is acting as if it wants to push to the limits of privacy compliance showing how distant the industry is from their customers. Hashing is an example of that distance.

Panels – These are small groups of people. See the scale issue from earlier. Ari mentioned this “back to the future” but most of these panel companies have been pushing into third-party data hard for the last decade. Not really sure what they were thinking, or why they had their eyes closed to ComScore settling multiple class-actions for privacy violations but I’m not long on panels and see a world of hurt coming for measurement companies of all stripes.

ML – Platform side this is already going on. Brand side it will but has not yet. There are huge unlocks that await ML centric-businesses like the attribution/mix model example from earlier. As brands move into cloud-based infrastructure (a theme of mine as you can tell) this capability will continue to improve business results. As tools, available datasets and modelers grow, differentiated success will be measured only by talent.


I’ll start with a comment – this is all about identity and we should call it identity in all conversation rather than more innocuous “cookies.”

If you’ve received this message like I did last week you can get a sense for where identity is going. Chrome is ramping up log-in and ID management. They are making it frictionless and creating value. They will own consumer data from both ends, consent and collection.

Screen Shot 2020-02-07 at 1.21.19 PM.png

Ari’s answer to the companies that are helping solve identity was publisher & adtech based. Identity does not scale on the publisher side and worse, its value in nominal. It paints a minuscule picture of who I am as a consumer. If it didn’t publishers would have owned Display & Video and all adtech would be SaaS. It’s not. Subscriptions are a non-advertising channel so nothing to see there for pubs and monetizing their data.

I think the Identity Apocalypse is more “pro-platform” than Ari believes it to be especially if Ari agrees on the brand data advantage. Identity in a first-party world as I mentioned earlier is the brand’s to own and keep. As a consumer I trust Nike to know me. I want Nike to know me. Nike shares data with the platforms already. We know they share customers. The platforms for all intents and purposes are “the cloud.”

Speaking of the cloud, Amazon is in a great position since like Google they are basically a utility for people. Wisely their supply-side business has been building server-to-server connections for the past five years. This creates an infrastructure for identity and transactions. While I agree with Ari it will short-term hurt their DSP I see a new channel for Amazon emerging that just might help publishers more than Amazon has done already. Also most publishers highly value their Amazon relationship and would welcome that non-threatening growth.

There was a question during the Q&A about agencies and brands pushing more dollars to the platforms. Ari said he hasn’t seen that happen yet but before these headwinds to the open web it was already happening based on efficiency of consolidating budgets and performance. Astute advertising nerds may have already noticed retargeting dollars are drying up which is arguably the only channel that cookies were able to deliver measurable results with. The re-targeting chickens have come home to roost.

As Ari notes the end of re-targeting may have a negative effect on Facebook short term but I think the real culprit to their DTC channel which has headwinds of its own before this announcement will be the end of VC investment for marketing. This pain will also be felt in the open-web ecosystem. Affiliate networks too will be crippled but those rev shares dollars will go right into the platforms. In fact this paves the way for a platform affiliate business, something Amazon knows a thing or two about.

Ari mentioned Live Ramp’s IdentityLink and the LR business has been doing very well in the face of all these odds. 2019 will be a pivotal year for them. They are a true wildcard in this ecosystem, tied into so many parts of it that detangling could be messy and costly. However, LiveRamp has historically been deeply integrated into many parts of Google for analytics, bidding and transaction data and has true omni-channel data. I would not rule out a path for them as the identity bridge for the open-web which, is basically what they have been all along.

I’ll end like Ari and say that “Apocalypse” is the wrong word. And like Ari I am actually very optimistic. I would call the this the dawn of a new era. A more producty name would be the “privacy by design” era. Legislation, technical limitations, market forces are all building around data sharing controls and limits.

This is long overdue. It was eight years ago that President Obama unveiled a Consumer Privacy Bill of Rights actually called “A Framework for Protecting Privacy and Promoting Innovation in the Global Digital Economy.” While I was excited about it, I may have been the only one. Nothing happened. Then after years of consumer awareness of being re-targeted we had Facebook/Cambridge Analytica. And now everything is happening. At once. So it is a horror to the adtech industry but I would suggest they listen to the immortal words of Colonel Kurtz…






2 responses to “Google Cookie Apocalypse Now Redux”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Create a website or blog at

%d bloggers like this: