So, an Academic Walks Into a W3C Meeting…

There’s still a lot of uncertainty about what happens once the third-party cookie completely goes away.

In the fallout, publishers will experience a significant loss in revenue, but things may settle a bit if there’s a viable alternative in place. Some pubs are already prepping Universal IDs and first-party data strategies, along with resurfacing some contextual targeting solutions. But will they be enough?

Whatever comes next, it’s going to take the cooperation of the industry to ensure the open web’s survival and a level playing field for pubs to compete with the likes of the triopoly—Google, Facebook, and Amazon.

I recently chatted with Garrett Johnson (no, not related)—a marketing professor at Questrom School of Business at Boston University, who focuses on privacy, the value of cookies, and the effectiveness of advertising—about these topics, as well as a presentation he made to the W3C called Economics of Digital Ad Identity.

I was very interested in these stats he presented:

  • In a cookieless world, ad prices will be 50-70% lower.
  • Identity increases ad value 2x-3x.

And so I wanted to learn more.

Lynne d Johnson: You’ve been making the rounds with your presentation on the Economics of Digital Ad Identity. First, I saw you made one to the W3C Improving Web Advertising Business Group and then I saw you were part of a W3C lineup at WebInnovationX, SymposiumX MAdTech: Back to the Future, hosted by the W3C  NY Metro Group.

We’ve been hearing a lot about the W3C not listening to voices outside of Google Chrome, yet, your presence within the group, as an academic, shows a more inclusive organization, actually listening to other voices. How did you get involved with the W3C and what do you hope to achieve by speaking to the ad tech industry about this?

Garrett Johnson: This W3C group has diverse participants grappling with post-cookie advertising proposals, but I am one of few members outside of the industry. I was invited to join the group as an independent subject-matter expert who did not have a “horse in the race,” so to speak.

Academics are sometimes criticized for being stuck in the ivory tower. Research is an enormous privilege, so I try to work on important topics and contribute what I’ve learned to the public discussion.

I recently presented to the W3C about the value of cookies and my economic evaluation of the post-cookie alternatives. I wanted to level set the discussion because I think we make better decisions when we know the facts and have a framework for evaluating our decisions.  This was really well received by the group: one participant expressed that he wished I’d done this months ago!

I’ve since given shorter such talks at a NY chapter W3C event and at the Marketing Science Institute.

LJ: In your presentation, you reveal that in a world without cookies, the ad revenue pie will be divided more unequally than it is today. How much of a loss would this be for publishers? How do you determine the value of a cookie? And, how do you think things will play out once there is a cookie alternative in place?

GJ: Those are big questions! I start from an understanding that cross-site identity is valuable to advertisers. It’s not just behavioral targeting: advertisers value cross-site frequency and reach management, as well as cross-site measurement.

Most data points—including academic and industry studies—agree that programmatic ads generate 50-70% less revenue without cookies. Our own study put the figure at 52%. I’ve been collating these studies in a Twitter thread and Google deck:

That said, when cookies disappear, the market will adjust so as to prevent CPMs from falling so hard. Where we land will depend on the quality of the replacement. I worry that the ad revenue pie will shrink if the industry has a worse offering for advertisers. I also worry that the pie will be divided more unequally because the walled gardens will be able to preserve many of their current advantages. The walled gardens will thus be even tougher competition for the open web.

It’s worth noting that this is exactly what is happening in my industry during COVID. Colleges are offering a worse in-class product due to the pandemic. The whole industry is hurting—especially the bottom end—whereas Harvard and Stanford are fine.

LJ: You mention that the replacements or tentpoles for the cookieless world are context, authenticated traffic solutions, the walled gardens, and the Privacy Sandbox. There’s a lot of scrutiny in the industry about context and scale. What do you think of this criticism? 

GJ: The challenge with contextual targeting is that users don’t spend much time on niche content of interest to advertisers. Two studies show that general and news content are even more reliant on cookies to generate revenue. I am particularly troubled by this given how hard news publisher revenues have already been hit. When you also lose cross-site reach and frequency management, this is a tough sell to advertisers for all but the largest publishers.

Authenticated traffic solutions are promising because they provide an identity solution that can solve the cross-device ID problem. Scale and adoption are currently an issue, though these solutions seem to have a lot of momentum behind them.

One interesting tension here is how much the industry will coordinate on sharing data. I know that publishers are worried about audience leakage, though another risk here too is less total value creation without some sharing.

LJ: What about the Privacy Sandbox…are you feeling like any of these solutions give the open web a fair shot? 

GJ: The open web needs a replacement for the cookie that can be competitive. I affectionately refer to the various proposals at the W3C as “The Birds” because of their acronyms: SPARROW, PARROT, FLOC, TURTLEDOVE, etc. All these proposals are trying to preserve the advantages of the status quo for the open web.

The good news for privacy is that advertisers care about audiences, not individuals. The challenge is that guaranteeing user privacy requires difficult trade-offs. Advertisers will need to rely on user segments, though they may or may not be able to create their own. Advertisers will have to make do with delayed rather than real-time ad measurement. Requiring minimum segment sizes also poses greater challenges for both smaller publishers and smaller advertisers.

Still, I credit all involved in recognizing the serious consequences of losing cookies for the open web. This industry has rarely seen so many participants working together and working proactively to build a replacement technology.