Rascasse› Glossary› Lookalike Audience

Audience Intelligence Glossary

What Is a Lookalike Audience?

A lookalike audience is a group selected because it resembles an existing source audience on characteristics believed to predict the same behavior. It is the standard mechanism for expansion — and it inherits every flaw in the audience it was built from.

How similarity is determined

The mechanism differs by system, and the difference matters more than the label suggests.

  • Platform lookalikes build similarity from the platform's own signals, inside the platform's inventory. Effective for activation, invisible in method, and unusable outside that platform.
  • Behavioral lookalikes identify audiences with similar affinity and interest structures across the open market, independent of any single platform's inventory.
  • Demographic lookalikes match on age, gender, income and region. The weakest form, because demographic similarity predicts behavior poorly.

The seed decides the outcome

Every lookalike inherits the biases of its source. This is the single largest source of disappointing expansion campaigns, and it is structural rather than technical.

  • A seed of existing customers produces more people like the ones already won — which is the definition of the group with the least headroom left.
  • A seed built from one channel reproduces that channel's skew. A newsletter list is not the audience; it is the part of the audience that subscribes to newsletters.
  • A seed that is too small yields similarity fitted to noise. Platform documentation generally recommends a minimum in the low thousands.
  • A seed from one market does not transfer. Expanding into a new country from a domestic seed assumes the audience has the same structure there, which is usually the assumption that was worth testing.

A more productive seed is often the audience of a property your customers overindex on, rather than the customers themselves — it points outward, toward people not yet reached, instead of inward.

Lookalike and overlap are opposites

A good lookalike has high affinity similarity and low overlap with the source: it behaves like your audience and consists of different people. High overlap means the expansion is mostly re-reaching the group already addressed, which will look efficient on platform metrics and produce no growth.

Where audience data helps

Rascasse is used upstream of the activation platform: identifying which entities, properties and interest groups an audience overindexes on, so the seed can be chosen deliberately rather than defaulting to the customer list. The resulting audiences are then built inside the activation platform, where the targeting actually runs.

Related terms