How to Build Custom and Lookalike Audiences on Meta

Custom audiences on Meta let you target people who've already interacted with your business — website visitors, email subscribers, past customers — while lookalike audiences let Meta find new people who share characteristics with an existing custom audience. Together, they form the foundation of most effective Meta ad targeting beyond basic interest-based options.

Key takeaways:

• Custom audiences work best for retargeting people who already know your business
• Lookalike audiences extend reach to new, statistically similar prospects
• Audience quality (a well-defined custom audience) directly determines lookalike quality
• A 1% lookalike audience is more precise but smaller than a 5-10% lookalike

Custom vs Lookalike Audiences Compared

How Do I Create a Custom Audience on Meta?

In Meta Ads Manager, navigate to Audiences and choose a source — website traffic via the Meta Pixel, an uploaded customer list, or engagement with your Facebook or Instagram content. Website-based custom audiences require the Meta Pixel or Conversions API installed on your site to track visitor behavior for retargeting purposes.

What Percentage Should I Choose for a Lookalike Audience?

A 1% lookalike audience is the most precisely matched to your source audience but reaches fewer people, while broader percentages (up to 10%) reach more people with progressively less precise matching. Most advertisers start with a 1% lookalike for the tightest match, then test broader percentages if they need more scale once the tighter audience is exhausted.

A business targeting a 1% lookalike based on their highest-value repeat customers, for example, will typically see stronger conversion rates than a 5% lookalike, but the smaller audience size may limit how much budget can be efficiently spent before frequency climbs too high.

How Large Does My Source Audience Need to Be for a Good Lookalike?

Meta recommends a source audience of at least 100 people, though lookalikes built from larger, higher-quality source audiences — 1,000+ engaged customers — typically perform more reliably than those built from very small or low-engagement lists. A lookalike built from your top 500 highest-lifetime-value customers, rather than your entire customer list, often produces a more precisely targeted result.

Should I Exclude Existing Customers From My Lookalike Campaigns?

Yes — since lookalike audiences are designed to find new prospects, excluding your existing customer list prevents wasted spend showing acquisition-focused ads to people who've already converted. This is a commonly overlooked setup step that quietly inflates cost-per-acquisition when skipped.

“"The quality of a lookalike audience is a direct reflection of the quality of the custom audience it's built from. Garbage source data produces a garbage lookalike, no matter how well the campaign itself is built." — the first thing we check when a lookalike campaign underperforms.”

Building a Strong Custom Audience Foundation

• Install the Meta Pixel or Conversions API properly before relying on website custom audiences
• Segment customer lists by value — build lookalikes from your best customers, not your entire database
• Refresh custom audiences regularly so lookalikes stay based on current behavior
• Exclude existing customers from prospecting lookalike campaigns
• Test multiple lookalike percentages to find the best balance of precision and scale

Why Data Quality Determines Targeting Success

Advertisers often focus heavily on ad creative and copy while treating audience setup as an afterthought, but a poorly built custom audience limits how effective even the best creative can be. Investing time in clean, well-segmented source data before building lookalikes typically produces better results than endless creative iteration on a weak audience foundation.

Frequently Asked Questions

Q: How long does it take for a lookalike audience to be ready after creation?

A: Lookalike audiences are typically ready within 6-24 hours after creation, depending on the size and quality of the source audience used.

Q: Can I combine multiple custom audiences into one lookalike?

A: Yes, Meta allows combining multiple custom audience sources when building a single lookalike, which can strengthen the match if the combined data is relevant and high quality.

Q: Do lookalike audiences update automatically over time?

A: Yes, lookalike audiences refresh periodically based on the current state of their source custom audience, so they evolve as your source data changes.

Q: Is customer list data safe to upload to Meta for custom audiences?

A: Meta hashes uploaded customer data for privacy before matching, though businesses should still ensure they have proper consent to use customer data for advertising purposes.

Want help setting up custom and lookalike audiences that actually convert? Get a free Meta ads consultation from Ostrune and we'll build your targeting strategy from the ground up.