Lookalike Audiences in Meta Ads: Complete 2026 Guide

Lookalike audiences remain one of the most powerful prospecting levers in Meta Ads. Meta’s platform reaches over 3.2 billion daily active users across its family of apps (Meta Investor Relations, 2024), giving the lookalike algorithm an enormous pool to find users who match your best customers’ profiles. But how to use them changed radically between 2021 and 2026. The impact of iOS 14.5, the launch of Advantage+ Audiences, and the shift to CAPI-first tracking transformed both the mechanics and the optimal strategy. For the complete Meta Ads ecommerce setup guide

If your knowledge of lookalike audiences comes from tutorials written before 2022, this guide is for you.

---

1. What are lookalike audiences in Meta Ads
2. Lookalike audiences vs. Advantage+ Audiences: which to use in 2026
3. How to create a quality seed audience
4. The impact of iOS 14.5 on your lookalike audiences
5. How to create a lookalike audience step by step
6. What size to choose: 1%, 3% or 10%
7. Value-based lookalike audiences
8. Best practices for 2026
9. Common mistakes to avoid

---

Key Takeaways

• Lookalike audiences remain a powerful prospecting tool, but the Conversions API (CAPI) is now essential: it recovers 15-30% of events lost to iOS 14.5+ tracking restrictions (Meta for Business, 2024).
• A lookalike based on 500 real buyers consistently outperforms one based on 10,000 newsletter subscribers — quality of the seed audience matters more than volume.
• On accounts with 50+ weekly conversions, Advantage+ Audiences typically outperforms manual lookalike targeting; start with 1% lookalike audiences if conversion volume is lower.
• Value-based lookalike audiences (seeded from high-LTV customers) improve prospecting ROAS in ecommerce stores with variable customer value.

---

What are lookalike audiences in Meta Ads

A lookalike audience is a targeting type in Meta Ads that lets Meta’s algorithm find people with similar characteristics to a seed audience you define. Meta reaches over 3.2 billion daily active users (Meta Investor Relations, 2024), giving the model enough scale to find your audience in virtually any market. The more precise and conversion-focused your seed audience, the more accurate the lookalike becomes.

The principle is simple: you tell Meta “my best customers are these people” and the system analyses their demographic, behavioural, and interest attributes within its platform to find other users with a similar profile — people who don’t yet know your brand but statistically have a higher likelihood of converting.

The seed audience can be:

• A customer list (CSV with emails, phone numbers, or identifiers)
• Buyers tracked through the Meta pixel or Conversions API
• Visitors to your website who completed a high-value action
• People who engaged with your content on Instagram or Facebook
• Your email list subscribers
• People who watched at least 75% of one of your videos

The quality of the seed audience directly determines the quality of the resulting lookalike audience. Garbage in, garbage out.

---

Lookalike audiences vs. Advantage+ Audiences: which to use in 2026

Since 2023, Meta has actively pushed toward Advantage+ Audiences, its AI-based targeting system. Meta’s own testing shows that Advantage+ Audiences outperforms manual audience targeting on accounts with sufficient conversion history — but the threshold matters (Meta for Business, 2024). Here’s what each option actually does before you pick one.

What Advantage+ Audiences does

Advantage+ Audiences uses the conversion signals from your account (pixel, CAPI, campaign history) to automatically find the audience most likely to convert, without you having to manually define interests, behaviours, or custom audiences.

You can provide a suggested audience (equivalent to your seed audience) as a starting point, but the system has the freedom to expand beyond it if it identifies conversion opportunities.

When to use manual lookalikes vs. Advantage+

Situation	Recommendation
Account with little conversion data (<50/week)	Manual lookalike — more control
Established account with good conversion volume	Advantage+ Audiences
New prospecting campaign without history	Start with 1-3% lookalike, migrate to Advantage+ as data accumulates
eCommerce with catalogue and many buyers	Advantage+ Shopping Campaigns (full automation)
Very specific niche where interests don’t work	Lookalike audience based on customer list

On accounts where I’ve been able to run a proper side-by-side comparison — manual 1% lookalike vs. Advantage+ Audiences with the same seed as a suggested audience — Advantage+ consistently wins on CPA once the account has 50+ weekly conversions. But it’s not magic: accounts below that threshold often see Advantage+ spend broadly and inefficiently, while a tight 1% lookalike from a buyer list stays disciplined. My rule now is to start new accounts with a 1% buyer lookalike and migrate to Advantage+ once conversion volume justifies it.

---

How to create a quality seed audience

The seed audience is the core of the system. A poor seed produces a poor lookalike audience, regardless of budget.

Seed audience quality hierarchy

From highest to lowest quality:

1. Best customer list: customers with the highest LTV (lifetime value), not all customers. Use only the top 20-25% by average ticket or purchase frequency.
2. Recent buyers: the last 180 days, tracked by pixel or CAPI.
3. Qualified leads: people who completed a form or started a purchase process.
4. Conversion page visitors: those who reached your product, service, or checkout page, even if they didn’t complete the action.
5. Active email subscribers: those who open or click, not your entire list.
6. Facebook page fans: the lowest quality option; avoid it unless you have no other source.

Recommended minimum size

Meta requires a minimum of 100 people in the seed audience to create a lookalike. But the practical minimum for quality results is:

• 1,000 people: functional lookalike audience
• 5,000-10,000 people: good quality lookalike audience
• More than 10,000: optimal for broad lookalike audiences (5-10%)

If your list has fewer than 1,000 records, consider using a behaviour-based source (website visits, Instagram interactions) that has more volume.

---

The impact of iOS 14.5 on your lookalike audiences

Apple’s iOS 14.5 ATT framework reduced measurable Meta pixel conversion events by 15–30% for most ecommerce advertisers. The Conversions API (CAPI) recovers those events by sending conversion data server-side, bypassing browser-level blocks — making it the single highest-priority technical fix before scaling lookalike-based prospecting budgets. (Meta for Business, 2024)

In April 2021, Apple launched iOS 14.5 with the App Tracking Transparency (ATT) framework, which required apps (including Facebook and Instagram) to ask users for explicit permission before tracking their activity. Most users declined. The result: pixel-based conversion tracking dropped significantly, especially on Apple devices, directly degrading the quality of custom and lookalike audiences built from web visitor data (Meta for Business, 2024).

The effect on lookalike audiences was direct: pixel signal was significantly reduced, especially for conversion events on Apple devices. This affects:

• The accuracy of custom audiences based on web visitors
• The quality of lookalike audiences derived from those lists
• The volume of events available to create conversion-based audiences

How to compensate in 2026

The right answer is the Meta Conversions API (CAPI). While the Meta pixel works from the user’s browser (and is therefore subject to iOS blocks), CAPI sends events directly from your server to Meta, without depending on the user’s device or browser.

With CAPI correctly configured:

• You recover 15-30% of events lost to iOS 14.5+
• Your custom audiences become more complete again
• Lookalike audiences derived from them are higher quality

If you don’t have CAPI implemented, it’s the first technical optimization you should prioritize before scaling budget on prospecting with lookalike audiences.

---

How to create a lookalike audience step by step

The process is done from Meta Ads Manager (ads.manager.facebook.com).

Step 1: Create your custom audience (seed)

Before creating a lookalike audience you need a custom audience as a seed. Go to Audiences within Ads Manager and create a new custom audience from:

• Customer list: upload a CSV with emails, phone numbers, or Meta user IDs
• Website: audience based on pixel events (buyers, visitors, etc.)
• Meta activity: people who engaged with your page, videos, or Instagram/Facebook forms

Step 2: Create the lookalike audience

With your custom audience ready:

1. In Audiences, click “Create Audience” - “Lookalike Audience”
2. Select the source audience (the custom one you just created)
3. Choose the target country or countries
4. Choose the size (from 1% to 10% of the selected country’s population)
5. Click “Create Audience”

Meta will take between 1 and 24 hours to process and make the lookalike audience available.

Step 3: Apply the audience to your ad set

When creating or editing an ad set, go to the “Audience” section, click “Custom Audiences” and select the lookalike audience you just created.

Important: Don’t combine lookalike audiences with detailed interest or behaviour targeting unless you have a specific reason. The lookalike audience is already built on behavioural attributes; adding interest filters reduces it unnecessarily.

---

What size to choose: 1%, 3% or 10%

1% lookalike audiences generate the highest conversion rates in Meta Ads prospecting but exhaust quickly in small markets. 5–10% audiences reach a larger pool at lower precision, making them useful for brand awareness or when the 1% tier is saturating. Frequency above 2.5–3 in prospecting is the clearest signal that an audience needs replacing. (WordStream, 2024)

The percentage you choose determines the size and precision of the resulting audience. WordStream’s analysis of Meta Ads prospecting campaigns shows that 1% lookalike audiences generate the highest conversion rates but are limited in scale, while 5-10% audiences work better for brand awareness and retargeting-exhausted markets (WordStream, 2024).

• 1%: the audience most similar to your seed. Smaller but higher quality. Ideal for starting prospection with the least possible risk.
• 2-3%: balance between precision and volume. Works well in medium-sized markets (UK, France, Mexico).
• 5-10%: broad audience, less similar to the seed but with much more reach. Useful for scaling when 1% is exhausted or in small markets with little volume.

Recommended scaling strategy

Always start with 1%. When the cost per result starts rising because the audience is saturating (high frequency, growing CPM), move to 2-3%. Don’t jump directly to 10% without having validated the message with more qualified audiences.

In countries with large populations (USA, Brazil, Mexico), 1% can represent several million people — more than enough for months of campaign before needing to scale.

---

Value-based lookalike audiences

One of the least used but most powerful features of Meta Ads is the value-based lookalike audience.

Instead of treating all clients on your list identically, you can assign a numerical value to each (their LTV, average ticket, order value) and upload that information to Meta. The system will prioritise the attributes of your highest-value customers when building the lookalike audience, resulting in prospection oriented not just towards “people similar to your customers” but “people similar to your best customers”.

To use this feature, your CSV must include a value column (labelled value or LTV). Meta will automatically detect the column and offer the option to create a value-based lookalike audience during the upload process.

Value-based lookalikes are significantly underused, particularly in ecommerce stores with wide customer value ranges. In my experience, the performance lift vs. a standard buyer lookalike is most pronounced in stores where the top 20% of customers account for more than 60% of revenue — which is more common than most advertisers realize. If your customer data shows that pattern, generating a value-based lookalike from your top-tier customers while excluding mid-range buyers from the seed can produce a meaningfully different (and better) prospecting audience.

This technique is especially effective in ecommerce where there is significant variability in value per customer.

---

Best practices for 2026

1. Use CAPI before scaling

Without the Conversions API, your event signal is incomplete due to iOS 14.5+. Prioritise implementation before investing significant budget in prospection. Full Meta Ads for ecommerce setup guide including CAPI

2. Prioritise buyer lists over prospect lists

A lookalike audience based on 500 real buyers consistently outperforms one based on 10,000 newsletter subscribers. Source quality matters more than volume.

3. Refresh your seed audiences regularly

Lookalike audiences are created at a specific point in time and don’t update automatically. If your buyer list grows, create a new updated seed audience and generate a new lookalike. Many advertisers work with seed audiences that are months old without realising it.

4. Always exclude your existing customers

When using lookalike audiences for prospection, exclude your custom audience of existing buyers from the ad set. Paying to show acquisition ads to people who are already customers is one of the most common waste points in Meta Ads.

5. Combine lookalike audience with Advantage+ when you have data

Once you have sufficient conversion history (more than 50 weekly conversions), consider using your customer list as a suggested audience within Advantage+ Audiences instead of as a manual restriction. The system has freedom to find conversions beyond 1% or 3%, and with sufficient signal it will do so at a competitive CPA.

6. Control costs with these metrics

Review weekly for prospecting campaigns with lookalike audiences:

• Cost per result (lead, purchase, etc.)
• Frequency: if it exceeds 2.5-3 in prospection, the audience is saturating
• CPM: a rising CPM is a signal that Meta is bidding against itself or the audience is exhausted

Once your lookalike audiences are running, the next priority is efficiency — see the full guide to ROAS optimization strategies to get more from the traffic you’re already paying for.

---

Common mistakes to avoid

Using page fans as your seed audience. A Facebook page’s followers are the worst starting point for a lookalike audience: they’re heterogeneous, many may be bots or passive followers, and they have no purchase signal. Always use sources based on conversion behaviours.

Creating lookalike audiences with lists that are too small. With fewer than 300-500 people in the seed, the resulting lookalike has poor statistical precision. Meta builds the profile on an insufficient sample.

Overlapping 1%, 3%, and 5% lookalike audiences in the same campaign. If you have multiple lookalike audiences active simultaneously without exclusions between them, Meta bids against itself, inflates CPM, and confuses campaign learning. Use exclusions or consolidate into separate ad sets with separate budgets.

Ignoring audience exhaustion. A 1% lookalike in a small market can exhaust itself in weeks. Monitor frequency and prepare the next layer (2-3%) before the cost per result spikes.

Not updating seed audiences. Custom audiences and their derived lookalikes are static. If your customer base grows, you need to regenerate the seed audience and create a new lookalike to capture the most recent patterns.

---

If you want to analyse how your prospecting audiences are performing in Meta or review the complete structure of your account, book a free audit. I review advertising costs, audience quality, and technical setup in a 30-minute session.

---

Frequently Asked Questions

How often should I refresh my lookalike audience seed?

Every 30-60 days for active ecommerce accounts. Lookalike audiences are created at a fixed point in time and don’t update automatically when new buyers are added (Meta for Business, 2024). Refreshing the seed audience monthly ensures the algorithm is working from current purchase behavior patterns, not a snapshot from six months ago.

Can I use lookalike audiences without CAPI?

Yes, but the quality will be degraded. Without CAPI, your buyer lists and pixel-based audiences are missing 15-30% of Apple device events due to iOS 14.5+ restrictions (Meta for Business, 2024). This means your seed audience is incomplete, which produces an incomplete lookalike. For serious prospecting investment, CAPI implementation is a prerequisite.

What’s the minimum list size to create a quality lookalike audience?

Meta requires 100 people minimum to create a lookalike, but practical quality starts at 1,000 records. Between 5,000 and 10,000 records is the sweet spot for good quality, and 10,000+ is optimal for broad audience sizes (5-10%). If your buyer list is under 1,000, use website events (30-day visitors, add-to-cart events) as your seed instead — they’ll have more volume.

Should I combine lookalike audiences with interest targeting?

No, for most cases. Lookalike audiences are already built on behavioral and demographic attributes from your seed. Adding interest filters reduces the audience size unnecessarily and can exclude people who would have converted. The exception: very broad 5-10% lookalike audiences in large markets, where additional qualifiers can improve precision without killing scale.

How do I know if my lookalike audience is exhausted?

Watch three signals: rising frequency (above 2.5-3 in prospecting), rising CPM (Meta bidding more aggressively for a shrinking eligible pool), and rising cost per result. When two or three of these trend upward simultaneously on a prospecting campaign, the audience is likely saturating. Move to the next size tier (1% to 3%, or 3% to 5%) and refresh the seed audience.

---

Sources

1. About Lookalike Audiences - Meta Business Help Center
2. Advantage+ Audiences - Meta Business Help Center
3. Conversions API - Meta for Developers
4. About Value-Based Lookalike Audiences - Meta Business Help Center
5. iOS 14 and Meta Ads - Meta Business Help Center
6. Meta Investor Relations - Annual Report 2024
7. WordStream - Facebook Ads Benchmarks by Industry