A lookalike audience is a targeting method used in paid advertising that finds new users who share similar characteristics with an existing group, such as your customers, website visitors, or social media engagers. Rather than defining an audience manually through interest or demographic targeting, the advertising platform analyses a source audience and identifies patterns it then uses to find similar people at scale.
How does a lookalike audience work?
On Meta platforms, you create a lookalike audience by selecting a source audience and choosing a similarity percentage between 1% and 10%. A 1% lookalike is the closest match to your source audience but the smallest in reach. A 10% lookalike covers a broader group with a weaker similarity to the original.
Meta analyses attributes including demographics, interests, behaviour, and device usage to find matching users in your chosen country or region. The platform updates the audience automatically as your source audience changes, so you do not need to rebuild it manually for each campaign.
What makes a strong source audience?
The quality of a lookalike audience depends almost entirely on the source audience provided. The most effective source audiences are:
- Customer lists of at least 1,000 contacts, ideally from buyers rather than all contacts on a mailing list
- Website visitors who completed a specific action, such as a purchase or enquiry form submission
- App users who reached a meaningful in-app event, such as a completed registration or purchase
- Video viewers or social page engagers who showed genuine interest rather than passive exposure
Where can you create lookalike audiences?
Lookalike audiences are primarily a Meta feature, available in Meta Ads Manager for Facebook and Instagram campaigns. LinkedIn offers a similar feature called Lookalike Audiences, built from matched audiences such as contact lists or website visitors, which tends to perform well for B2B campaigns. Google Ads previously offered Similar Audiences but deprecated the feature in August 2023, shifting advertisers towards automated targeting options such as Optimised Targeting. In Thailand, Meta lookalike audiences are widely used given the large Facebook and Instagram user base, and they are a standard part of campaign setup for most consumer-focused clients.
What are the limitations of lookalike audiences?
Lookalike audiences work best when your source audience contains clear signals of purchase or intent behaviour. A small or broadly defined source audience will produce unreliable results. Increasing the similarity percentage widens reach but reduces precision, which can lead to wasted spend on users with only a superficial resemblance to your best customers.
Privacy changes from iOS 14 onwards have reduced the accuracy of web-based source audiences by limiting the tracking data available to Meta, which means customer list-based audiences now tend to outperform pixel-based ones. Testing different source audiences and similarity percentages against actual campaign results is the most reliable way to find what works.