Two mountain goats with curved horns standing on a snowy rocky ledge overlooking a fog-covered valley at sunset with snow-capped peaks in the background.

Lookalike targeting

Lookalike targeting finds new audience members who resemble an existing high-value segment, using matched signals inside a data clean room instead of a shared customer list.

A partner's data is compared against the seed audience inside a confidential computing environment, and only the resulting lookalike segment is returned as output.

Contact
Abstract geometric design with circles and arcs representing secure data collaboration by Decentriq.
Abstract black and white geometric patterns representing secure data collaboration solutions by Decentriq.

How it works

The process starts with a seed audience, typically an advertiser's highest-value or most engaged customers. That audience's characteristics are matched against a much larger pool of potential customers inside the clean room containing both user identifiers and user attributes. A lookalike model is trained entirely within the secure environment. The output is a new audience segment sized to the advertiser's targeting needs, ready for activation across the partner's inventory. For a deeper look at the modeling technique, see our guide to lookalike modeling. For the broader cookieless targeting context, see how to future-proof your lookalike audience strategy.

See it in action: Lookalike targeting in a Decentriq data clean room

AI-powered lookalike audiences on the Decentriq platform, walking through campaign reach expansion using matched, privacy-safe lookalikes.

Case Study: New York Times Advertising

+61%

international CTR lift

+25%

performance uplift in U.S.

New York Times Advertising (NYTA) and a major financial institution wanted to reach a highly specific audience, households with $1M+ in investable assets in the US and $100k+ internationally, without relying on third-party cookies. Working inside a data clean room, the two built lookalike audiences from first-party data at two levels of precision: a narrow 1% lookalike and a broader 5% lookalike, both matched against the bank's own criteria.

The most precise segment, the 1% lookalike, delivered a 25.12% uplift over the control group in the US, while the international campaign saw a 61% lift in click-through rate.

REad Full Case Study
Close-up of The New York Times newspaper with date Saturday, October 12, 2024, featuring a headline that reads 'The War in Darkness' and a blurred image of two people embracing.

Frequently asked questions

How is lookalike targeting different from remarketing or affinity segments?

Remarketing re-engages an advertiser's own existing customers, and affinity segments define a new audience from the overlap between two parties' data. Lookalike targeting instead extends reach to new prospects who share characteristics with an existing high-value audience.

Does the seed audience ever leave the clean room?

No. The seed audience and the partner's underlying data stay in their original environment. Only the resulting lookalike segment is returned as output.

What data is used to build a lookalike audience?

The clean room compares an advertiser's existing customer data (seed audience) against a publisher or partner's larger audience pool, using patterns in behavior or characteristics to identify similar prospects.

What kind of results can lookalike targeting achieve?

In addition to the case study outlined above, a major Swiss bank reduced its cost per page view by 44% and cost per qualified visit by 31% using tailored lookalike audiences built this way.

View Other Targeting use-cases

ID graph onboarding

Bring an identity graph into the clean room for matching and activation without sharing PII with the identity provider

explore ID graph onboarding ->

Affinity segments targeting

Match first-party signals to uncover top affinity segments to target

explore Affinity segments targeting ->

Remarketing

Build retargeting pools from matched audiences without sharing full raw customer data with the other party

explore Remarketing ->

Exclusion targeting

Suppress existing customers from a campaign without handing over the full customer list

explore exclusion targeting ->
Person in low light with colorful screen reflection; conveys focus and modern technology in data collaboration.

Talk to us about lookalike targeting

Contact