Glossary entry
What is a Lookalike Audience?
A lookalike audience is an algorithmic match generated from a seed list (your customers, leads, or website visitors) that finds new users with similar behavior. Lookalikes from a high-quality seed (top-LTV buyers, not all leads) outperform broad targeting by 2 to 5x.
How lookalikes work
You upload a seed (200 to 50,000 profiles work best). The platform finds users sharing demographic, interest, and behavior signals. You choose match precision: 1% (closest match, smallest pool) to 10% (loose match, biggest pool). Start with 1% to 3%.
Seed quality is everything
A lookalike of all email subscribers is a lookalike of curiosity. A lookalike of paying customers from the past 90 days is a lookalike of intent. The difference is 3 to 10x ROAS at scale.
Lookalikes after iOS 14
Meta's signal degraded but did not vanish. Pair lookalikes with a Conversion API feed and modeled conversions to restore most of the lost precision. TikTok and LinkedIn lookalikes were less affected.
How Unled Network helps
Managed Campaigns rebuilds your audience stack monthly with high-LTV seeds, layers lookalikes by intent tier, and pairs them with platform creative variants tuned to each tier.
Related Unled Network services
Frequently asked
How many people should be in my seed?
Minimum 200, ideal 1,000 to 10,000. Larger seeds dilute precision.
1% or 10% lookalike?
Start at 1% and expand only when ROAS stays acceptable at 3% and 5%.
Can I lookalike from website visitors?
Yes, but visitor lookalikes are weaker than customer lookalikes. Use a tighter percentage.
Does Google have lookalikes?
Similar concept: Optimized Targeting and Audience Expansion. Same rules apply - seed quality drives output quality.