U.S. publishers experience testing fatigue as they evaluate alternative IDs to third-party cookies

Condé Nast is one publisher that’s only testing “the most vetted and premium” potential alternative ID solutions in the market that “its core set of advertisers” have asked to use in campaigns, according to Deb Brett, the company’s chief business officer, digital. Brett declined to share which identifiers are currently in that cohort and how many they’re testing.

Publishers’ first-party data solutions and alternative identifiers are both aimed at replacing third-party cookies, which will soon come to their ultimate end once Google removes them from its Chrome browser by the end of 2024. But alternative IDs are more appealing to buyers than publishers, according to Mike Nuzzo, Hearst’s vp and head of audience intelligence and insights. That’s because alternative ID providers boast capabilities like frequency capping, attribution, cross-site analytics and activation (similar to third-party cookies), but require publishers to give up some control.

First-party data offerings, on the other hand, build upon a publisher’s existing content, allowing them to maintain control of what goes into measuring a successful ad campaign. Similar to Condé, its in-house first-party data offering is the priority for Hearst over alternative identifiers, Nuzzo said. That doesn’t mean that Hearst won’t also use both solutions in tandem, however.

“A lot of us are not all-in on one solution because [alternative identifiers] are not even one [solution],” Nuzzo said, adding that there are dozens being touted around the industry at the moment. “And instead of testing we’re actually having to invest in them. I think we’re starting to see testing fatigue because it actually requires investment. It’s not to say we won’t [use alternative identifiers]. It’s just we’re going to test them all” before backing only one or two, he added.

But even if every one of Hearst’s advertisers requested the same alternative ID, Nuzzo and his team would still have to vet it and consider whether it balances privacy regulation and regulatory constraints, before adopting it as its own.

“Having the bandwidth to be testing everything is not within our core focus, because, foundationally, at our core is [our] first-party solution, [Spire],” said Brett, adding that she has faith that the industry will surface the best options over time and those ID solutions will be the ones her team considers adding into their advertising business model. 

Publishers like Dotdash Meredith and Insider are in the camp of entertaining any and all alternative IDs that advertisers are asking to use, though they’re hoping that their first-party offerings will cover up any lags in ad revenue lift. 

“It’s not up to us. It’s up to our clients. And If we do our job, whatever happens to these third-party identifiers is not our issue,” said Neil Vogel, CEO of Dotdash Meredith. 

Vogel’s plan is to show advertisers that the company’s new intent-targeting tool, D/Cipher, will be the best way to buy the publisher’s ads and reach their target customers most effectively, though he added that his team will not “bully the market” to get advertisers to use D/Cipher.

Insider’s Chao Liao, svp of advertising, said it’s rare for an advertiser to request a particular ID. But he did note his concern that publishers are waiting for buyers to align their ad dollars and, collectively, that will decide the future of the alternative IDs in the post-cookie ad industry.

“That’s such a chicken or the egg thing,” Liao said. “By the end of day. I think everyone’s lukewarm about it, so that’s why I’m not focused on figuring out which one of the IDs really work.” 

Taking matters into their own hands

Not all publishers are willing to leave it up to the buy-side to decide which third-party cookie alternatives will stick around, however.

Justin Wohl, CRO of Salon, TVTropes and Snopes, has taken the more drastic measure of only using Unified ID 2.0.  

“UID is for me the number one and only ID solution that I expect to see a result from at this point,” said Wohl. “I’ve tested so many I have extreme frustration with how little results have happened. I got so heavily sold — honestly I ate it, hook, line and sinker — on some of these ID solutions that have led to absolutely nothing.”

As it stands, Wohl and his team are not able to see which advertisers transacting within a DSP are asking to target users using a specific alternative ID, like LiveRamp’s RampID or ID5’s ID, and the proprietors of said alternative IDs “are being extremely fucking coy about confirming who is using it on the other side” of the transactions.  

Over the last two years, there’s been an apparent lack on the part of alternative ID vendors proving that their solution is being not only adopted but used by advertisers, Wohl continued, which gives him enough pause to step away from those IDs almost entirely.  

“It’s just maddening to me because I really took the time to integrate them early and try to match IDs to users and we’ve just had no transactions that were linked to that effort. So super frustrating,” Wohl said. 

Testing takes time and isn’t always easy

Publishers that are open to any and all alternative IDs may not be necessarily implementing them properly, which is why Mike O’Sullivan, co-founder of ad tech tracker Sincera, said that he’s noted frustration amongst publishers who claim they are not seeing consistent ad revenue results, despite testing alternative identifiers for months.

According to Sincera data, a lack of results from the alternative IDs on the publishers’ end could be a result of improperly testing them in the first place. That wouldn’t be surprising, however, considering how many IDs there are and how difficult it is for a publisher to conduct a proper test, O’Sullivan added. And yet, segmenting traffic and isolating a single identifier to that traffic cohort, though a risk, is the only way to prove that it can generate a revenue lift.

TVTropes is currently in the midst of a six-week-long test that’s running UID 2.0 against the brand’s logged-in users, isolating a small percentage of those users as the control sample, Wohl said. After the six weeks are over at the end of July, his team will determine whether to register Salon’s and Snopes’s users without having them pay for a subscription, to run UID 2.0 against those subsets as well.

Once he has solid confirmation that the UID 2.0 option is more valuable than no IDs at all, Wohl said his team will roll it out across the company, which would happen before the fourth quarter.

“Most of the time, just turning on that toggle isn’t enough to actually get an ID up and running. Publishers will [say,] ‘I want all the lift possible’ so [they] turn on everything … [but] the result of that is a lot of identifiers being turned on and off and it’s sometimes difficult to tell if this is the result of a very intentional [action],” said Ian Meyers, the other co-founder of Sincera.

Anecdotally, Meyers estimated that fewer than 5% of publishers in the U.S. have properly segmented their traffic and tested a single ID within each traffic cohort — how TVTropes is currently testing UID 2.0 — which is what Meyers claims is the most effective way to prove if an identifier is working. 

“I think you would struggle to find any publisher that was perfectly configured,” O’Sullivan said. “Publisher testing with holdout groups is orders of magnitude more complex in terms of time to invest, so I just don’t see a lot of publishers doing it.”