Tired: Companies investing in Chrome Privacy Sandbox testing.
Wired: Companies trying to justify the time and money they spent testing the Sandbox APIs now that Google isn’t deprecating third-party cookies in Chrome – or even releasing a new cookie consent mechanism for its browser.
Google says it’s going ahead with the Privacy Sandbox, but the program’s future feels in doubt while cookies are still in play.
What are the implications for publishers and ad tech companies that pitched in on Privacy Sandbox testing?
A few lamented their seemingly wasted time and effort. But everyone who spoke with AdExchanger said they remain committed to post-cookie targeting and measurement – even if Google eventually abandons the Sandbox entirely.
A massive commitment to research
Publishers and tech companies were candid about the resources they’ve spent supporting the Chrome Privacy Sandbox’s development since it was announced in 2019.
Demand-side platform Adform devoted “north of 1,000 days of pure development effort” from its engineers since 2019, according to CTO Jochen Schlosser.
Adform has around 230 developers, and roughly 50 to 70 engineers worked part-time on Sandbox testing at various points. That many engineers spending that many hours on testing was a massive commitment for a research project with “zero commercial objective,” Schlosser said.
Although Adform received a Sandbox testing grant from Google, the grant required a matching commitment. According to Schlosser, Adform spent “more than a few hundred thousand” to participate in the grant program.
After Google turned off cookies for 1% of Chrome traffic to support Sandbox testing at the beginning of last year, Adform’s advertisers committed between $30,000 and $100,000 apiece to testing, Schlosser said.
Adform was upfront with clients that they might not see a satisfactory return. “We needed to set expectations that we can’t promise the same cost per click, the same CPMs, so they might be disappointed by the results,” Schlosser said. “And, honestly, some clients were disappointed.”
And, honestly, so was Adform.
Although Sandbox testing “was an exciting journey that drove discussions around trusted execution platforms and encryption,” Schlosser said, “as a research project, from a software technology perspective, it was a waste of time.”
Laying the groundwork for addressability
Others were less down on their experience, like c which developed bidding infrastructure for the Privacy Sandbox.
“It was a chance to modernize our tech stack, so there were ancillary benefits,” said Andrew Pascoe, VP of data science engineering.
Still, NextRoll had to devote a fair amount of engineering resources to the task. Starting in 2023, it reoriented roughly two-thirds of its DSP team to work on Sandbox testing.
But Pascoe doesn’t have regrets. For example, he said, “we’ve found out which ad tech players are actually invested in making the web more private.” And testing has led to some lasting business relationships for NextRoll, he added.
“I’ve seen the takes on LinkedIn: ‘This was a giant waste of time.’ That’s an extremely cynical view,” he said. “I absolutely do not think it was a waste of time, at all.”
Plus, Pascoe said, although the Privacy Sandbox wasn’t ready for prime time, other companies could improve on the open-source APIs. Besides, he said, the industry needs addressability solutions even more now than in 2019 when the Sandbox project kicked off.
Criteo, another leading Privacy Sandbox tester, shared similar sentiments during the company’s Q1 earnings call last week.
According to Chief Product Officer Todd Parsons, investing in Sandbox testing helped Criteo improve its platform’s audience setups, product recommendations, bidding and optimization – and that work will continue.
But, he added, “we will not have to continue to invest in Privacy Sandbox APIs.”
Curation and contextual targeting’s gain
Some testers came to that conclusion even sooner, like supply-side platform PubMatic, which spun up a team of 30 to 40 engineers that solely worked on Protected Audiences testing for a period of six to nine months last year, said Mike Chowla, VP of product management.
At the height of testing, PubMatic devoted, at most, 10% of its engineering resources to the project.
However, “as testing progressed and the feedback came in,” PubMatic scaled back this work significantly due to “anemic” demand for Protected Audiences, he said.
Overall, PubMatic didn’t see “a ton of direct benefits from Privacy Sandbox,” Chowla said.
“But we have seen real benefits from the broader industry push for addressability alternatives to the cookie,” he noted, pointing to the push for cookieless targeting, which has helped boost PubMatic’s curation and contextual offerings.
Frankly, Chowla is just relieved that publishers won’t have to deal with the deprecation of third-party cookies in Chrome.
Publishers didn’t generate much revenue from the Sandbox tests, but at least they only affected 1% of Chrome traffic, he said. “The revenue hit from when Safari and Firefox dropped cookies was much greater, and those impacts carry through today.”
The publisher POV
Speaking of the effects on publishers, Google’s pivot doesn’t take the sell side back to where it started in 2019, because third-party cookies have become even less available on Chrome since then, said Amanda Martin, CRO at Mediavine.
“A lot more users are opting out, even if it’s not prompted uniformly across the browser,” she said. Google’s about-face might perpetuate the cookie status quo in the short term, Martin added, but “I don’t think any publisher or advertiser is planning a strategy around cookies.”
Mediavine tested the Topics and Protected Audiences APIs, but it never had any engineers working on the Sandbox full time. DSPs, SSPs and data platforms did most of the development work, Martin said.
Although Mediavine discovered that contextual targeting via Topics does function well, she said, “it doesn’t drive a ton of monetization value for publishers.” Meanwhile, Protected Audiences “never really scaled,” Martin said, and she sees no reason why Google would continue to support it if the industry doesn’t.
Even so, she said, Sandbox testing wasn’t a bad thing, because it made publishers, SSPs and DSPs emphasize first-party data strategies, “which is what we’re already pacing toward.”
But, while Chrome keeping cookies is a good thing for publisher revenue, she said, it’s a bad thing for user privacy. “Information is still openly transmitted across the ecosystem,” she said, “and if publishers don’t put all that information out there, their inventory is not as valued.”
Unlike Mediavine, Raptive spun up a small pod of engineers that focused roughly 80% of its time on Sandbox development during testing sprints, said Jon King, EVP of revenue operations. These resources came from Raptive’s audience identity team, and the Sandbox work was part of its overall task to develop Raptive’s audience ID graph.
Like Mediavine, however, Raptive found the Topics API to be the most fully realized part of the Sandbox, King said. He believes it could remain a viable contextual targeting tool – assuming the buy side, particularly Google Ads and DV360, supports it.
Fact is, the continued viability of the entire Privacy Sandbox project hinges on whether Google’s buy side contributes strong demand, King said, which remains a huge question mark.
Still, despite all the drama and the open questions, King believes the Privacy Sandbox was mostly positive because it created momentum behind industry efforts around privacy and new monetization strategies.
“For good or bad,” he said, “Google is the 100,000-pound gorilla that forced the conversation.”
