You have a list of past customers, a website that gets visitors, or an ad account with a few months of data. Somebody told you Facebook lookalike audiences can find more people like your best customers. Somebody else told you they are going away, or that Meta ignores them now. You are about to spend real money on an ad, and you would rather not spend it on a setting that no longer does anything. The short version: Facebook lookalike audiences are still available in Meta Ads Manager, and they work from a source audience you choose. What changed is how much weight they carry when Advantage+ audience is switched on. This post walks through the job in the order you do it: check the source, build the lookalike, pick a size, understand what Advantage+ does with it, and run a fair test against broad targeting. If your source audience is weak, the free answer comes first, and you may not need a lookalike yet.
Key Takeaways
Meta builds a lookalike audience from a Custom Audience source and finds new people similar to it in the locations you pick. Meta needs at least 100 people from the same country in the source.
A lookalike is generally an audience suggestion, not a strict delivery boundary, so Meta can reach beyond it when its system predicts better results.
Keep the objective, creative, landing page, conversion event, exclusions and budget method the same, and judge qualified leads or purchases, not clicks.
Are Facebook lookalike audiences still around, and how do they work today?
Yes. A Meta lookalike audience starts with a Custom Audience, which Meta calls the source. Meta then finds new people who are similar to the people in that source. Meta's guidance on defining your ad audience and its help article on creating a lookalike audience both describe it this way. The source can be a customer list, website visitors, app users, offline activity, or people who engaged with your Facebook or Instagram content.
You then make two choices. First, you pick one or more target locations. Second, you pick an audience size as a percentage of the people in those locations. Meta's size control runs from 1% to 10%. A smaller percentage is more similar to your source. A larger percentage reaches more people and is less similar. Meta's Ads Manager labels and controls can change, so match these steps to what you see on your own screen.
A lookalike is a modeled group, not a list of named people and not a promise. Meta is making a prediction from your source. It does not guarantee that each person it reaches will resemble your customers, and it does not tell you who is inside the group. That is why the quality of the source matters so much, and why the next section comes before any setup.
What source audience should a small business use?
Use the group that looks most like the customers you want more of. A small business usually has a few candidates, and they are not equal.
| Source | What it holds | What to check before using it |
|---|---|---|
| Customer list | People who bought or became clients | Whether it is recent, whether it mixes big and small customers, and whether you have the right to upload it |
| Qualified leads | People who asked and fit what you sell | Whether unqualified or spam contacts were removed |
| Website visitors | People who visited pages you choose | Whether the visitors are buyers or casual browsers |
| Facebook or Instagram engagers | People who interacted with your content | Whether engagement tells you anything about buying |
| App users or offline activity | People from your app or in-person sales | Whether the data is current and covers one country at a time |
A purchaser, a qualified lead or a high-value customer source usually tells Meta more than an undifferentiated pool of page engagers. Someone who liked a funny post and someone who paid you for a roof are different signals. If you can pick, choose the customers you would happily get ten more of.
Meta requires at least 100 people from the same country in the source audience before it will create a lookalike. A list of 100 mixed contacts from three countries does not qualify, and a bigger list is not automatically a better one. A smaller list of real buyers can be a stronger signal than a large list of weak ones. If you have not built a customer list yet, the guide on making a Facebook custom audience from your customer list covers that step.
Then there is permission. Meta's Customer List Custom Audiences Terms require you to have the rights, permissions and lawful basis to upload and use that data. If you are not sure you have them for a list, sort that out before the upload, not after.
How do I check my source audience in ten minutes?
You can do this without hiring anyone and without changing a thing in your account. Write down what you find before you touch any setting. The steps below take you from the source to the results.
Open the source and read its basics
In Meta Ads Manager, open Audiences and find the Custom Audience you plan to use. Note its estimated size, where the data came from, how recent it is, and which countries the people are in. Confirm that at least 100 of them are from one country. If the source mixes purchasers with people who only clicked once, split them into separate sources if you can.
Ask what the group represents
Ask whether these people are customers, qualified leads, or only engagement. A source made of people who paid you is a different thing from a source made of everyone who ever opened your page. If you cannot say what the group represents, Meta cannot either.
Compare last month's results by business outcome
If you ran a lookalike ad set last month, pull its results next to a broad ad set from the same period. Record spend, impressions, clicks, leads or purchases, cost per result, revenue where you track it, and lead quality. Do not decide anything from one small or incomplete set of numbers. If tracking was patchy, fix that first, and what to set up first with the Meta Pixel and Conversions API is the place to start.
How do I create a lookalike audience in Ads Manager?
Once the source passes the check, the build is short. In Ads Manager, go to Audiences, choose Create Audience, then choose Lookalike Audience. Pick your source, pick the location or locations, and pick the audience size. Then create it and give it a name that tells you the source and the size, so you can find it later.
Name it for the source and the size, not for the date alone. A label like "purchasers 1%" tells you what you are looking at six months from now. A label with only a date tells you nothing.
Choose the location with care. The lookalike is built in the places you select, and your ad can only reach people there. If you serve a handful of towns, a lookalike across a whole country adds reach you cannot use. If your ad account runs across more than one country, remember that the source has to hold at least 100 people from one country for the lookalike to be made. For radius and place settings in general, see how to target Facebook ads to a local radius.
What does the 1% to 10% size setting mean?
The percentage is a slice of the people in your chosen location. At 1%, Meta takes the group it predicts is most similar to your source. At 10%, it takes a much larger group that is less similar. Meta offers a range from 1% to 10%, and no single percentage is the best choice for every small business.
That matters because "always use 1%" is an old habit that gets repeated. The right size depends on your source quality, your conversion data, your location, your creative, your objective and your test results. A small local business in a small place may find that a larger slice is the only way to get enough reach. A business with a strong purchaser list may be happy with a tight one. You find out by testing, not by copying a number from someone else's post.
Does Advantage+ audience ignore my lookalike?
No, but it changes the job the lookalike does. Meta's help article on Advantage+ audience and its guidance on planning your campaign audience describe lookalikes under Advantage+ as audience suggestions. Meta says it prioritizes people who match the suggestions, then searches more broadly when its system predicts better results. Suggestions can include custom audiences, lookalike audiences, age, gender and detailed targeting.
Some controls still hold. Location, minimum age, language where it is available, and excluded Custom Audiences stay as restrictions. So, a lookalike you choose gives Meta a starting point. It does not fence in every impression. A selected 1% lookalike may guide the first delivery without limiting everything to that modeled group.
So, are Facebook lookalike audiences going away? The feature is still there, and your first-party customer data is still a useful signal. What has changed is that broad delivery can overlap with, or reach beyond, the lookalike you picked. If you set a tight 1% and expect it to behave like a wall, you will be surprised by who sees the ad.
When does a lookalike help a small business, and when should you skip it?
A lookalike is most likely to help when three things are true. The source represents customers you actually want. The source has enough people to model. And your campaign has a conversion event you can measure, such as a purchase or a qualified lead.
It is a poor fit when the source is small, stale or mixed. It is also a poor fit when tracking is broken, when the offer is weak, or when the landing page does not turn visitors into inquiries. A lookalike cannot repair any of those. It cannot make an unqualified lead form qualified, and it cannot make a weak offer strong. If your ads run and the numbers miss, the ad set may have stalled for another reason, and an audience change will not fix it.
So, the free route comes first. If your source has fewer than 100 people from one country, or it is mostly people who never bought, use broad targeting while you build better customer data. Fix tracking. Clear up the offer. Come back to a lookalike when you have a source worth modeling. Skipping it for now is a reasonable decision, not a failure.
How do I test a lookalike against broad targeting?
A fair test changes one thing: the audience. Everything else stays the same, or the result tells you nothing about the lookalike. Meta lists audience testing as a valid experiment dimension, and you can use Meta's A/B testing tool where it is available, or run two separate ad sets with controlled variables. For the wider process, see how to A/B test Facebook ads.
| Setting | In both ad sets |
|---|---|
| Campaign objective | Same |
| Creative and offer | Same |
| Landing page | Same |
| Conversion event | Same |
| Exclusions and placements | Same |
| Budget method and time period | Same |
| Audience | Different: lookalike in one, broad in the other |
Judge the test on what your business cares about. That means qualified leads, booked appointments, purchases and revenue, and the quality of the leads you got. Do not decide on cost per click, impressions or raw lead count alone, since cheap clicks and a high number of unqualified leads can look good and mean nothing. Give each ad set enough time to produce real outcomes before you call a winner. No universal setting wins for every business, so your own numbers are the evidence.
If you are acquiring new customers, exclude your existing ones from both ad sets so the comparison is fair. And if your budget is small, decide the spend first. How much to spend on Facebook ads explains how to work that number back from what a customer is worth.
Can I use a lookalike for housing, employment or credit ads?
Special rules apply. Housing, employment, credit, and political or social-issue advertising have additional restrictions on Meta. Meta removed ordinary lookalike treatment for certain regulated advertising use cases and introduced Special Ad Audiences as an alternative in 2019. In its June 2022 post on expanding its work on ads fairness, Meta announced it would retire Special Ad Audiences while continuing to apply fairness and delivery safeguards. Meta's January 2023 post on its variance reduction system describes how ads are delivered across demographic groups.
If your ad falls into one of those categories, choose the matching Special Ad Category in Ads Manager and confirm the current restrictions before you launch. Do not treat it as ordinary targeting.
Check your lookalike basics
Pick an answer to begin.
1. How many people from the same country does Meta require in a lookalike source audience?
2. Under Advantage+ audience, what is a lookalike generally treated as?
3. What should stay the same when you test a lookalike against broad targeting?
Frequently Asked Questions About facebook lookalike audiences
Are Facebook lookalike audiences going away?
No. Meta still supports them in Ads Manager. Under Advantage+ audience, a lookalike is generally used as a suggestion rather than a strict limit, so how much it shapes delivery has changed.
How large must the source audience be?
Meta requires at least 100 people from the same country. A bigger list is not automatically better than a smaller, higher-quality customer source.
What can I use as a source?
Customer lists, website activity, app activity, offline activity, and Facebook or Instagram engagement can all be used where available. Purchasers and qualified leads usually say more about your best customers than page engagement does.
What does 1% mean?
It selects a smaller, more similar modeled group within the location you choose. Meta offers settings up to 10%, and no single percentage is right for every business, so test.
Does Advantage+ stay inside my lookalike?
Not necessarily. The lookalike may act as an audience suggestion, while controls such as location and excluded Custom Audiences can remain boundaries.
Can a small business use a lookalike in a regulated category?
Special Ad Category rules apply to housing, employment, credit, and political or social-issue ads. Choose the right category in Ads Manager and confirm the current restrictions first.
Final Thoughts
Lookalike audiences are not dead, and they are not magic. They are a tool that works from the source you give it. Check that source first: confirm it holds at least 100 people from one country, that it represents customers or qualified leads, and that you have the right to use the data. Then build the lookalike, pick a size on purpose, and remember that under Advantage+ it guides delivery without fencing it in.
You can do all of that yourself this week, for free. Compare a lookalike ad set with a broad one on qualified leads or purchases, keep every other setting the same, and let the results decide. If the source is thin or the tracking is shaky, spend your effort there before you spend more on ads. That is a better use of the budget than a clever audience.
If you want a second pair of eyes, Web Leveling can look at your source audience, your tracking and your test setup, and tell you whether a lookalike is worth running yet. Our Facebook ads service covers the audience work, measurement and testing, and if the answer is "not yet," we will say so. We work with small and medium businesses across the country and overseas. Send us your audience and results questions and we will help you read them.
Terms
Lookalike audience words in this post
Tap a term to see what it means.
Lookalike audience. A modeled group Meta builds to find new people similar to a source audience you choose.
Source audience. The Custom Audience a lookalike is built from, such as customers, qualified leads or website visitors.
Custom Audience. A group made from your own data, such as a customer list, website visitors or app users.
Advantage+ audience. A Meta setting that treats your audience choices as suggestions and can reach beyond them.
Broad targeting. Letting Meta find the audience with few or no audience suggestions from you.
Special Ad Category. A Meta category for housing, employment, credit, and political or social-issue ads with extra restrictions.
