Lookalike Audiences
Lookalike Audiences are Meta's machine-learning-generated audiences that share characteristics with a source list — your customers, leads, or website visitors — letting you reach prospects who behave like your best existing clients.
What it is
You upload a source audience (a customer email list, a pixel-based audience of past converters, or an engagement audience) and Meta builds a lookalike of users in your target country who share statistical similarities. You pick the similarity tier from 1% (closest match, smallest audience) to 10% (broadest, less precise).
Why it matters
The hardest part of paid social is finding warm prospects in a cold-traffic ocean. A well-built lookalike often outperforms interest-based targeting because it's anchored to your actual conversion data, not Meta's interest categories.
How to build effective lookalikes for healthcare
The quality of the source list determines the quality of the lookalike. For behavioral health, the best source is your verified admit list — patients who actually paid for services. Form-fill lookalikes are noisier. Avoid building lookalikes from low-intent audiences like newsletter sign-ups.
Frequently asked questions
Can I run lookalikes for healthcare without HIPAA issues?
Yes if done right. Don't upload PHI or any data that could identify a patient's health condition. Use hashed contact lists with first-party consent. For HIPAA-covered entities, work with legal counsel and consider Meta's healthcare advertising restrictions.
Should I use 1% or larger lookalikes?
Start with 1% for highest match precision when budget is small. Scale to 2–5% once 1% saturates and CPMs spike.
How often should I refresh lookalikes?
Meta auto-refreshes most lookalike types when you update the source. Manually refresh quarterly even if Meta does it automatically — and rebuild after major service line changes.
