Third-party cookies may be on the way out in Chrome, but that doesn’t mean the end of data-driven advertising.
There are still many other types of audience data that marketers will have at their disposal, including identity graphs and consented third-party data, household-level data, second-party data, contextual data, cohorts and, of course, first-party data, including login data.
Marketers now have the complex task of figuring out how these components fit into a sound identity strategy, said Adam Broitman, an associate partner focused on digital marketing at McKinsey & Company.
“In the near-term, the cookieless world will be complex,” Broitman said. “Advertisers will have to test many solutions.”
Many solutions, indeed. Sometimes it feels as if there’s a new purported cookieless solution born every minute (and maybe a sucker, too).
Here’s a cheat sheet for the six primary types of targeting data marketers will need to get comfortable with before third-party cookies go away for good.
1. First-party data
First-party data – email addresses, in particular – will cure all that ails you. Well, almost.
Login data is often considered to be the gold standard of consumer data, said Rohini Sen, managing partner and head of audience science and measurement at Wavemaker. It’s even more valuable when consumers validate their information, such as signing up to receive email or text messages in return for a coupon or some other benefit.
(It’s still – seemingly eternally – up for debate, though, as to how valuable consumers find the more classic value exchange of data collection and targeted ads in return for access to free content.)
“Once someone reaffirms their information, you can see that, yes, this is a real person,” Sen said. “And then when they actually log in you know you have at least one reliable method to get their direct attention.”
Pros: Consent, accuracy, helpful for measurement and handy as a truth set to train machine learning models.
Cons: Scale and potential privacy concerns.
“There are two main questions: scalability and are we littering the web with PII?” said Nishant Desai, director of technology and partnerships at Xaxis. “If every site requires something like Unified ID 2.0, scale will grow, but this data is arguably more exposed if everyone is out there flogging it.”
2. Identity graphs
Identity graphs will continue to be viable, and even vital, following the deprecation of third-party cookies. But some of the inputs will have to change, as will the linkages that are made among different data points.
Today, third-party cookies are a linchpin of most identity graphs.
“At Xaxis, we do a bunch of cookie matching and we’re going to have to do similar bridging and matching post-cookie between multiple sources,” Desai said. “There is an infrastructure cost to that and identity graphs need to be maintained, so I’d say this is a solution for companies and large publishers that can afford it.”
Pros: Marketers and publishers have an opportunity to rebuild their identity graphs on a stronger foundation for the long term.
Cons: Identity graphs are resource intensive.
“If you can bridge enough data, they’re useful,” Desai said. “But you have to ask yourself if you’re getting the returns and if it’s worth it financially.”
3. Household-level data
Household data has been “the workhorse of media and advertising for years,” McKinsey’s Broitman said, especially for full-funnel marketers looking to build brand equity, because they need to cast a wide net.
But there’s a catch, of course. Most household targeting relies on an IP address, and the future of IP address targeting is shaky at best, said Wavemaker’s Sen. Just because it’s common practice today to use IP addresses to associate multiple devices that connect to the same WiFi network doesn’t mean there’s a future in it.
“I’m advising my brands to steer clear of it, because it doesn’t seem to respect the intent and direction of where privacy is going in the industry,” Sen said. “On the flip side, though, anything with an opt-in is still very much on the table, including the sort of shopper data that a company like Catalina collects.”
Pros: Enhanced reach and has worked well enough on TV for years.
Cons: Diminished personalization and potential privacy risks associated with cross-device tracking methods that use device fingerprinting.
“Still, I don’t think less individualized ads necessarily has to mean less personalization,” Desai said. “When audience targeting and programmatic took hold, the industry largely abandoned household targeting; but now, with the end of cookies, it feels like everything old is new again.”
Speaking of breathing new life into legacy methods, contextual targeting is undergoing a renaissance as well as marketers assess their post-cookie options.
Data collaboration has a lot of potential – if the data sets are merged safely and stored securely.
And data consortiums could shift power back to publishers and away from the obsession with audience-based targeting. “You can think of it almost as an extension of the first-party relationship,” Sen said.
The danger is that second-party data partnerships often aren’t clear to consumers. “They don’t know which sites are tracking their data, how it’s happening or who’s involved,” she said.
Pros: Power to the publishers, scale, improved ad targeting and measurement capabilities.
Cons: Consumers might not be clear about how data is being shared on the backend. Proper data security and governance is paramount in light of platform changes and increased regulatory scrutiny.
Who could forget about cohorts?
Google has been dominating the headlines with FLoC [Federated Learning of Cohorts], but Google’s proposal isn’t the only game in town.
Cohort-based targeting – using common behaviors and interests to create anonymized groupings rather than tying these attributes and experiences to an individual identifier – is a key concept of post-cookie product development.
Pros: Theoretically privacy safe …
Cons: … but could introduce new privacy risks of its own. Cohorts, for example, do nothing to eliminate predatory targeting. They’re also unproven in a cookieless context. Google’s controvertial FLoC experiment, for example, relied on real-time access to cross-site publisher data that won’t be available after Chrome removes third-party cookies.
“There are subtle differences in how FLoCs and cohorts are collected and identified,” Desai said. “I don’t see any good way, at least right now, to be able to forecast the performance of a FLoC.”