Transforming Trust in Advertising: How Privacy Engineering is Changing the Game

Yes, through AdTech solutions that prioritize privacy and create value industry wide, empowering advertisers and publishers to grow using their first-party data.

Our Lead Privacy Engineer, Goran Stipcich, explains how advanced Privacy Engineering, including techniques like federated learning, can unlock valuable insights for audience targeting without sharing user data. This approach empowers brands and publishers to maintain their competitive edge with first-party data while fostering customer trust. The future of digital advertising hinges on getting this balance right, and we believe unlocking this future lies in privacy engineering.

The privacy shift

The shift towards privacy over the last decade has been dramatic. Data protection regulations like GDPR have been brought in to address public concern over the volume of data being harvested and misuse of sensitive data.¹ They demand users have control over how brands use their personal information. At the same time, key players Apple and Firefox have made privacy-first a core selling point ², limiting advertisers’ access to user data and forcing a shift in targeting strategies. Consumers, increasingly aware of data misuse, are actively limiting data sharing, disabling cookies or using ad-blockers. So, while the volume of data people are creating continues to grow, traditional methods of raw data sharing, uniformed data tracking and indiscriminate data collection are no longer viable.³ This new reality will require advertisers to fundamentally change how they reach their audiences. Innovative targeting solutions are emerging, yet the challenge lies in how we can avoid moving or sharing any raw data, without impacting performance.

Privacy Engineering: Building trust through technology

This is where Privacy Engineering comes in.

Privacy by Design and Privacy Enhancing Technologies (PETs).⁴

Privacy by Design is a design philosophy used to build privacy directly into technology, ensuring data protection is fundamental, not an afterthought.

Privacy Engineering puts this philosophy into practice through a set of development practices applied throughout the entire product lifecycle. This includes requirements like purpose specification, data minimization (collecting only essential data) and anonymization techniques.⁵ Already tried and tested in finance and healthcare⁶, at Resolve we’re using these techniques to extract insights and value without ever sharing any personal data, with the ultimate goal of re-building the ad tech landscape into a trusted and collaborative ecosystem:

• Federated learning: Allows multiple parties to collaboratively train machine learning models without sharing any sensitive user data.
• Secure multi-party computation: Facilitates collaborative analysis of data owned by different parties, while maintaining each party’s local inputs privately. It guarantees that only the previously agreed upon outcome can be learned – nothing more – and uses advanced encrypting techniques to guarantee a high level of security. This is used alongside federated learning and works for complex use cases such as measurement and analytics.
• Differential privacy: This mathematical framework allows us to protect the privacy of individual records in a dataset while keeping the general utility of the data for analysis, by adding carefully chosen noise (additional meaningless data). Differential privacy is also complementary to both federated learning and secure multi-party computation and we layer them to increase privacy.
• Threat Modeling: A proactive approach for identifying potential privacy threats in a software tool or system that works by focusing on the data flows involved.⁷ Each privacy threat is assigned a priority proportional to its risk, and a suitable fix using PETs is proposed and implemented. At Resolve, each product undergoes this analysis until no significant risks remain.

Together, these privacy-enhancing technologies offer achievable and robust technical solutions to protect sensitive data without limiting scalability or growth opportunities. This “hard privacy” approach can provide a sturdy foundation for digital advertising that could revitalize the entire ecosystem.

Hard privacy: Removing the need for blind trust

Resolve’s mission goes beyond navigating the data privacy shift to create a new ecosystem built on privacy and trust, championing a “hard privacy” approach that we believe will futureproof the industry’s success.

While many solutions only offer “soft privacy” – using methods where data is shared with third parties relying on privacy policies and offering opt-out options – they fall short. Advertisers and publishers are still forced to blindly trust that third parties centralizing the data are handling data responsibly and complying with regulations.

Hard privacy, on the other hand, has trust built in. It actively protects privacy by never sharing raw data in the first place. Instead of data constantly being moved to external models for analysis, as it is today, Resolve brings its AI models to the data. Resolve then utilizes PETs to guarantee that only specific insights are computed and shared, allowing advertisers and publishers to collaborate without risk. Our transparent approach decentralizes control, removing the need for blind trust between partners. This safeguards everyone and unlocks new value: advertisers can target efficiently, and publishers preserve their competitive advantage – extracting valuable audience insights without compromising the unique value of their individual data sets – and can support quality journalism.

Imagine an ecosystem where this is the new norm.

If the Ad tech industry can fully embrace privacy Engineering and hard privacy principles, then together we can build a more ethical and sustainable AdTech ecosystem that benefits everyone – users, advertisers, publishers and even the quality of advertising itself. At Resolve, we are building the foundations to support this reality. A reality that publishers and advertisers can be a part of today.

Find out more about our technology: Our Products