Quick answer. A lookalike audience is a group Meta builds by finding new people who resemble a source audience you provide, like your purchasers or high-value customers. The quality of that source matters far more than its size. A 1 percent lookalike is the closest match and wider percentages trade precision for reach. These days broad targeting with strong conversion signals often performs as well or better, so use lookalikes when your signal is thin or you want a clean test.

When I train new buyers, lookalike audiences are one of the first ideas they latch onto, and one of the first they overthink. The concept is simple. You hand Meta a list of people you like, and Meta goes out to find more people who resemble them.

The trick is not the setup, which takes a couple of minutes. The trick is feeding Meta the right people to copy, and knowing when a lookalike still earns its place next to broad targeting. Let me walk you through both.

What is a lookalike audience and how does it work?

A lookalike audience starts with a source, which is just a group of people you already know something about. Meta studies the traits that group shares and then finds new users who look similar across the things its system can see, like behavior and interests.

The source is called a custom audience. It can come from a customer list you upload, people who took an action on your site through the Meta pixel, or folks who engaged with your page, your videos, or your Instagram profile.

You are not telling Meta who to target by hand. You are showing it a sample of good people and asking it to go fishing in a much bigger pond for more of the same. The better your sample, the better the catch.

What makes a good source audience?

This is where most beginners go wrong. They pick the biggest list they have, like all website visitors, and wonder why results are soft. Size is not the goal. A tight, high-quality source is.

Here is the order I reach for, strongest first:

  • Purchasers: people who actually bought, ideally recently. This is gold because it copies real buyers, not browsers.
  • High-value customers: if you can tag your top spenders or repeat buyers, a lookalike of them chases more of your best customers, not just any customer.
  • Engaged users: add-to-carts, lead form completions, or people who watched most of a video. Useful when you do not yet have enough purchases.

You usually need at least a thousand or so people in the source for Meta to build a solid match, but do not pad the list to hit a number. A clean group of true buyers beats a bloated list of casual visitors every time.

What do the lookalike percentages mean?

When you create a lookalike, Meta asks for a percentage between 1 and 10. That number is the slice of a country's population you want to reach, sorted by how closely they resemble your source.

A 1 percent lookalike is the closest match. It is the smallest group and the people who look most like your source. As you move to 3, 5, or 10 percent, the audience grows and the resemblance loosens, so you trade precision for reach.

If you are testing or your budget is modest, start narrow with 1 percent. If you have proven the offer and need more volume, you can widen out or stack a few ranges. Wider is not worse, it just asks Meta to reach further from the original sample.

How do I use customer lists the right way?

Uploading a customer list is one of the strongest sources you can give Meta, but it comes with a real responsibility. You can only use data you are allowed to use. That means contacts who gave you their information and understood it could be used for marketing.

Do not buy lists, scrape emails, or upload data people did not knowingly share with you. It breaks Meta's terms, it can violate privacy rules in the US, and it tends to perform poorly anyway because those people have no relationship with you.

When you upload, Meta hashes the data to match it against accounts, then builds the lookalike from the matches. Keep your list current, keep your consent clean, and you get the upside without the headache. The same rule applies to any custom audience: it has to come from data you have the right to use.

When do lookalikes still help, and when should I go broad?

Meta's system has gotten very good at finding the right people on its own, especially when you feed it strong conversion signals through the pixel and the Conversions API. With Advantage+ style setups, Meta often treats a lookalike as a suggestion and looks beyond it anyway, which is why broad targeting now performs as well or better in a lot of accounts.

So when do lookalikes still earn their spot? A few cases:

  • You are early and your pixel does not have enough conversion data yet, so a lookalike gives Meta a head start.
  • You sell to a narrow or unusual audience that broad targeting struggles to find.
  • You want a clean test of one variable against your broad campaign.

My honest take after running both: let strong signals and a clear offer do the heavy lifting, and treat lookalikes as a tool you reach for when the situation calls for it, not a default you bolt onto every campaign.

Key takeaways

  • A lookalike copies a source audience, so the quality of that source matters more than its size.
  • Start with a 1 percent lookalike of purchasers or high-value customers, then widen for reach.
  • Only use customer lists you have permission to use, and lean on broad targeting plus strong signals when your data is solid.

Frequently asked questions

How big does my source audience need to be?
Aim for at least a thousand people, ideally more, but quality beats size. A clean source of real purchasers builds a stronger lookalike than a huge list of casual visitors. If you do not have enough buyers yet, use a high-intent action like add-to-cart or lead completions as your source instead of padding the list.
Is a 1 percent lookalike always better than a 5 percent one?
Not always. A 1 percent lookalike is the closest match and great for precision or testing, but it is a small pool that can get expensive as you scale. A wider percentage gives Meta more room to find volume. Start narrow, prove the offer, then widen when you need more reach.
Can I upload my email list to build a lookalike?
Yes, as long as those contacts gave you their information and understood it could be used for marketing. Never buy or scrape lists. Meta hashes your data to match it to accounts, then builds the lookalike from the matches. Clean consent keeps you compliant and tends to perform better anyway.
Do lookalikes still matter with Advantage+ and broad targeting?
Less than they used to. With strong conversion signals from the pixel and Conversions API, Meta often finds the right people on its own and broad targeting performs as well or better. Lookalikes still help when your data is thin, your audience is unusual, or you want a clean test against broad.