How to Create Meta Lookalike Audiences in 2026 (And When to Use Them)
Meta lookalike audiences let you show ads to people who share behaviors with your best existing customers. You provide a source audience, a list of buyers, email subscribers, or high-value visitors, and Meta's algorithm finds users across Facebook and Instagram who match that profile. The setup takes under five minutes. Picking the right source is what actually determines whether it works.
This guide covers the step-by-step setup, source audience selection, size recommendations, and where lookalikes still fit in a 2026 Meta strategy that's increasingly built around Advantage+ Shopping.
What Are Meta Lookalike Audiences and How Do They Work?
A lookalike audience is a targeting method in Meta Ads Manager. You provide a source audience, a Custom Audience built from your customer data, and Meta identifies people on Facebook and Instagram who share similar characteristics. Meta's algorithm considers demographics, online behavior, and purchase signals to find the match.
The size percentage controls how wide the net goes. At 1%, Meta finds the closest match to your source in a given country, typically 1-3 million people in the US. At 10%, you're reaching a much broader group that shares fewer traits with your source. Most campaigns start at 1% and expand once performance is proven.
Meta refreshes lookalike audiences automatically every 3-7 days, pulling in new users who match the evolving source profile. You do not need to rebuild them manually unless you're switching source audiences entirely.
Are Meta Lookalike Audiences Still Effective in 2026?
Yes, in the right context. Meta's Advantage+ Shopping campaigns have absorbed a lot of the work lookalikes used to do. Andromeda, Meta's 2026 algorithm update, pushed the platform further toward broad targeting and machine-driven audience selection. For accounts spending $300+/day with solid Pixel signal, the algorithm often finds your best customers without manual audience setup.
Lookalikes still outperform broad targeting when you have CRM data the Pixel has never seen: offline buyers, loyalty members, high-LTV email segments. They also work better at smaller budgets where the algorithm does not have enough data to self-optimize. See how this fits the broader shift in Meta's 2026 Andromeda update.
| Factor | Lookalike Audiences | Advantage+ Shopping |
|---|---|---|
| Best for | Smaller budgets, CRM-sourced audiences | Scaled accounts, $300+/day with strong Pixel signal |
| Targeting control | You select source and size | Algorithm controls all targeting |
| CPA (2026) | Strong with quality source audience | 17% lower CPA vs manual campaigns (Meta data) |
| Creative dependency | Moderate | High. Creative is the targeting signal. |
| Learning phase | 50 conversions/week to exit | Faster with consolidated ad sets |
| Campaign type | Standard campaign, manual ad set | Advantage+ Shopping Campaign (ASC) |
| 2026 recommendation | CRM segments and accounts under $5K/mo | Default choice for $300+/day ecommerce |
How to Create a Meta Lookalike Audience: Step-by-Step
Before you start, make sure you have a Custom Audience ready to use as your source. If you're uploading a CRM list, format it as a CSV with email, phone, or both. A Pixel-based Custom Audience built from Purchase events in the last 180 days also works well as a starting point.
Open Audiences in Meta Business Manager
Log in to business.facebook.com. In the left navigation, go to Advertise, then Audiences. Confirm you're in the correct ad account before proceeding.
Click Create Audience and select Lookalike Audience
Click the blue Create Audience button. From the dropdown, select Lookalike Audience. This opens the lookalike creation dialog.
Choose your source audience
This is the most consequential step. Select a Custom Audience built from buyers or high-value customers, not general site visitors. The minimum is 100 people, but 1,000-50,000 gives Meta enough signal to find quality matches. Source and target country must be different.
Select your target country or region
Choose where you want to find new audiences. Lookalike audiences are country-specific. If you target multiple countries, create a separate lookalike for each. Do not combine markets into a single audience.
Set the audience size (1-10%)
Use 1% to start. It finds the people most similar to your source. Expand to 2-5% if 1% is too small to exit the learning phase. Avoid 5-10% until you have validated performance at smaller percentages.
Click Create Audience and wait for it to populate
Meta typically populates the audience within 6-24 hours. Check the status in your Audiences dashboard. Do not add it to an ad set until it shows a Ready status.
Add to an ad set and test against broad targeting
In Ads Manager, create an ad set targeting the new lookalike. If you want to compare, run a separate ad set with no audience restrictions. Use the same budget, same creative, same bid strategy, and let the data decide.
How to Choose the Right Source Audience
The source audience is what Meta uses to find your lookalike users. Quality beats size. A list of 500 actual buyers will outperform a list of 10,000 generic website visitors almost every time.
| Source Type | Quality | Why It Works |
|---|---|---|
| Top 20% customers by LTV (CRM upload) | Best | Signals your highest-value buyers directly to Meta's algorithm |
| Purchase event (Pixel, last 90 days) | Excellent | Tracks actual buyers; 90 days captures recent purchase intent |
| Email purchasers (CSV upload) | Excellent | Works well for brands with offline or subscription-based sales |
| Add to cart (Pixel, last 30 days) | Good | High-intent users who came close to buying |
| All website visitors (Pixel, last 180 days) | Weak | Too broad; includes browsers, bounces, and accidental clicks |
| Video viewers (50%+, last 60 days) | Weak | Engagement signal is too soft for purchase intent modeling |
| Page Likes | Avoid | Skewed heavily by organic engagement, not buying behavior |
Meta Lookalike Audience Size Guide: Which Percentage Should You Use?
Lookalike audience size determines how closely Meta matches your source. Lower percentages are more precise but smaller. Higher percentages reach more people but share fewer traits with your source audience.
| Size | Approx. US Reach | Match Quality | When to Use |
|---|---|---|---|
| 1% | 2-3 million | Highest | Default starting point for all campaigns |
| 2-3% | 4-9 million | High | Scale up if 1% is too small to exit learning phase |
| 3-5% | 9-15 million | Moderate | Good for upper-funnel awareness alongside retargeting exclusions |
| 5-10% | 15-30 million | Lower | Only after you have exhausted smaller, better-performing percentages |
| Stacked (1%+2%+3%) | Combined | Mixed | Avoid. Stacking sizes in one ad set sends conflicting signals to Meta. |
Lookalike Audience Mistakes That Kill Performance
Most lookalike campaigns that underperform are not set up wrong. They're built from the wrong inputs. Here's what to avoid.
Using Page Likes as the source. Page engagement audiences are built from people who liked or commented on posts, not buyers. The behavioral signal is completely different, and your lookalike will reflect that.
Stacking multiple sizes in one ad set. Running 1%, 2%, and 3% in the same ad set sends conflicting signals to Meta's delivery algorithm. Test sizes in separate ad sets so you can identify what's actually working.
Running lookalikes alongside Advantage+ Shopping simultaneously. ASC already performs lookalike-style matching. Adding a separate LAL campaign at the same time splits your budget and creates audience overlap without giving you clear learnings from either.
Forgetting to exclude existing customers. Always add your current customer list as an exclusion at the ad set level. Without it, you're spending acquisition budget on people who already know you, and likely paying higher CPMs for it.
Killing the campaign before the learning phase exits. Meta needs 50 optimization events within 7 days before delivery stabilizes. Early results are unreliable. Give a lookalike campaign at least 7-10 days before making a pause-or-scale decision.
We inherited an account once that had a lookalike built from all website visitors in the past 180 days. It looked solid on paper: 120,000 people in the source. But when we dug into who those visitors actually were, about 60% had bounced in under 10 seconds. When we rebuilt the lookalike from the top 10% of customers by lifetime value (about 800 people), the performance was completely different.
Lookalike audiences work best when paired with a retargeting structure that re-engages users who did not convert the first time. If you're building a full Meta funnel, see our guides on ecommerce marketing strategy and CPG retail marketing strategy for how paid social fits into a broader acquisition framework.
What is the minimum source audience size for a Meta lookalike?
Meta requires a minimum of 100 people in your source audience, but that is just the floor. For reliable results, aim for 1,000 to 50,000 people who share a meaningful common action: purchases, subscriptions, or high-value behavior. Below that threshold, Meta's algorithm does not have enough signal to build a consistent match.
Are Meta lookalike audiences worth using in 2026?
Yes, in specific scenarios. If you're spending under $5K/month on Meta, or you have CRM data the Pixel has never captured, lookalikes can outperform broad targeting. For accounts at $300+/day with strong Pixel signal, Advantage+ Shopping typically handles audience discovery more efficiently. That does not make lookalikes obsolete. It changes when to reach for them.
How often should I refresh my lookalike source audience?
Meta automatically refreshes the lookalike audience every 3-7 days, so you do not need to rebuild it. What you should update periodically is the source Custom Audience itself. If it is Pixel-based, it updates automatically. If it is a CRM upload, refresh it every 30-90 days to add new buyers and remove people who have already converted.
What is the difference between a Custom Audience and a Lookalike Audience?
A Custom Audience is built from your actual data: a customer list, Pixel events, or video engagement. A Lookalike Audience is built by Meta to find new people who match that Custom Audience. Custom Audience is who you already know. Lookalike Audience is who you should probably know.
Should I use a Lookalike Audience or Advantage+ Shopping?
It depends on your budget and Pixel signal quality. Advantage+ Shopping is the better default for ecommerce accounts spending $300+/day with 50+ monthly purchases tracked. Lookalike audiences win when you have offline buyer data, are below that spend threshold, or want explicit control over targeting. Test both and let CPA at a consistent 7-day click window tell you which to scale.
Where Lookalike Audiences Fit in Your 2026 Meta Strategy
Lookalike audiences are not the default starting point they were in 2022. Advantage+ Shopping has taken over that role for most ecommerce accounts. But dismissing lookalikes entirely is a mistake. They are a precision tool for situations where the algorithm needs your data instead of inferring from behavior.
The playbook: start with Advantage+ Shopping as your primary acquisition campaign if you're spending $300+/day. Use lookalikes as a complementary layer for CRM-based segments not reflected in your Pixel events. Test both, measure CPA at a consistent 7-day click window, and let performance decide the budget split.
Want a Free Meta Account Audit?
We'll review your audience strategy, ad structure, and creative mix and show you exactly where lookalikes, Advantage+, and broad targeting should each be doing the work.
Get Your Free AuditLaunch into Success
Tell us a bit about yourself and your business. We are just one message away from the perfect partnership!