Google’s decision earlier this year not to replace third-party cookies in Chrome with any of a variety of new proposed personal identifiers rocked the ad tech world. Coming on top of Apple’s moves to require that people opt in to any tracking tech, it signals the end of personal identifiers across platforms and publishers.
But despite all the hand-wringing, the “cookieless future” is far from the end of the world for marketers. Indeed, for many marketers and marketing generally, it could make things better.
The reality is that digital targeting has often been woefully bad—serving people ads for things they just bought, in sizes and varieties they didn’t want, or solving problems they never had. That’s when the ads reached real people at all. Third-party cookies have been a key tech enabler of digital fraud. And many marketers already have been building their own consumer databases anyway, in part because of the poor quality of third-party data that fueled cookie-driven ad buys.
To be sure, the effects are uneven. Marketers most positively affected are those who are essentially media players too—the retailers whose first-party transaction data and direct consumer relationships will become more valuable to other brand marketers. Brand marketers who’ve invested in their own consumer databases—such as Procter & Gamble Co., Unilever and L’Oreal, each numbering more than a billion consumers globally—will see the value of that data grow as it’s used to target and model ad buys. But some smaller direct-to-consumer players who rely more on targeting using third-party cookies may be at a disadvantage.
Varying impact
“The advertisers that are going to feel the most amount of pain are the small and medium-size businesses, because those companies have been heavily reliant on third-party cookies,” says Rachel Tipograph, CEO of MikMak, provider of e-commerce enablement and analytics software, including shoppable ad formats that let brands direct purchases to shopping carts at a variety of retailers. “The behemoth CPG brands that didn’t build their customer acquisition strategies around third-party cookies, I don’t think this affects them much at all.”
Indeed, the outcry from big brands has been relatively muted. For example, Unilever Executive VP for Global Media Luis Di Como in a statement says his company will lean on its first-party data and relationships with retailers to reach people and “as a responsible marketer, we will continue to support partners that are committed to better data privacy.”
“Over the course of the past couple years, the mindset we have simply taken is that we will have a more privacy-focused internet with more opt-in, that’s consent-based, with consumers having a say in what data they provide,” says Procter & Gamble Co. Chief Brand Officer Marc Pritchard. “As our good friend and CEO of Gillette Gary Coombe says, ‘If it’s inevitable, get enthusiastic.’”
Yes, P&G is part of a group of 30 advertisers in China that have explored working around Apple’s opt-in requirements for sharing identity information across apps. But less than 5% of P&G’s consent-based data comes from app-to-app sharing anyway, Pritchard says. And P&G took issue with a recent Wall Street Journal story characterizing the effort as a furtive attempt at sidestepping Apple’s privacy controls, saying that consumers would have the ability to turn data-sharing on and off under the plan, and that P&G is working on it directly with Apple “or anyone else who shares our values of giving consumers control and transparency.”
Meanwhile, P&G has its own global database of more than 1.5 billion consumers, a combination of third-party data and first-party data gathered everywhere from direct e-commerce sales to expectant moms sharing information on the Pampers app to quick quizzes people fill out online facilitated by ad tech player Jebbit. The approach was never to fuel individual ad targeting, though, but rather to help create “smart audiences” that can be reached in digital and other media based on context of the media or what P&G learns about the group’s media habits, Pritchard says.
“Those smart audiences, which are really cohorts, we can reach through programmatic means, and frankly in an anonymous way,” Pritchard says. “With Google, we’re going to be able to match our smart audiences with their FLoCs [Federated Learning of Cohorts] to reach people. That’s really how it’s going to work. I gotta tell you, I think it’s going to be fine.”
P&G will reach its smart audiences in some cases via contextual targeting. like cooking content for Dawn Power Spray, and in places where people provide opt-in access to their data “because they’re trusting you.”
And yet for one of P&G’s much smaller, mainly direct-to-consumer competitors, household products marketer Dropps, the crumbling of cookies looks crummier.
“In our case, not having the cookie really impacts re-targeting and display prospecting. It will be more complex to build customer personas,” says Dropps CEO Jonathan Propper. Ultimately, he thinks the loss of personal digital identifiers will force Dropps to “create better experiences for consumers online, which may ultimately have an extra effect of customers becoming evangelical for the brand.”
Google Ads is about half Dropps’ ad budget, with Facebook and Instagram advertising—which should be less affected by loss of identifiers—composing most of the rest.
Dropps also benefits from a first-party database of past buyers, many of whom are subscribers essentially, Propper says. But even though business for his e-commerce laundry and dishwash determent pods has boomed during the pandemic, that doesn’t mean he can give up on prospecting. Amazon, which was always part of his business beyond direct sales, may become more important and certainly won’t be less so.
Tina Pozzi, chief brand officer of beauty brand Bliss, a direct-to-consumer player with considerable retail distribution at Ulta and elsewhere, says the brand has relied heavily on cookies for re-targeting in the past, but increasingly has been building its first-party relationships through programs like a skincare quiz via Jebbit, distributed through online influencers. Bliss also has been building its community through TikTok with agency Movers & Shakers, generating some 6 billion views.
“In some ways I think this is a really big challenge for brands to ensure that they’re developing that relationship with their consumer,” Pozzi says. “Ultimately this will be about understanding the consumer first and building that first-party data to augment our targeting strategy.”
But having a d-to-c business doesn’t mean heavy reliance on cookies. An executive who has started a handful of successful d-to-c brands who asked to speak not for attribution said he long-ago rejected programmatic buys using cookies on the open web because it simply didn’t work as well as Facebook and Instagram advertising.
His most successful media strategy is one that somewhat resembled old-fashioned consumer products marketing applied to social media—putting as much money as he can into reaching as many women ages 18-to-49 as he can on Facebook, Instagram and, increasingly, TikTok. That’s augmented by rapid-fire testing of ad creative to see which ads generate the most sales, and then putting more money against them. He doesn’t expect Google or Apple’s privacy moves to affect that strategy.
Leaving behind a world of bad targeting
One reason for marketers to dial back panic is that personal identifiers have fueled years of really bad advertising, as Norm de Greve, chief marketing officer of CVS, sees it.
“I spent 14 years at Digitas, and one of my roles there was running the analytics group,” de Greve says. “I saw the development of all the programmatic targeting technology, the data exchanges, all of this. All that was about putting ads in front of people most likely to buy. And I don’t think the manifestation of that has really been that great. We’ve all seen the re-targeting ad for the product we’ve already bought or have no interest in. I don’t know of anyone who says, ‘Wow, my advertising experience is just awesome now.’”
Some targeting is just inexplicable, such as display ads that have been stalking an Ad Age reporter in recent weeks with information about a blood disease he doesn’t have—paroxysmal nocturnal hemoglobinuria (PNH).
Indeed, almost no one has the disease—as few one in a million people, according to some estimates. But biotech firm Apellis is out to raise awareness about the illness as it recruits patients to an early acceptance program for its new PNH drug with help from demand-side platform MediaMath.
Apellis didn’t respond to email requests for comment on what someone might have done to elicit ads about the rare disease. Nor did MediaMath have any specific explanation, but said it uses third-party identifiers to help advertisers “drill down into audience segments and build specific profiles based on location, demographic data, interests, etc.” MediaMath says it encourages clients to use “reasonable frequency capping.” Even so, the reporter has been seeing Apellis’ PNH ads several times daily for six weeks.
Even when targeted ads reach the wrong target, that’s still probably better than reaching no one at all, which is what cookie-driven programmatic ad buys often do. Cookies underpin a programmatic marketplace rife with fake identities making fake clicks on fake websites—but at really favorable prices if you don’t count the fact that they don’t reach real people, says Augustine Fou, cybersecurity and ad fraud consultant. He’s a decided skeptic of widely used fraud detection technology that his research shows is regularly outsmarted by fraudsters.
“Doing away with third-party cookies from Chrome will not harm advertisers, just their egos, because ad tech targeting was awful to begin with,” Fou says.
“Advertisers don’t need to know which individual bought their product,” he says. “They just need to know that individuals (plural) that were exposed to ads ultimately bought their product. Third-party cookies allowed ad tech to prolong the fiction that microtargeting was magical, so they could keep selling microtargeting services.”
That doesn’t’ mean the new world order will be fraud-free, Fou warns. Google’s FLoC IDs can be forged just like individual identities, which will allow bots to pretend to be FLoC audience groups rather than individuals.
Rise of retail data
Whatever happens with cohorts, retail databases will still be rich with data about real people buying real products.
“For us, it raises the value of first-party data” to have personal identifiers go away in Google properties, says CVS’ de Greve. “There continues to be a growing strength between the retailers and the CPGs to use the first-party data to drive more sales. It’s going to push everyone more into their first-party data.”
For retailers, in particular, Google’s announcement has only heaped further fuel on fast-growing interest in finding new ways to apply all their data to ad sales. Amazon, whose 77% growth in ad revenue last quarter is already outpacing growth of its retail sales, stands to be possibly the biggest winner, with growing opportunities outside its usual search, display and video offerings. While its recent exclusive deal to stream “NFL Thursday Night Football” is widely seen as a way to build Amazon Prime membership, the deal could also let Amazon sell individually targeted, un-skippable connected TV ads based on sales data.
Other retailers keep piling harder into selling targeted audiences based on their data too, such as Walmart, which earlier this year enlisted The Trade Desk to help it sell programmatic advertising that plays off its growing first-party database of sales. Even DoorDash is developing an ad network built largely around first-party data about what and when people order from restaurants.
Potentially, all these things elevate databases and ad platforms other than Google’s. That leads to some speculation that Google will eventually restrict the ability of retailers or publishers to trade in their first-party data within Chrome. But a Google spokeswoman said the company has “noted explicitly that we’re going to continue to respect direct first-party relationships that brands and publishers have with users. Having first-party data that is collected in a privacy-safe manner is increasingly important given regulation, privacy concerns and browser changes.”
Regardless, personal IDs aren’t the only way marketers can target likely retail purchasers. Galderma’s Cetaphil skincare brand earlier this year targeted people based on time of day—such as based on phone movement to indicate when they’re waking up and starting their skincare routines, or when research shows they’re likely to be making weekly online stock-up buys. Cetaphil used Aki Technologies to serve them shoppable ads from Arc Worldwide that then let people drop products into Amazon, Target or Walmart shopping carts.
Though Galderma didn’t have access to data on whether those items ultimately got checked out, estimates based on industry standard benchmarks show the mobile campaign generated $1.3 million to $2.7 million in sales for a return on ad spending of $1.51 to $3.25 per dollar spent. The campaign worked because it made things convenient for shoppers at times they wanted to shop, says Beth Clevidence, shopper marketing lead at Galderma, who plans to keep using the approach. And Aki says its “moments first” daypart targeting is entirely unaffected by loss of cookies.
Back to the future
One of the biggest potential threats to marketers from losing personal identifiers comes in analyzing how well advertising works broadly. Moves by Google and Apple are likely to prevent marketers or their third-party number crunchers from anonymously tracking the same people across platforms owned by different companies to know what ads they see and what they buy. Even if you’re not out to measure individual clicks, that makes apples-to-apples comparisons harder.
Pritchard believes that ongoing talks with Google, Apple, Facebook, the Association of National Advertisers, World Federation of Advertisers and others will ultimately lead to a privacy-compliant solution that will make it possible to track individuals across platforms for the purpose of measuring ad effectiveness and limiting ad frequency.
But one outcome of the likely shift of more money to advertising via retailers could also reinforce a longstanding marketing analytics fallacy: last-click attribution. That’s the practice of automatically giving all credit for a purchase to the last ad someone saw before doing the deal. So an Amazon ad will likely get all the credit for an Amazon purchase even if someone saw dozens of ads previously that led up to the purchase. Since those prior stops on customer journeys will be invisible, it will be harder to give them credit. That’s likewise true for sales or other actions that happen right after someone sees Facebook, Google or other “walled garden” ads, since only what happens within each media owner’s property can be easily attributed to an individual.
“It’s going to drive a lot of last-click attribution, because that’s the only thing available,” says Jeff Greenfield, a longtime marketing analytics executive and former CEO of multi-touch attribution firm C3metrics, who’s now senior VP at ad management firm WideOrbit.
The end of personal identifiers may render multi-touch attribution of the sort Greenfield did at C3 obsolete, he acknowledges. MTA had become the buzzy next generation of marketing analytics only a few years ago, but Greenfield believes eliminating personal identifiers will give new life to an older approach—marketing mix modeling—which combines statistical modeling with more limited information to estimate the impact of spending.
Thanks to technical progress by modelers and much more spending on digital platforms, marketing mix models are better positioned to evaluate media effectiveness now than a decade ago, Greenfield says.
He also expects a return to “incrementality testing,” or small, quick media experiments that evaluate how specific media channels and ads do at increasing sales or responses. These tests have some of the same advantages of attribution, he says, but don’t require personal identifiers.
Brand tracking surveys, whose demise has also long been predicted, will also likely see a renaissance with marketers, he says.
“Everything old is new again,” Greenfield says. But that’s not necessarily a bad thing, because he believes the targeting capabilities that cookies created fueled short-term transactional thinking that hurt marketing.
“The biggest beneficiaries are going to be the brands,” Greenfield says. “If they’re smart, instead of being upset that data is being taken away from them, they’ll step back and experiment more with brand building and see what happens.”
And ultimately that just reinforces the power that bigger brands have, says Tipograph. They already had more recognition, so didn’t need to try to save money by targeting so tightly using personal identifiers. Yes, bigger brand marketers increasingly are using MikMak’s shoppable ads to get more first-party data about their consumers. But even before the loss of cookies, marketers such as P&G were using that data primarily to inform buys aimed at cohorts, segments or “smart audiences.”
“I’m excited about this new world,” Tipograph says, “because it’s going to force marketers to be creative and care about brands and understand people’s relationship to the brands.”