Differential Privacy Guide

Last updated: January 23, 2024

About the Differential Privacy Guide

The goal of privacy is to limit and minimize the collection and sharing  of personal data between business partners. However, some of the core functions of digital advertising such as ad targeting, measurement and attribution wholly rely on the ability to identify and share data collected at the individual user level. 

Different techniques and heuristics have been applied to protect identity of individuals and prevent identification of individuals by linking multiple data sets, for e.g. anonymization of identifiable attributes. But none of these provide a robust or rigorous guarantee of privacy. 

Differential Privacy, a rigorous mathematical definition of privacy has emerged as a leading technique to analyze and draw inferences from data sets in a way where one cannot determine if a particular individual was present in the data or not. As the use of differential privacy grows, it is necessary to understand:

  • What is Differential Privacy
  • How is it applied to ad tech use cases with deeper dive into attribution
  • Privacy vs. utility considerations
  • How is it different from other anonymization techniques
  • What are the considerations when using Differential Privacy in advertising 

IAB Tech Lab’s Privacy Enhancing Technologies (PETs) Working Group has developed an informational guide on Differential Privacy for decision makers, analysts and product developers working with advertisers, publishers and ad tech providers to demystify the technology, scope and application of Differential Privacy.

The Public Comment period ended on December 14, 2023. Updates are being made for finalization.

This is part of a series of guidance and specifications being developed by IAB Tech Lab PETs initiative to provide the industry with definitive solutions to applying and integrating Privacy Enhancing Technologies (PETs)  in digital advertising.