Is First-Party Data Collateral Damage For Stricter Third-Party Privacy Policies?

"Brand Aware” explores the data-driven digital ad ecosystem from the marketer's point of view.

Today's column is written by Sachin Puri, senior director and global head of performance marketing at StubHub.

The cookie, the technology that powers the internet and digital advertising, has never faced such serious risk. The burgeoning concerns about user privacy and online fraud, such as cookie hacking, have led to stringent laws, strict platform policies and a rise in users preventing cookie tracking.

The loss of ability to retarget and measure the impact is one of the biggest challenges for performance marketers. And with these fast-changing developments, much of the real impact still remains unclear.

As we understand the impact and adjust the operation, it’s vital to recognize that recent policy changes could seriously slow down cross-domain tracking, including site remarketing. In my opinion, this balancing of user retargeting and privacy is one of our most important responsibilities in the marketing ecosystem.

Recent tweets by Apple indicate that first-party tracking would likely be collateral damage to the emerging closed-tracking system, and other developments will likely jeopardize the measurement and ROI tracking on first-party domains. The marketing community should recognize this as a wake-up call and opportunity for key players to collaborate on the solution.

This is not just about one platform or browser, but about the whole marketing ecosystem built around delivering a delightful user experience.

Cookies and browser tracking or prevention

While many first- and third-party vendors were working on prospective solutions for Intelligent Tracking Protection (ITP) 2.1, Apple released ITP 2.2, another strict update that limits first-party persistent cookie tracking to one day (vs. seven days two months ago) for domains classified with cross-site tracking capabilities, such as Google and Facebook.

For marketers with purchase cycles longer than one day, this means channel attribution will likely not be accurate beyond one day, limiting their ability to truly measure the impact of marketing campaigns.

It’s evident that the solutions to stop cross-domain tracking to protect user privacy are moving digital toward a closed ecosystem for legitimate first-party tracking, which is an undesirable outcome for the marketing community.

Intermediate impacts: measurement and analytics

Web analytics or attribution tools, such as Adobe Omniture and Google Analytics rely heavily on first-party cookies to measure and analyze the full funnel once the user visits the domain. The full funnel analysis starts from visits but continues beyond the checkout to product usage, customer support or repurchase.

To effectively grow active users, analyzing the full funnel by user cohorts, such as new vs. repeat visitors, is critical. Whether you are a startup or a big enterprise, it’s key to know if marketing campaigns are attracting enough new users to grow a brand, or if they are just driving revisits by existing users. Are those visits incremental reach or just increased frequency of the same users?

Not only is this reach-frequency tradeoff core to any media plan, but to develop a data-informed marketing investment strategy, knowing the average number of visits, the order value in dollars and attributed marketing channel by user is necessary. This information is key to forecast, measure and deliver optimal customer acquisition costs and ROI.

Under ITP 2.2, if a user visits a site via display ad on day one and later revisits by clicking on a social post on day nine, he or she will be counted as a new visitor in both visits. Thus, the number of visits and average order value per visitor will be incorrect, and the order will be attributed via last-click attribution to social, even if the first visit originated from a display ad. The data would then indicate display ads are not performing well, leading to incorrect spend allocation to display and social, which may or may not be the case.

Merchandising and personalization

Many advanced marketers, especially in ecommerce, have developed machine-learning based algorithms to merchandise products and personalize customer journeys based on browsing history stored in first-party cookies. For example, a brand may offer a product promotion or pricing incentive via an onsite banner ad to improve conversions or lifetime value.

In my opinion, as the number of SKUs and visitors increase, algorithmic user journey management is not just good-to-have but a necessity to grow the business at scale.

Under ITP 2.2, in the same scenario outlined when a user visits a site twice, seven days apart, the merchandising algorithms would not be able to run personalization effectively because cookie-based day-one visit history would be lost on Safari. Any A/B test on this pricing promotion would have a high degree of noise as the same user can fall under different control and exposed cookie pools in his or her two visits. Thus, not only will the test results be contaminated, it will also be a suboptimal experience for the user.

What can we do now?

The goal to safeguard user privacy can’t be solved in silos and with unilateral view; it requires a thoughtful discussion and collaboration across government agencies, industry bodies, such as IAB, platforms, browsers, advertisers, agencies and ad tech companies.

The cost of unilaterally cannibalizing legitimate first-party cookies will deliver a bad user experience, which is neither the goal nor a desired outcome.

Let’s collaborate as we protect user privacy and solve for solutions that serve both marketers and the user.

For now, we can get to work by paying close attention to measurement reports and inquire with analytics teams about how GDPR and ITP are impacting metrics, such as channel attribution or unique visitors.

Follow Sachin Puri (@spuri79), StubHub (@StubHub) and AdExchanger (@adexchanger) on Twitter.

 

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>