How accurate and precise data sets drive effective first-party data activation

Sponsored by Adstra • April 24, 2024 • 3 min read •

Emily Kistner, director of new business and identity solutions, Adstra

The landscape of data-driven advertising is evolving at a breakneck pace, with connected TV and programmatic advertising leading the charge.

Scalable identity, coupled with the adoption of machine learning, has unlocked marketers’ ability to reach increasingly specific audiences. These improvements, along with the massive COVID-era shift in eyeballs and subsequent investment dollars, have rebranded CTV as a performance channel.

At the heart of this transformation lies the critical role of seed data sets. A seed audience typically consists of existing customers’ or authenticated visitors’ data. It serves as the sample for identifying common attributes among a given audience and enabling the lookalike modeling used for targeted ad campaigns and syndicated audiences. These data sets are the foundation upon which machine learning algorithms predict, learn and scale.

However, in a world of deprecating signals, the accuracy of seed data sets is increasingly critical — yet it’s often overlooked in a world of buzzier concepts and news cycles such as AI.

Without precise data, marketers risk going off course with targeted advertising

The 1-in-60 rule is an aviation navigation principle relevant to various analytical and planning scenarios, including targeted advertising. A mere one-degree deviation at takeoff takes a plane one mile off course for every 60 miles traveled. A seemingly inconsequential variance at the outset can result in an aircraft miles away from its intended destination by the end of the journey.

In data-driven advertising, the 1-in-60 rule underscores the importance of precision in targeting strategies and data sets used for measurement. This requires accurate, declared and consented consumer data with high-fidelity matching and validation between online and offline data sets.

The richest authenticated consumer data has historically persisted almost exclusively across walled gardens, but it doesn’t have to stay this way.

Publishers can access strong first-party data, such as authenticated identity or viewership behavior across devices; brands receive rich first-party data sets through a direct consumer relationship and value exchange, giving insight into identity, behavior and preferences.

Publishers, brands and their agencies must take a more active role in collecting and using this first-party data to be best positioned in the era of performance marketing — but accuracy is critical.

High-fidelity data resolution is unlocking improved campaign reach and performance

Many vendors in the data ecosystem, like customer data platforms and onboarders, have incentive structures built on scale instead of accuracy. A CDP may charge for the total number of identifiers matched, regardless of the accuracy of the match. An onboarder may use fuzzy logic or shared identifiers like email to increase a match rate. However, advertisers can’t run high-performing campaigns with poor accuracy.

Marketers need visibility and understanding of identity resolution services. Various use cases demand different balances between scale and accuracy. Many black box solutions in the market don’t afford brands this control. The real differentiators to these solutions are often the underlying graphs mapping individuals to households (handling examples such as individual- and household-level moves) and the digital IDs associated with those individuals (like cookies, MAIDs and HEMs).

In digital, significant universes of false digital IDs often enter the equation due to online visitors not using a legitimate email when logging in to view content and other factors. Or, in CTV, a shared identifier like an email used to log into a streaming app adds scale without necessarily adding precision to link individual-level demographics. And if marketers use this fuzzy first-party data as a seed audience for a lookalike model, they’re only compounding the problem.

Just like a slight navigational error can significantly mislead an aircraft, leveraging low-fidelity seed data in performance marketing can lead to substantial inefficiencies, impacting the reach and effectiveness of advertising efforts. Focusing on first-party data management and accurate identity resolution ensures that marketers stay on course.