Skip to content Skip to sidebar Skip to footer

Data Privacy and AI: Where Are My Customers Coming From?

The folks over at Analytics Power Hour decided to bring me back for another episode. This time, we talked about how companies use cookies to determine where their customers come from (the technique is called, “attribution”)—and how to approach attribution in an environment where traditional tracking methods, like cookies, are becoming obsolete.

With the rise of privacy-focused browsers and company standards (i.e. Apple’s new privacy-first positioning), stricter data protection regulations, the digital marketing industry faces significant challenges in tracking and attributing user behavior.

The fading relevance of cookies and the stripping of referring site information are not just minor inconveniences. They represent a fundamental shift in how we must approach data collection and analysis.

One might wonder if server-side technical solutions could be the silver bullet for attribution in this new landscape. While these solutions offer a part of the answer, they are not the panacea for the broader challenge of multi-touch attribution. We need to think beyond just technical fixes and consider methodological innovations.

Interestingly, a solution proposed centuries ago by Jan Baptist van Helmont—randomized controlled trials—holds considerable promise for today’s attribution challenges. These trials, a cornerstone of experimental design, provide a robust framework for understanding causal relationships, a critical aspect often overlooked in the rush to attribute marketing success to various touchpoints.

In our podcast discussion, we explored how problem formulation and experimental design play crucial roles in effective attribution. Understanding the cost of data and how to draw causal inferences are fundamental to this process. It’s not just about collecting data; it’s about collecting the right data and interpreting it within the correct framework.

The industry stands at a crossroads where adopting rigorous experimental approaches to attribution could significantly enhance the accuracy and reliability of our insights. However, the challenge remains to shift the industry’s mindset and practices towards these methodologies.

As we move forward, it’s essential for analysts, marketers, and industry stakeholders to embrace these changes and the opportunities they present for more accurate, privacy-respecting attribution methods. The future of attribution is not just about adapting to new technologies but also about revisiting and applying time-tested scientific principles to our digital age challenges.

If you’re interested in digging into the implications of changes like these for your business, I encourage you to join one of my upcoming talks, or book an appointment with me.

Leave a comment