Similar Segments vs Affinity Audiences: What to Use Now
In short: Both were prospecting tools for reaching people beyond your own lists, which is why they still get compared - but only one of them exists. Similar segments, Google's lookalike mechanism that derived new audiences from a remarketing or Customer Match seed, was retired in 2023; affinity audiences remain live and selectable. The lookalike job did not disappear, it moved: optimized targeting and Smart Bidding now do that expansion invisibly, using your lists as signals rather than generating a named audience. If you want interest-based prospecting today, affinity is still on the menu; if you came looking for similar segments, what you actually want now is optimized targeting fed by fresh seed lists.
By the AdFlint research team · Fact-checked against current Google and Meta platform behavior · Last reviewed July 2026
Affinity Audiences
Reaches people with a durable long-term interest or lifestyle, such as cooking enthusiasts or luxury shoppers, rather than anyone showing immediate purchase intent.
Google builds affinity segments from sustained browsing and viewing habits, so they describe who someone is rather than what they are shopping for this week. You use them on Display, Video, and Demand Gen for upper-funnel reach when a product needs awareness before consideration. The mistake is expecting direct-response performance and then judging them on last-click conversions they were never going to win.
Full definitionSimilar Segments
Google's retired audience expansion feature, which generated lookalike-style segments from a seed list before being removed in 2023 in favor of automated targeting.
Similar segments once derived a resembling audience automatically from a remarketing or Customer Match seed, and Google sunset them so campaign-level automation could do that expansion instead. Your seed lists remain valuable as signals feeding optimized targeting and Smart Bidding rather than as sources for a derived list. Advertisers still hunting for the setting almost always want optimized targeting or Customer Match.
Full definitionSide by side.
The differences that actually change what happens in your account.
| Affinity Audiences | Similar Segments | |
|---|---|---|
| Status today | Live, selectable, and unchanged in concept. | Retired in 2023 - the setting no longer exists in the interface. |
| How the audience was formed | Google groups people by sustained interests from its own behavioral data. | Google derived a resembling audience from a seed remarketing or Customer Match list. |
| What it needed from you | Nothing beyond picking a segment off the menu. | A healthy, qualifying seed list - no seed, no similar segment. |
| Where the job lives now | Still its own feature, unchanged. | Inside optimized targeting on Display, Demand Gen, and Video, with your lists demoted from audience source to signal. |
| Reporting visibility | A named segment row you can judge on its own numbers. | Gone as a row - lookalike-style expansion now blends into automated delivery and cannot be isolated cleanly. |
| What your seed lists are worth | Nothing to this feature - affinity never used your data. | Still valuable: fresh Customer Match and remarketing lists sharpen the automation that replaced it. |
What actually separates them.
One is selectable today and one is a retired feature, so any playbook still telling you to add Similar Audiences is describing a Google Ads that no longer exists.
Similar segments mirrored your customers while affinity mirrors Google's interest taxonomy - resemblance to your seed versus membership in a category are different targeting logics.
The lookalike job moved from a visible, named segment into invisible expansion inside optimized targeting, so you can no longer report on similar-audience performance as its own line.
Similar segments were only as good as the seed feeding them, while affinity never needed any first-party data at all.
Your lists did not lose their value when the feature died - they now improve automated expansion and Smart Bidding as signals instead of generating a separate audience.
Which one should you use?
Use Affinity Audiences when
- You want interest-based prospecting you can set up today with no first-party data.
- You need a named, reportable segment whose performance you can read as its own row.
- Your seed lists are too small or too stale to be useful signals, so resemblance-based expansion has nothing good to learn from.
- You are building top-of-funnel reach for a lifestyle product where an interest bucket honestly describes the buyer.
Use Similar Segments when
- You are following a pre-2023 playbook or audit checklist that says to add Similar Audiences - the modern equivalent is optimized targeting seeded by your lists.
- Your remarketing pool is exhausted and you want new people who resemble it; that is now optimized targeting's job, not a segment you select.
- You inherited an old campaign that used similar segments and need to understand what took over its delivery after migration.
- You run Lookalike Audiences on Meta and expected a Google twin - the honest answer is that Google folded the concept into automation rather than keeping a named feature.
Common questions.
What happened to Similar Audiences in Google Ads?
Google retired them in 2023. The company shifted lookalike-style expansion into automated features - optimized targeting for Display, Demand Gen, and Video, and Smart Bidding with broader matching on Search - with your remarketing and Customer Match lists acting as input signals. The named segment type was removed rather than replaced one-for-one.
What should I use instead of similar segments?
Keep your seed lists fresh and let automation use them. Attach your best lists as signals, enable optimized targeting where it fits your goal, and make sure conversion tracking is solid, because the expansion optimizes toward whatever conversions you report. The resemblance logic still runs; you just steer it with inputs instead of selecting its output.
Does Meta still have lookalikes even though Google removed theirs?
Yes. Meta continues to offer Lookalike Audiences, though its Advantage+ automation increasingly performs the same expansion on its own. The platforms converged on the same idea from different directions: seed data in, automated resemblance out - Google simply stopped exposing the intermediate audience as a thing you pick.
Will similar segments still show up in my old reports?
Historical rows can still appear when you pull date ranges from when the feature was live, but nothing new serves against them. Treat those rows as archaeology - useful for understanding past performance, not something you can act on or rebuild today.
Or stop choosing between them.
AdFlint picks the setting, writes the ads, and keeps optimizing inside the Google and Meta accounts you already own.
Related comparisons
- Affinity Audiences vs In-Market Audiences
- Custom Segments vs In-Market Audiences
- Customer Match vs In-Market Audiences
- In-Market Audiences vs Remarketing Lists
- In-Market Audiences vs Similar Segments
- Demographic Targeting vs In-Market Audiences
- In-Market Audiences vs Life Events Targeting
- In-Market Audiences vs Optimized Targeting