Learn how Meta lookalike audiences work, which customer source to use and when broad targeting may be the better choice.
A Facebook lookalike audience is a group of new people Meta identifies as similar to an audience you already have.
You choose a source, such as a customer list or website audience. Meta uses shared characteristics to find other people who may be interested in the business.
A lookalike audience is a prospecting tool. It is not the same as retargeting people who already visited your website.
The Simple Answer
Lookalike audiences work best when the source represents the customers you want more of.
A list of paying customers is usually more useful than a list of every email address collected. A list of repeat or high-value customers may be even better if it is large enough to match reliably.
Meta can also expand beyond a defined lookalike in some Advantage audience settings. Always review the current ad-set controls instead of assuming the audience is a hard boundary.
Source Audience vs. Lookalike Audience
| Audience | Who it contains | Main use |
|---|---|---|
| Source or Custom Audience | People already connected to the business through a list, website, app or Meta engagement | Retargeting or building a lookalike |
| Lookalike Audience | New people Meta considers similar to the source | Reaching possible new customers |
People in the source are excluded from the resulting lookalike audience, according to Meta. You can still set additional exclusions where the current campaign controls allow them.
For ads aimed at previous visitors, read Darlington's Facebook Retargeting guide.
Choose the Right Source
Source quality matters more than collecting the largest possible list.
| Source | What it teaches Meta | Watch out for |
|---|---|---|
| Paying customers | Traits shared by buyers | Mixes small and high-value customers unless segmented |
| Repeat or high-value customers | Traits shared by the best customers | May be too small or slow to update |
| Qualified leads | People who fit the business | Includes people who did not buy |
| All leads | People who submitted a form | Can include spam and poor fits |
| Website converters | People who completed a tracked action | Depends on accurate tracking and event quality |
| Website visitors | People who visited | Many may have little buying intent |
| Video or social engagement | People who interacted on Meta | Engagement does not necessarily mean purchase intent |
Do not upload customer information unless the business has the right permissions and processes to use it. Follow Meta's Custom Audience terms and applicable privacy rules.
Lookalike Size and Similarity
Meta lets advertisers choose a percentage range for a lookalike audience.
Smaller percentages are generally closer to the source but include fewer people. Larger percentages create a broader audience with less similarity.
That is a tradeoff, not a universal ranking. A smaller audience is not always better, and a larger audience is not always worse. Market size, location, budget, source quality and Meta's delivery system all affect the result.
Meta's help material recommends a source audience of roughly 1,000 to 5,000 people but also emphasizes that quality matters. It lists at least 100 people from one country as a source requirement. Platform rules can change, so confirm the current requirement in Ads Manager before planning a campaign around an exact count.
How to Create a Lookalike Audience
The exact interface can change, but the basic process is:
- Open Audiences in Meta Ads Manager.
- Create or select a Custom Audience as the source.
- Choose the location where you want to find new people.
- Select the lookalike percentage or range.
- Create the audience and allow it to populate.
- Select it in an eligible ad set.
- Review any Advantage expansion settings and exclusions.
The campaign's objective should match the actual result you want, such as leads or sales.
Lookalike or Broad Targeting?
Broad targeting gives Meta fewer audience restrictions. A lookalike provides a source signal based on people you already know.
Test broad targeting when:
- Conversion tracking is strong
- The market is large enough
- The offer has broad appeal
- The campaign can generate enough customer signals
Test a lookalike when:
- You have a useful source of real customers
- The customer type is relatively specific
- You want to compare a source-led audience with broad delivery
Read Darlington's Facebook Broad Targeting guide for the tradeoffs.
A Simple Test
Suppose a business tests the same offer and ads in two prospecting ad sets.
| Result | Lookalike | Broad |
|---|---|---|
| Spend | $1,500 | $1,500 |
| Clicks | 300 | 375 |
| Qualified leads | 30 | 25 |
| Customers | 6 | 5 |
| Cost per customer | $250 | $300 |
The broad audience gets more clicks. The lookalike produces more customers at a lower cost.
The calculations are:
- $1,500 ÷ 6 = $250 per lookalike customer
- $1,500 ÷ 5 = $300 per broad customer
This one result is not a permanent rule. Delivery changes, and a different source, creative or period may reverse the outcome. Repeat useful tests and judge them by customers and revenue.
How to Improve the Source
- Remove obvious spam and invalid records.
- Separate customers from unqualified leads.
- Use recent data when the business changes quickly.
- Consider a source of repeat or high-value customers.
- Keep website and sales tracking accurate.
- Refresh customer lists through an approved process.
Better source data gives Meta a clearer example of the customer you want.
Common Lookalike Audience Mistakes
Using Every Lead as the Source
If half the leads are poor fits, the source teaches Meta the wrong lesson.
Changing Audience and Creative Together
You will not know whether the audience or the ad caused the difference.
Assuming the Audience Is Fixed
Advantage features may expand delivery beyond a selected lookalike in eligible campaigns.
Using Too Many Similar Ad Sets
Overlapping tests can split budget and slow learning. Keep the structure simple.
Judging by Click Cost
A cheap click from the wrong person is still waste. Track qualified leads, customers and revenue.
Facebook Lookalike Audience FAQ
What is the best source for a lookalike audience?
Usually a clean list or tracked group of the customers you want more of. The best source depends on data quality and size.
Is a 1% lookalike always best?
No. It is narrower and more similar, but a broader percentage can sometimes give Meta more room to find customers. Test against your business result.
Can I create a lookalike from website visitors?
Yes, if the website audience is set up and eligible. A purchase or qualified-lead source may carry a stronger customer signal than all visitors.
Does a lookalike include my existing customers?
Meta says the source audience is excluded from the lookalike. Review campaign exclusions and Advantage settings to confirm the delivery you want.
Are lookalike audiences still useful with Advantage+?
They can still provide a useful source signal, but Meta may use more automated expansion. Compare the result with broad targeting instead of assuming one method wins.
The Bottom Line
A lookalike audience helps Meta find new people who resemble an existing source.
Choose the source carefully, understand that delivery controls can expand, and compare it with broad targeting. The useful audience is the one that produces qualified customers at a sustainable cost.
Sources
- Meta Business Help Center: About lookalike audiences
- Meta Business Help Center: About reaching new audiences
- Meta Customer List Custom Audiences Terms
Want Better Results From Paid Media?
Darlington manages advertising and measurement for businesses that want clearer decisions and more profitable growth.
Talk With Darlington