What actually are ... lookalikes?

Lookalike audiences are used to reach new target groups or to maximize the retargeting potential at a meaningful frequency. The creation of lookalikes is therefore primarily about enlarging an existing audience with the aim of generating more reach for the advertiser. Starting from a basic audience that has been identified through online behavior, this target group is expanded by adding additional audiences that have similar characteristics. Depending on how high the similarity must be, the reach can be larger or smaller. In German-speaking countries, the term "statistical twins" or "user twins" is also used in connection with lookalikes.

Important to note: Prediction is not the same as lookalike, but describes a method based on statistical procedures and algorithms to predict future behavior or characteristics. With the help of machine learning, the audience is continuously updated and optimized.

Lookalike audiences are mainly used to pursue certain campaign goals: Targeting people who are similar, such as people who have downloaded a coupon, liked certain fan pages or engaged in shopping behavior. Lookalikes are based on large amounts of data and information and are created accordingly via data management platforms. Modern data management platforms offer the possibility of creating lookalikes based on individual target groups.

Lookalike audiences have gained new popularity, particularly through Facebook, where Facebook, as a profile-based data management platform, offers the possibility of creating lookalikes on profiles and expanding the campaign target group within the Facebook universe.

On Facebook, lookalike audiences are based on users who may be interested in a company, as they are similar to users who have already shown interest. The basis for lookalike audiences is the so-called source audience. When a lookalike audience is created, a source audience (at least 100 people from a specific country) is selected - a custom audience that was not created with a data partner; pixel data, mobile app data or fans of the page. The similarities between the people are determined, such as demographic data or interests. People are then searched for who are similar to those in the selected country.

Term explained by: Stefan Aichner, Head of Sales at Booost.Digital.