"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Ionut Ciobotaru, Chief Product Officer at Verve Group.
The past has finally caught up to the marketing industry. We can’t pretend we didn’t know it would. As an industry, we perverted identifiers and cookies, we failed to proactively address concerns raised by regulators and platforms, and now – like children who failed to heed the first dozen warnings from their parents – our much-abused toys are being taken away.
Although Google’s decision to delay final cookie deprecation until 2023 prompted a sigh of relief in many circles, the writing is still on the wall: The way we attribute ad performance, particularly omnichannel measurement, needs to fundamentally change — and soon.
Good. We had it all wrong anyway.
Cookies decay quickly and fail to aid advertisers in making cross-channel and multi-touch connections. Plus, third-party cookies are already irrelevant in Safari and Firefox. Reliance on this dated technology is resulting in increasingly skewed and inaccurate audience understandings.
It’s time to stop taking comfort in meaningless metrics. We now have the opportunity to take a step back and rebuild our approach to metrics – properly this time.
We need to start with a few simple truths
Let’s face it: No one knows exactly what omnichannel measurement and attribution will look like five or 10 years from now. But we do know a few things.
First, inventory is going to be mostly sold without identifiers. Attribution will shift to probabilistic approaches that rely far more on econometrics and media mix modeling than they do today. Second, first-party data and direct user relationships will become more valuable than ever. They will help improve conversion modeling within each channel, increasing the effectiveness of media buying.
The shift away from deterministic attribution tactics will disproportionately affect brands that centered their media buys on web and mobile. We’ll likely see a growing willingness to experiment within other channels, including emerging opportunities like connected TV.
As the attribution playing field levels across the advertising landscape, advertisers will rebalance media strategies to include a more productive blend of awareness-driving tactics vs. an overreliance on direct response channels.
To understand full-funnel advertising, marketers will need to put a greater emphasis on lift-measurement studies and incrementality.
Let’s come to terms with necessary complexity
I’d love to tell everyone that the future of omnichannel measurement and attribution will be vastly simpler in the future. But that’s not likely to be the case, at least in the short term.
Every channel brings unique strengths and limitations to the table. We need to recognize the best practices for measurement and attribution within each channel and ensure we connect the dots in a logical and sustainable way for a privacy-first world.
I predict we will see a rise in privacy-forward tactics like:
- On-device measurement capabilities, which can generate audiences at the device level and make the audience segments, rather than the individuals, available for targeting
- Indirect feedback loops, where personal user information is masked but outcomes are still fed back into attribution understanding
- Aggregated or modeled conversions measurement, which is a specific indirect feedback loop that uses known user data to infer conversion rates where privacy or technical limitations limit direct observation
All of these approaches protect user identity but still deliver valuable insights to the advertiser, and some of them are already available within the walled gardens.
Here are channel-specific considerations for attribution we need to recognize as we move forward:
Mobile: Mobile attribution has come a long way, despite regular policy shifts by Apple and Google. These days, multi-touch attribution, including incrementality, is possible within mobile channels.
Now the focus will become extending that understanding into other channels to form a truly omnichannel picture. As an industry, we need to embrace cohorts (most visibly embodied by Google’s Federated Learning of Cohorts), but advertisers also need to start crunching their own numbers versus relying on self-attributing networks. We also need to move away from our obsession with instant feedback loops and adapt to longer – and aggregated – user journeys.
Web: Google’s shift to cookieless will deliver the heaviest blow to the digital web. Alternate identifiers and privacy-first measurements that recognize incrementality must replace cookie-based measurement. New solutions in this space must be built with non-web connection points – places that don’t rely on cookies, such as mobile, CTV and offline data.
We should also embrace web solutions, like Safari’s Private Click Measurement and Google’s Federated Learning of Cohorts, and experiment with Facebook’s and Google’s modeled conversions measurement, while at the same time using or building more holistic solutions for the open web that enable unbiased, omnichannel measurement.
Connected TV: Without clicks to track, CTV will continue to focus on statistical inference, delivered by models that can demonstrate the value (i.e., awareness or sales lift) of spend by using control groups. But with digital capabilities comes the possibility for external digital connections as well. QR codes and coupons spring to mind as a potential way for connecting CTV to mobile devices in the user journey, similar to how SMS was used within TV.
Digital OOH: In a privacy-first world, footfall attribution is likely to take a backseat to more traditional awareness metrics when it comes to DOOH measurement. And that’s OK, provided our industry puts more effort into understanding the interplay between DOOH and other channels.
Gone are the days of dropping a pixel or tag, then matching for attribution. As we look for new ways to connect the dots for a more holistic omnichannel understanding, media mix modeling and control groups will play a vital role in full-funnel measurement and attribution.
Panel-based measurement, such as that conducted by Nielsen and Comscore, will add value by evaluating statistically relevant samples from which advertisers can extrapolate results and feed those into their conversion models. Likewise, feedback forms — such as those employed by platforms like YouTube — will help fill in key gaps in brand affinity, brand lift and other top-funnel outcomes driven by ad exposures across digital and mobile channels.
Don’t think of the next two years as a reprieve from a cookieless reality. Rather, think of them as much-needed time to reimagine campaign performance across channels. Past campaign tactics are going to become irrelevant. Let’s put in the hard work to recalibrate what “success” looks like in a privacy-first digital world.