After months of testing, Google faces growing pressure to revamp the Privacy Sandbox

Six months after Google began phasing out third-party cookies in Chrome, the verdict on its alternatives is unchanged: they’re a financial wrecking ball for publishers and another lever for Google’s advertising dominance. The difference? There’s now even more proof to back it up.

Ad execs might be tempted to throw in the towel. Yet, many aren’t giving up. They’ve invested too much time, money, and effort in building tech for Google’s alternatives to let cynicism turn into apathy. Besides, it seems Google is at least extending them an olive branch — or rather a twig.

Consider this: Google hired an ad tech veteran to help with partnerships around its Privacy Sandbox the same month it delayed killing cookies for the third time in four years. If that doesn’t sway skeptics, Google is also working with the industry to establish a clearer timeline for the cookie phase-out. And, last but by no means least, ad tech execs are actually getting time with Google engineers — actual engineers — to talk about all things Sandbox.

“We’ve had a two-day workshop with Google engineers and we’re working with them to lock up a roadmap that everyone can understand regardless of their interest in the ecosystem,” said Todd Parsons, chief product officer at Criteo.

Even the smaller players are getting the VIP treatment these days.

“They’ve [Google] has done more outreach in recent weeks,” said Mark McEachran, vp of product management at ad tech vendor Yieldmo. “I even talked to three or four of them at the TechLab summit two weeks ago.”

And it’s not just face time that’s improved; several ad tech execs report that the Sandbox team is now responding to emails in days instead of weeks. 

Hopeful as this all sounds, no one in ad land is popping champagne just yet. They know none of this necessarily means anything. Not when it’s still a tightrope walk to satisfy themselves, Google, and regulators alike. At least now, though, they have a lot more evidence to say the quiet part out loud: the Sandbox, in its current form, just isn’t cutting it.

The most notable piece of this evidence comes from Criteo — Google’s biggest ad tech ally in the protracted rollout of its alternatives to third-party cookies.

After eight weeks of testing from March 18th to May 12th, the ad tech vendor concluded the Sandbox would do more harm than good. Publishers would lose 60% of their ad revenue, and worse still, they’re losing it to Google. As a result, the Sandbox would skyrocket Google’s market share from 24% to a staggering 83%. 

In short, the Sandbox would make publishers — and the entire industry — more dependent on Google than ever before.

That’s if it ever properly works. Criteo’s tests also revealed that the Sandbox would slow down publisher sites by more than 100%, resulting in lost impressions, lost revenue, non-viewable ads, and a terrible consumer experience. 

No wonder publishers think the Sandbox is quicksand for everyone but Google. Adoption remains below 55%, according to Criteo, and that’s unlikely to change anytime soon. They want solutions, not sand traps.

“We’ve chosen not to test the Sandbox at this stage,” said a publisher exec, who was not authorized to speak to Digiday. “Most of our traffic is Safari-based, and the tools for testing available in Google Ads Manager are far from adequate to do a proper test, so — in our opinion — testing would be a waste of time.”

If all this all sounds a tad over the top — it’s not. These viewpoints have solidified over the past six months, and even more so in the last three months, as the Sandbox entered its most critical testing phase yet. 

Just look at Criteo. Its tests covered the lion’s share of its 18,000 advertisers and 1,200 publishers, crunching over 100 million weekly ad impressions. That’s a sample size too significant to brush off. And it’s even tougher to dismiss when looked at alongside the IAB Tech Lab’s own analysis.

Here’s the short version: in its current form, the Privacy Sandbox may stifle the industry’s ability to deliver relevant, effective advertising, placing smaller media companies and brands at a significant competitive disadvantage. The stringent requirements could throttle their ability to compete, ultimately impacting the industry’s growth.

“When the IAB Tech Lab originally came out with its report Google tried to dismiss it, claiming that it contained ‘many misunderstandings and inaccuracies’,” said James Rosewell, founder of the Movement for an Open Web (MOW) — a coalition of anonymous businesses and industry players. “Here we are four months later and — following a thorough reassessment by the Tech Lab in collaboration with Google – we find out that very little has changed.”

In an emailed statement, a Google spokesman pushed back on these claims: “While we appreciate the IAB Tech Lab for inviting Chrome to join task force discussions, we stand by our initial assessment of the report as not much has changed. This is not surprising given the bar for the report was for everything to work exactly as it does today, which was never the intention of Privacy Sandbox. Looking forward, we’re encouraged that the Tech Lab is turning its Privacy Sandbox focus toward discussing new capabilities and sharing integration guidance with the industry, with input from Chrome. The Chrome team remains committed to its goal of collaborating with the industry, including many IAB and Tech Lab members, who are working to create innovative solutions using the building blocks provided by Privacy Sandbox.”

Nevertheless, the industry expects more from Google.

Execs at Audigent echo this frustration. The ad tech company has developed a Component Buyer — the Sandbox equivalent of a demand-side platform — to grasp Google’s vision of advertising without third-party cookies. Yet, they’ve hit a wall.

“Right now the answer is nobody knows,” said Drew Stein, founder and CEO at Audigent.

That’s a bleak status update from a company that’s invested nine months, millions of dollars, and countless engineering hours to future-proof their clients’ media buying on Chrome.

“What we found in the process were massive gaps in the capabilities in Privacy Sandbox,” said Stein. “Even in PAAPI, one of the areas of Privacy Sandbox that holds the most promise, the known gaps are still too large to be scalable for the industry. Google has yet to announce the changes needed in order to make Privacy Sandbox a reality and has yet to share a timeline for engaging with the industry on these topics.” 

But it would be wrong to say frustrations like this mean ad execs lose faith in the Sandbox completely. If anything, they’ve clarified why it can’t continue in its current guise. Sure, its design has some gaping holes, and its premise is questionable at best, but these are issues that can be fixed — assuming Google chooses to address them within legal limits. And let’s be honest, the industry isn’t exactly lacking in suggestions for Google’s engineers.

These recommendations tend to fall into one of four categories:

  • Performance features: Enhancing machine learning performance to drive more spend on the open web and higher CPMs for publishers.
  • Audience qualification: Allowing for better qualified, more valuable audiences, translating into higher revenues for publishers.
  • Critical functionalities: Providing essential capabilities to ensure transparency, avoid fraud, and maximize sustained improvement and competition.
  • Governance optimizations: Improving decision-making, accountability, and efficiency to boost publisher performance.

Or to put it another way, there’s a blueprint for fixing the Sandbox is there — if Google is willing to listen.

“None of the results from our test or the subsequent recommendation mean the Sandbox is a failure,” said Criteo’s Parsons. “Ultimately, the Privacy Sandbox is a product that is evolving, but it requires different configuration changes to meet the goals of it being shipped — It’s as simple as that.”

It’s a view echoed by those at programmatic specialists MiQ, whose ongoing tests of the Attribution Reporting API (ARA) in the Chrome Privacy Sandbox reveal a mixed verdict: promising advancements tempered by significant concerns.

Contrast to received wisdom, this part of the Sandbox does solve for privacy-first optimization, according to the MiQ’s chief strategy officer John Goulding, who is tracking the performance across several clients involved in the test. Based on what he has seen so far, the API captures 84.9% of the same unique converters as cookies — a particularly positive outcome when compared to the typical match rates seen with first-party data. However, the attribution API  does not provide a complete measurement dataset, with data loss occurring between conversion events and report generation, and it is present across only 25% of total ad impressions. 

This means marketers will need to model data to achieve accurate ROI representation.

Moreover, the attribution API changes campaign optimization best practices, as it requires reporting decisions to be made upfront. Marketers and operational teams must make calculated tradeoffs between speed, accuracy, and detail across various settings.

Despite these challenges, the attribution API is not as flawed as recent industry opinions suggested — far from it, said Goulding.

That’s the thing about the narrative around the Sandbox — and by extension the end of third-party cookies in Chrome. It swings between extremes when it really needs a nuanced throughline. It’s not an ideal scenario, of course, but it’s far from a doomsday one. Remember that 60% drop in publisher revenue Criteo flagged? That only happens in a world without any alternatives to third-party cookies or the Sandbox.

Fortunately for publishers, that world doesn’t exist. There’s a cacophony of alternatives — probabilistic, authenticated, and contextual — that are trying to prove they can make ad inventory addressable, measurable, and, crucially, maintain CPMs.

Where Google’s own alternatives fit within that remains to be seen. But in the meantime, it continues to mount a robust defense of them.

In an emailed statement, a Google spokesman said: ““We are encouraged to see companies building with the Privacy Sandbox and other privacy enhancing technologies. But it’s not possible to predict publisher performance based on effectiveness of a single buying platform, as publishers typically work with dozens of demand sources. In addition, we expect performance numbers to evolve, and they currently don’t reflect how the overall ecosystem will perform in a true marketplace – which won’t exist until adoption expands alongside third-party cookieless traffic. We look forward to the ecosystem continuing to share valuable insights and feature requests for Chrome and the industry.”

Editor’s note: this story was updated on July 1 to add comments from Google.

https://digiday.com/?p=549233

More in Marketing

Agency group Common Interest is on an acquisition drive

Agency group’s CEO outlines plan to redefine brand building with cultural relevance in a performance-driven world.

How fast-food chain Carl’s Jr. is pushing CTV to become a performance channel

For the last year, fast-food restaurant Carl’s Jr. has been spending more on the channel, pushing for it more for performance than anything else.

Marketing Briefing: Marketers race to use generative AI tools, find ways to make humans ‘smarter, faster, better’

Some already seem to believe. Marketers including the likes of Klarna, U.S. Bank, and, most recently, Toys ‘R’ US have been touting the use of generative AI and the ability to speed up processes as something that’s already helping them.