‘There is movement now’: Advertisers cautiously step into Google’s Chrome’s cookieless era with the Privacy Sandbox

For many performance advertisers, this is an uncomfortable new reality. Brand advertisers, on the other hand, find themselves at a crossroads. The Sandbox doesn’t just limit their reach across websites; it also hampers their ability to prevent ad fatigue among those who visit.

Yes, it’s a broad brush. But while some brand marketers are tentatively dipping their toes into the Sandbox, the reality stands stark: the Sandbox is far from being a universally perfect alternative to third-party cookies for all marketers.

This is why the current experimentation in the Sandbox feels like a patchwork effort, akin to piecing together a puzzle without the full picture in sight.

Nowhere is this piecemeal approach more evident than in the testing of the Protected Audience API — the Sandbox’s way of trying to make retargeting more privacy-friendly. While you can delve into the specifics here, the gist is that it moves ad auctions from web page to browser so marketers can advertise to people who have previously visited their site.

This is a monumental change, impacting everyone in the ad auction ecosystem, not just the advertisers. All of the intel the demand-side platform (DSP) has to inform their bid is pushed to the browser, and the browser runs an auction in a protected environment. Only the consumer’s browser can see what’s happening. Navigating such a vast change was always going to be intricate. 

Without audience data being shared between publishers, supply-side platforms (SSP) or DSPs, the DSP won’t know which consumers belong to each interest group. But the browser will know, and it will handle all the ad logic for bidding. The SSP won’t see the interest group bids either, but again, the browser will and it’ll manage the auction logic for scoring bids. All while keeping consumer info private and on the device.

In simpler terms, it’s currently a complex and fragmented scenario, yet it’s the most viable path forward.

“We need other industry platforms to get on board with these tests because this is something we can’t do ourselves due to it being such a massive change to how the auction works,” said Georgiana Haig, global strategy and partnerships director at MiQ, who hopes to begin testing the Protected Audiences API within the next three months. In fact, she and her team are currently in talks with at least one DSP to collaborate on some of these tests.

However, there’s extensive groundwork to be done. Advertisers need to grasp how to categorize their target audiences into so-called interest groups — think of these like audience segments, for instance, individuals who added items to their cart but didn’t buy. 

Haig clarified this process: “We’re going to be working with advertisers on how to add browsers to interest groups and then we’ll be working with buying platforms on how to buy against them.”

It’s no surprise that ad executives like Haig are testing parts of this feature incrementally, rather than all at once.

“We’re bringing partners to the table [for the Sandbox] and are actively testing it,” said PubMatic’s svp of marketplaces and addressability Andrew Baron. “There is movement now, timelines are becoming more real and not getting kicked down the road.”

None of this movement really matters without measurement. The effectiveness of the Protected Audiences auctions in Chrome hinges critically on marketers’ ability to measure the impressions they buy. This is where the Attribution Reporting API part of the Sandbox becomes vital. It’s the only way marketers can accurately track whether someone viewed, clicked or converted after seeing a Sandbox ad. Protected Audiences and the Attribution Reporting APIs are, in effect, two sides of the same coin, both critical to the success of the other.

Rob Webster from Goodway Group breaks it down: His agency is comparing tracking ad performance (or conversions) from a client’s DSP or ad server with what’s happening in the Sandbox to get the lowdown on the differences.

MiQ is on a similar track. It kicked off its own comparison in the U.K. just a couple of weeks back and is looking to expand this testing worldwide pretty soon. As Haig put it, her team is inviting clients to place new tagging for the Attribution Reporting API testing. 

However, these tests are more than mere comparisons; they aim to illuminate a suspicion long held by marketers regarding this part of the Sandbox: that a lot of advertisers are going to struggle to get reasonable reports out of it because they can’t generate sufficient conversions — as an example, by default, this would be a thousand for daily reports. Falling short of this threshold results in the introduction of additional “noise” data into the campaign data to maintain anonymity or the need to simplify the report. 

For marketers, it’s a bit of a conundrum. They have to grapple with the reality that without reaching those conversion numbers, their campaign data won’t offer much detail. Instead, it might get drowned in ‘noise,’ making a substantial portion of recorded conversions unreliable. 

“Unless you’re a big advertiser you might struggle to get the numbers out of this part of the Sandbox,” said Adform’s chief technology officer Jochen Schlosser.

To some, this may not be as much of an issue if the performance is there. As Epsilon’s chief analytics officer Loch Rose said, good performance will still be good, and great performance will still be great. However, where it does hurt is that marketers are essentially being made to report on relatively large segments of pre-determined users, via Protected Audiences interest groups, making it challenging to derive actionable insights from the reporting for campaign optimization. For instance, if a marketer guesses wrong on the segments to report on, they may all perform similarly but collectively not well enough, Rose continued. 

It looks like the biggest headache in the Sandbox is also the most crucial bit: comprehending the measurement results.

“The measurement implications of all of this are really important because we need to have an infinity loop so that our clients know that if they’re going to be targeting a certain audience that they can say with some certainty whether it’s worth what they’re paying — are they getting the payoff they were expecting,” said Goodway Group’s chief media officer Stephani Estes.

And then there’s Topics. Marketers are already drawing up plans to test how to target the general subject of categories of the sites someone visits like sports, fashion or cooking.

Word has it, some of those Sandbox brainstorming sessions took place in the meeting rooms of Consumer Electronics Show in Vegas last week. Marketers and agencies met with Google execs and ad tech vendors like Index Exchange to hash things out, with a keen focus on the targeting-driven aspects of it. 

“Testing is being lined out now, specifically when it comes to the Topics and Protected Audiences parts of the Sandbox,” said Andrew Casale, CEO of ad tech vendor Index Exchange, who was in some of those meetings. “One of them has already happened where the agency talked through testing designs. They’re bringing a couple of clients to the table to bring test budgets to run fully cookieless campaigns in Chrome.”

When it comes to those tests for Topics, the focus is going to be on comparing the performance of those ads with those bought on the back of third-party cookies. Regardless of the outcome, marketers appear to have almost made up their minds about Topics. They’re eyeing it as just one piece of the targeting puzzle, not really a game-changer by itself. 

Just to set expectations, we’re not talking about massive tests here; think more along the lines of campaigns in the five-figure range, at best. Spending more would be foolhardy given that only one percent of traffic in the Chrome browser will remain without third-party cookies. Really the aim of these tests isn’t scale, it’s to get a feel for how this sort of advertising works. 

Goodway Group’s Estes makes this crystal clear: “These tests aren’t just fluff. We’re not in a situation where marketers can say, ‘Let’s see what happens and figure it out.’ They’re not testing a new channel here. This isn’t optional; it’s an essential part of a marketer’s strategy to transition away from what has existed.”