This article summarizes Google’s Q1 2023 interest-based advertising experiment. It describes a limited test of alternative audience signals, not a forecast of every advertiser’s results or a current Chrome implementation guide.
Table of Contents
At the time of the 2023 experiment, Google was exploring interest-based audience solutions in anticipation of changes to third-party cookies. The test covered affinity, in-market, custom and demographic audiences on its display inventory, using combinations of privacy-preserving signals.
Google’s Experiment
In Q1 2023, Google’s ads platforms conducted an experiment to analyze the performance of interest-based audience (IBA) solutions when reliant on privacy-preserving signals. These signals included contextual information, the Topics API from the Privacy Sandbox, and first-party identifiers such as Publisher Provided IDs. The research assessed a wide range of signals available in a privacy-first environment. Rather than comparing third-party cookies to the Topics API alone. The experiment involved two arms on a portion of Chrome traffic. One arm maintained the use of third-party cookies, while the other replaced them with privacy-preserving signals for IBA solutions, keeping third-party cookies for purposes like measurement and remarketing. 
Source
The results
The findings revealed that IBA solutions with privacy-preserving signals show promise in comparison to third-party cookies. Google Display Ads advertiser spending on IBA, as a proxy for scale reached, decreased by 2-7% compared to third-party-cookie-based results. Conversions per dollar, representing a return on investment, decreased by 1-3%. Click-through rates remained within 90% of the status quo, and similar performance was observed for Display & Video 360.
The results also highlighted the positive impact of AI-powered optimization solutions on campaign performance. Campaigns utilizing optimized targeting or Maximize conversions bid strategies were less affected by the removal of third-party cookies, suggesting that machine learning can play a crucial role in driving results. Although the findings are encouraging, they should not be regarded as a definitive indicator of Google’s IBA performance post-third-party cookie deprecation. The current experiment evaluates the effectiveness of serving interest-based audiences, but results may vary in future experiments considering measurement, remarketing, and other use cases.
In a nutshell
Google Ads tests privacy-focused interest-based advertising and shows promise with reduced ad spend and minimal impact on conversion rates. Google’s interest-based audience solutions experiment, which uses privacy-preserving signals, demonstrated mixed results compared to third-party cookies.
For an advertiser, the next step is to test measurement and campaign outcomes in the browsers and consent conditions your customers actually use. Separate audience delivery from conversion reporting, and do not read this one experiment as proof of all targeting strategies. Our Google App campaigns resource provides channel context, while your own experiments should decide whether a signal improves valuable acquisition.
