The history of innovation follows a familiar pattern: New technology comes along that improves our quality of life—like the car—then time reveals its unintended consequences, and innovators jump in to create something even better.
In ad measurement, what once was innovative became commonplace, and then concerning, with 1:1 tracking tactics like third-party cookies and hashed emails that were the accepted industry standard now under fire for their privacy risks.
Just as car companies are committing to zero-emission vehicles in the coming years, innovative ad-tech platforms must commit to zero cookies and find a better way to resolve the trade-off between addressability and privacy.
We still aren’t sure exactly how the landscape will change, and there are still many unknowns. But we do know that the loss of cookies means that marketers should rethink their strategy and at the most basic level, understand and integrate the use of cohorts into their revamped approach.
Invest in cohort-level measurement
To innovate beyond the current problematic industry standard, brands must find a way to implement accurate, efficient, and privacy-compliant cohort-level measurement into their advertising strategy. Cohort-level measurement is the future-proof alternative to 1:1 identifiers. It provides brands with significant cross-channel lift and targeting accuracy analytics, without sacrificing any individual consumer’s privacy.
Advertisers can take a page from Google and find a middle ground that allows for accurate targeting that doesn’t identify individuals in any way. Google’s FloCs are cohorts, and advertisers can work with their own data and with partners to develop a similar approach to solve for the same problem.
Cohorts can be as small as a household, a workplace or any collection of people within a tight geographical area. With the right cohort-level measurement technology, these audience segments are measured as a privacy-first substitute for the unidentified individuals. Although you won’t know exactly which individuals were exposed to an advertising campaign, you can still determine the exposure probability of each cohort.
If you’re worried about making sure your driver analysis shows actual causal impact instead of just correlations, look no further. Accurate cohort-level measurement ensures that measuring incrementality is not any different than with individual-level tracking. You can then create hold-out groups at the cohort level to conduct controlled experiments or use a larger set of cohorts within a campaign to analyze with casual machine learning.
Cohorts are all-encompassing
As more users continue to become unidentifiable through 1:1 identifiers, non-reliance on such identifiers, including hashed emails and device IDs, mitigates the individual-level contamination of test and control groups.
Cohorts also offer an effective approach to cover an entire digital media plan. With only 10-30% of the internet logged in to an email identifier, cohorts can fill in the gaps without compromising any sense of privacy for users or accuracy for advertisers.
So, who is innovating with cohort-level measurement? Clean rooms are a central source of innovation in cohort-level measurement. Many clean rooms, like Ads Data Hub, use query thresholds to constrain data sharing to a cohort-level, and more advanced clean rooms, like InfoSum and Habu, are exploring ways to join data on a cohort-level. Browser-based measurement approaches, such as the Webkit Privacy Preserving Click Attribution, measure impressions and conversions at a cohort-level.So while cohorts are still a new measurement concept, there are many players in the space that are leveraging and investing in this privacy-alternative to hashed emails and cookies.
If you’re still a little skeptical about cohort-level measurement, try using the individual-level and first-party data you’ve collected over the years to compare to cohort-level measurement techniques. Not only will this validate the cohort-level models, but you can also leverage such first-party data — or opted-in second and third-party data — to calibrate and improve the techniques. For example, by measuring basic reach or engagement among first-party audiences, for whom you have 1:1 data, and comparing it with results from a clean room, a browser attribution API, or another cohort-level measurement platform, you can validate and calibrate the results.
Innovation in privacy-friendly, cohort-level measurement depends on whether the Ad and MarTech industries are ready to move on from yesterday’s cross-domain, individual-level tracking methods and embrace the future. Those who successfully implement meaningfully segmented audience cohorts and measurement models will have a firm grasp on the impending privacy-first advertising future.