As the sun begins to set on the cookie-powered web, some of the brightest minds in the industry have been reimagining what the next chapter of the internet will look like.
The answer needs to come sooner rather than later. Though talk of phasing out the cookie has been circulating the industry for years, the introduction of the The Banning Surveillance Advertising Act could mean the moment of reckoning is coming sooner than once thought. Though the bill faces strong opposition from the IAB and more recently, the U.S. Chamber of Commerce, the proposed bill is the most restrictive to hit the Senate’s floor to date, in keeping with rising social and political backlash against the advertising industry’s data collection tactics.
This has many companies and businesses that rely on existing cookie infrastructure concerned about how their businesses will be impacted. According to research by Statista, more than 80% of marketers are dependent on the cookie in some way. To back that up, results from an A/B test run by Google saw a 52% decline in revenue for publishers with cookies disabled as compared to the control group.
But solutions are emerging, and many of them are already bearing promising results. Finding viable alternatives to cookies and third-party data all starts with understanding the current roles they play, why they are problematic, and how to build back in a way that is mindful of user privacy.
In many ways, cookies and third-party data are the building blocks of the web as we know it. With most users logged simultaneously into several applications – Gmail, Facebook, Twitter – around the clock, sharing data and tracking users across sites allowed for unprecedented levels of granularity in profiling audiences.
From the standpoint of a web-based business, the more that is known about the user, the better they could match audiences with the right content and products. It was a win-win.
But as is often the case, there can be too much of a good thing. Years and years of amassed cookie-collected data based on billions of people ultimately grew alarmingly concentrated into the hands of a few.
What users liked on Facebook, the historical log of their Google searches, which sites they visited and what content they engaged with, were all pieces of sophisticated audience segmentation that at some point crossed the line between useful data collection and into invasive profiling.
Additionally, the tactics for collecting ever more data grew increasingly questionable, stretching into private communications, using device speakers and cameras, and tracking user locations, even when the user wasn’t using their device. Advertising increasingly resembled surveillance with implications for both individuals’ most personal experiences and society’s fundamental organization.
The next generation of web-based data collection has to find solutions that factor in the consequences of current data-collection infrastructure.
The crux of the matter boils down to: If data collection can’t be based on personally identifiable traits such as age, gender, location, cross-site activity and user engagement history, etc. – then what should it be based on?
On-site user behavior is a compelling solution to restore balance between privacy concerns and more palatable use cases for collecting data.
On-site user behavior allows for site owners to optimize their funnels, content, and products based on aggregated, first-party data without knowing any personal attributes about the user. In other words, the user maintains full anonymity, but the data from their site sessions can be leveraged into better serving future audience needs.
Moreover, on-site user behavior reflects important considerations not necessarily accounted for in demographic or interest-based information.
A 28-year-old male gamer in Texas, who we’ll call User A, is a casual web surfer with a proclivity to make purchases on a whim. Another 28-year-old male gamer in Texas, who we’ll call User B, is more scrupulous. User B generally requires in-depth research and a good night's before making purchasing decisions.
User A has a tendency to scroll through pages fairly quickly and hits multiple site pages in their first visit to the site. User B stays on one page for longer and is less likely to click through as many pages as User A on their first visit. User A is more likely to make a purchase on the first visit than User B.
Under existing, cookie-based and third-party data collection infrastructure, these two users might be segmented together, because their personal attribute profiles overlap. But with behaviors-based analytics, they would wind up in different audience segments.
User A would be designated into an audience segment that is more likely to convert on their first visit. The website’s user experience could be optimized accordingly by recommending shorter articles, upselling related products, and focusing on a sale as a conversion for future users that demonstrate similar behavioral patterns as User A.
User B, on the other hand, could be delegated into an experience that would cater to their web-based behavior and longer cycle decision-making processes. For example, their user experience could include longer-form article recommendations and a conversion goal of getting an email sign up to maintain contact in the subsequent days where they might be more inclined to make a decision.
These segments can be generated without knowing whether either user is a 28-year-old male gamer in Texas. Instead, users are parsed into an audience segmentation composed of the site’s historical users whose on-site behavior correlated with theirs.
Neither User A nor User B is tracked across sites, yet the funnel can be optimized while privacy is preserved.
While the phase out of third-party data and cookies is expected to fundamentally change the nature of web-based advertising and commerce, there’s no need to throw out the baby with the bathwater.
The machine-learning technology that exists around current data-collection methods can be transposed onto behavior-based models to improve data accuracy and granularity. The key difference will be defining what behavior-based data points to collect, translating correlations into probabilities and predictions, and deepening data sets over time. It’s a similar process with a different starting point and opportunities to learn new things about audiences, particularly as it relates to predicting behavior and decision making.
Though the impending overhaul of cookies and use of third-party data is a shift that will inevitably cause disruption, it’s also an opportunity to rebuild an internet that is more user centric and respectful of privacy.
And while this fundamental shift in data collection will inevitably bring on a host of new challenges, it’s important to remember that all great innovation requires continuous iteration.