A lookalike audience is a targeting tool that finds new potential customers who share similar interests and behaviors with your existing customers. By analyzing your current audience, advertising platforms identify common traits to reach highly-qualified prospects who were previously difficult to find. This approach helps you scale your reach while maintaining the relevance of your ads.
What is a Lookalike Audience?
Digital advertising platforms use lookalike audiences to expand your reach beyond known customer lists. The tool anatomizes existing user profiles to find commonalities and then searches the broader platform for people with matching characteristics.
Facebook first initiated this feature in 2013, and other major platforms like Google Ads, LinkedIn Ads, Outbrain, and TikTok have since adopted similar models. These audiences help you enter new markets or segments by finding people who act like your most valuable users.
Why Lookalike Audience matters
Using lookalike audiences shifts targeting from manual guesswork to algorithmic precision. This offers several practical benefits:
- Higher efficiency: Identifying qualified leads becomes faster, which frequently lowers the cost of acquiring new customers.
- Scale: You can expand into different countries or new audience segments that mirror your current successful demographics.
- Automated discovery: Platforms find prospects that you might not have identified through traditional interest or demographic targeting.
- Performance improvement: Algorithms study custom audience attributes to look for users with the highest potential to improve ad performance.
How Lookalike Audience works
The process relies on a "seed" audience, which is a group of users you already know. The platform uses this seed as a blueprint to build the larger lookalike group.
- Select a seed audience: You choose a source group, such as your email list, website visitors, or people who have completed a purchase.
- Choose a location: You define the specific country or region where you want the platform to find similar users.
- Set the audience size: You determine how closely the new audience should match your seed.
On Facebook, size is customized via percentiles from 1% to 10% of the total population. A 1% audience is the most similar to your seed but the smallest in size. As you move toward 10%, the reach increases, but the similarity to your original customers decreases. TikTok uses three distinct tiers: Narrow, Balanced, and Broad to provide similar flexibility.
Types of Seed Audiences
The quality of your lookalike depends on the data source you provide. Common seed types include:
- CRM-based: Created from uploaded email or phone lists. You can segment these further by targeting only your highest lifetime value customers.
- Conversion-based: Built from users who performed a specific action, such as submitting a lead form or making a purchase on your site.
- Engagement-based: Based on how users interact with your content, such as those who watched a certain percentage of a video or spent a specific amount of time on your website.
Best practices
To get the most out of these audiences, focus on the quality and consistency of your data.
- Prioritize homogeneity: Ensure your seed audience shares a specific, consistent behavior. A small, highly similar group is often more effective than a large, diverse one.
- Use the right sample size: While Facebook allows a minimum seed of 100 users from one country, they generally recommend a seed of 1,000 to 5,000 users.
- Update your seeds: Regularly refresh your source lists to ensure the algorithm is targeting based on recent customer behavior.
- Balance size and similarity: Start with a narrow 1% audience for high conversion and move toward broader percentages only when you need more volume.
Common mistakes
Mistake: Using a seed audience that is too small. Fix: Ensure you meet platform minimums. For instance, TikTok requires a minimum source audience size of 1,000.
Mistake: Including "outliers" in your seed data. Fix: Clean your lists to remove one-time customers who don't fit your target profile so the algorithm doesn't learn from the wrong signals.
Mistake: Ignoring regulatory restrictions. Fix: Be aware that platforms have restricted targeting for sensitive sectors. Since 2019, Facebook does not allow age, gender, or ZIP code targeting for housing, employment, or credit ads to prevent discrimination.
Examples
Example scenario (eCommerce): A specialty coffee brand uploads its list of "top 20% spenders" as a CRM seed. They create a 1% lookalike audience in the United Kingdom to find new premium coffee buyers.
Example scenario (SaaS): A software company creates a conversion-based seed from users who completed a 14-day trial. They use a "Balanced" setting on TikTok to reach a broad but relevant group of professionals interested in productivity tools.
FAQ
How many people do I need for a lookalike audience? Minimums vary by platform. Facebook lets you start with 100 people from the same country, but advises 1,000 to 5,000 for better results. TikTok requires at least 1,000 users in your source audience.
Is search targeting better than lookalike targeting? Both serve different purposes. Lookalike targeting uses behavioral and interest-based commonalities from a seed, whereas search targeting typically focuses on intent via keywords.
Do lookalike audiences work for startups? Startups may struggle with lookalike audiences initially because they lack a large existing customer base. Small sample sizes can lead to insufficient data and interference from outliers, making the audience less accurate.
What happened with the Meta lawsuit regarding lookalikes? In June 2022, the U.S. Justice Department settled a lawsuit with Meta alleging that its lookalike tool discriminated against users by distributing housing ads based on protected characteristics like race and religion.
