Similar Audiences in Google Ads were user segments generated automatically by Google’s advertising platform from a customer list provided by the advertiser. Instead of simply showing ads to the people already on the list, Google used machine learning to find new users with characteristics, browsing behavior and purchase patterns similar to those of the source audience. This allowed a campaign to expand its reach without sacrificing the relevance of the advertising message. Google retired Similar Audiences in 2023: they are no longer created or used for targeting, and have been replaced by automated signals such as optimized targeting and audience segments in Performance Max and Demand Gen campaigns.
How it worked was based on comparing the data provided by the advertiser with signals collected across Google’s network, such as searches performed, sites visited and apps used. The system never shared the identity of the people in the original list, but created an anonymous statistical model. Similar Audiences were therefore an intermediate tool between keyword targeting and interest targeting, and were used mainly in search, display, video and shopping campaigns.
What a similar audience identified
A similar audience (similar audience or similar segment in Google’s documentation), was a group of users that Google considered akin to a reference audience already known to the advertiser. The reference audience could consist of customers who had made a purchase, visitors who had filled out a form, newsletter subscribers or users who had installed an app. From this data, the platform built a predictive model and identified other users with a similar likelihood of taking the same action.
The logic was that of so-called lookalike targeting, already widespread on other advertising platforms. The main difference lay in the quantity and variety of signals available to Google, which ranged from search engine queries to videos watched, from maps consulted to apps downloaded. The higher the quality of the starting list, the greater the precision of the similar audience generated.
How it differed from affinity audiences
Affinity audiences group users based on stated or inferred interests, such as sports, technology or cooking. Similar Audiences, by contrast, were not based on predefined thematic categories, but on behavioral similarity to a specific advertiser list. The first type is broader and more generic; the second was more targeted and personalized.
What data was needed to generate it
To create a similar audience, you needed a source audience with a sufficient minimum size. Google generally required a few hundred or thousand active users, depending on the type of list and the country of reference. Lists that were too small or not recent enough did not allow the system to build a reliable model, and the similar audience might not be generated.
The source data could come from different sources: manually uploaded lists, remarketing audiences based on tracking tags, customer match audiences via email addresses or phone numbers, or lists from audience management platforms. In all cases, it was necessary to comply with data protection regulations and obtain consent from the users concerned.
Why the quality of the starting list mattered
A list made up of actual customers, with recent and high-value purchases, generally produced more effective similar audiences. Conversely, outdated lists or lists containing contacts not genuinely interested in the product tended to generate irrelevant segments and reduce campaign performance.
The variants that were available and the data they were based on
Google Ads offered several variants of similar audience, differentiated by data source and intended use. The choice depended on the campaign objectives and the advertiser’s industry.
| Type of similar audience | Data source | Typical use | Level of precision |
|---|---|---|---|
| Similar to customers | Customer list uploaded by the advertiser | Acquisition and sales campaigns | High |
| Similar to site visitors | Tracking tag on the website | Extended remarketing | Medium-high |
| Similar to app users | App usage data | Install promotion | Medium-high |
| Similar to YouTube subscribers | Interactions with the video channel | Video and awareness campaigns | Medium |
| Similar to qualified leads | Completed forms or sales contacts | Lead generation | High |
Opportunities and precautions for use
The main advantage was expanding reach without losing relevance. A similar audience allowed you to reach users who would hardly have been intercepted with keywords or generic interests, while maintaining a good probability of conversion. In addition, automatic optimization reduced manual segmentation work.
However, there were also significant limitations. The precision of the model depended on the quality and size of the source data. In niche industries or very small geographic markets, the similar audience could be insufficient or not meaningful. Moreover, changes to privacy policies and the reduction of third-party cookies progressively reduced the availability of signals, making the use of first-party proprietary data more important.
What changed with privacy regulations
European regulations and the choices of major browsers limited data collection across different sites. As a result, similar audiences based on lists provided directly by the advertiser, such as customer lists, became more relevant than those built on third-party tracking.
What is used today instead of Similar Audiences
With the retirement of Similar Audiences in 2023, Google Ads no longer offers the ability to create them or define their size. In their place, the platform uses automated signals such as optimized targeting and audience segments in Performance Max and Demand Gen campaigns, which leverage available data and signals to identify users most likely to convert.
Frequently asked questions about Similar Audiences
What is the difference between a similar audience and remarketing?
Remarketing shows ads to users who have already interacted with the site or app, while a similar audience reached new people who had not yet had direct contact with the brand but had similar characteristics.
Do Similar Audiences still exist in Google Ads?
No. Google retired Similar Audiences in 2023: they are no longer created or used for targeting. Today, automated signals such as optimized targeting and audience segments in Performance Max and Demand Gen campaigns are used.
Did Similar Audiences work for small businesses too?
Yes, provided the business had a sufficient data base. For organizations with few contacts, it could be useful to start from site visitor or newsletter subscriber audiences.
Could Similar Audiences be used together with keywords?
Yes, it was possible to combine the two targeting criteria. This further narrowed the audience and favored users who met both conditions.
Did Similar Audiences violate user privacy?
No, if used correctly. Google did not share the personal data of the source list and created anonymous models. It was still necessary to comply with current regulations and collect appropriate consents.
How long did it take to see results?
Timing varied based on industry, budget and the quality of the starting audience. Generally, a few weeks of data collection were needed before evaluating the effectiveness of the segment.
Facing this problem?Want to understand how to reach the right people on Google Ads today after the retirement of Similar Audiences?With the retirement of Similar Audiences in 2023, relying on poorly configured automated signals risks wasting budget on users unlikely to buy. I analyze your data and set up campaigns to maximize conversions and ROI.Discover Google Ads Management