Segmenting Customers Without PII: The Next Evolution of Marketing Analytics

Digital advertising spend is expected to exceed $700 billion by 2025. However, with the growing privacy concerns and strict regulations, the traditional ways of advertising are facing new challenges. For decades, advertisers have relied on collecting personal data from users to target ads effectively. But there have been many changes with privacy regulations globally which limit the use of user level data. Working with Meta, one of the largest advertising platforms in the world, Varun Chivukula is leading the charge in this area.

"As regulatory changes (GDPR, CCPA) and ecosystem changes (Apple ATT, Cookie deprecation from Google) make it hard and unsustainable for advertising platforms to ingest user level personal identifiable information (PII), there was a need for an alternative which doesn't require exchange of user level data," he shared.

This is where privacy-enhancing technologies (PETs) come in. These new technologies, including methods like multiparty computation (MPC) and federated learning, let advertisers and platforms keep working without needing to exchange personal data. They make it possible to measure ad performance and target the right audience, while still respecting user privacy. Across the industry, companies are starting to adopt PETs not only to stay compliant, but also to future-proof their systems. With cookies and device IDs on the way out, technologies that protect user identity while still allowing useful insights are becoming essential.

Chivukula played a key role in the first industry-wide tests of these privacy-focused technologies, showing how they can be used at scale. His work has been crucial in proving that it's possible to run ad tests and optimize campaigns without exchanging sensitive user data. Additionally, he has assisted in building systems that calculate the effectiveness of digital advertising without needing to know individual user details. For example, he developed a method to calculate match rates, which is a way of measuring how well two sets of data align when working with privacy-enhancing technologies. This was a major advancement because understanding match rates is essential to knowing how well advertising campaigns are performing.

Discussing his work, he mentioned, "This for the first time provided advertisers and ad platforms a way to evaluate the representativeness of their computation and improve it. This, for example, enabled Meta's largest advertiser to understand that their true match rate was 10% vs. an expected 60%; this resulted in work to expand representativeness which improved the eventual match rate to 70%." This change alone helped his organisation generate an additional $500 million in revenue, showing just how impactful these new privacy methods can be.

Industry-wide, there's growing interest in how match rates and representativeness affect the accuracy of marketing outcomes. As advertisers adopt PETs, they also need better ways to understand how much signal is being lost and how to improve it-without compromising privacy.

Emphasising on adapting the new approach to advertising, the professional has authored papers including, "Use of Multiparty Computation for Measurement of Ad Performance without Exchange of Personally Identifiable Information (PII)" and "Impact of Match Rates on Cost Basis Metrics in Privacy-Preserving Digital Advertising" among several others.

Along with his team, he has also worked with Amazon to create privacy-enhancing products like AWS cleanrooms. These tools allow companies to handle data securely while still being able to use advanced analytics and machine learning for better decision-making. Cleanrooms and similar secure environments are becoming a core part of ad tech. They allow companies to collaborate on insights and targeting without exposing raw data. More platforms are now building such features directly into their advertising ecosystems.

Talking about the different side of the coin, the challenges in the world of privacy-enhancing technologies are worth mentioning. For instance, understanding how accurate the data is when two parties are working with it. Before the professional's contributions, it was difficult to assess and improve the match rates that are critical for ad platforms to know if their measurements are correct. By developing a way to evaluate and improve these match rates, Chivukula helped solve a key issue that had been holding back the industry. His work has been crucial in showing that privacy and effective digital advertising don't have to be at odds. As privacy laws continue to change and tighten, these new technologies can be the key to ensuring that advertisers can still run successful campaigns without violating user privacy.

The broader trend in the industry is clear: the future of advertising depends on maintaining user trust. With more consumers aware of how their data is used, advertisers who adopt privacy-first approaches are more likely to succeed in the long term. To put these insights together, it's not surprising to learn that privacy-enhancing technologies will be a major part of the future of digital advertising-like Google Search Ads, Ads on Facebook etc. These tools can help the industry grow in a way that balances user privacy with the need for personalized ads. With AI-driven solutions and better cybersecurity practices, the industry can continue to innovate while respecting users' right to privacy.

It's also worth noting that advertising can still be effective and profitable, without relying on personal data that users aren't comfortable sharing. For the digital advertising world, these new technologies might just be the solution to the challenges it's facing today.

About Varun Chivukula

Varun Venkatesh Chivukula is a data and marketing analytics expert with over a decade of experience working with leading tech companies like Dropbox, Facebook, Microsoft, and Box. He currently leads the Marketing Data Science team at Dropbox, helping improve business performance through data insights. Before this, he worked at Facebook, where he managed large enterprise accounts, developed tools to measure marketing performance, and worked on privacy-focused solutions to adapt to industry changes.

Chivukula holds an MBA from UC Berkeley's Haas School of Business and an engineering degree from BITS Pilani. He is skilled in tools like SQL, Python, and R, and has used data science to support decision-making, improve marketing results, and build better customer experiences. His background also includes consulting at McKinsey, where he advised senior leaders on marketing strategy and analytics. This blend of technical expertise and business understanding has made him a trusted leader in using data to drive smarter marketing decisions and long-term growth.