What Is a Lookalike Audience on Meta? Complete DTC Explainer

A lookalike audience on Meta is a targeting tool that identifies and reaches new users who share statistical similarities with your existing customers, website visitors, or any other source audience you define.

Last updated: February 2026

Table of Contents

• How Lookalike Audiences Work
• Creating a Lookalike Audience
• Lookalike Percentage Sizes
• Best Source Audiences for DTC
• Lookalike vs Broad Targeting in 2026
• Performance Data
• When to Use Lookalikes
• Common Mistakes
• Key Takeaways
• FAQ

How Lookalike Audiences Work

When you create a lookalike audience, Meta analyses the people in your source audience (e.g., your customer list) and identifies hundreds of characteristics: demographics, interests, online behaviour, device usage, purchase history, content consumption patterns, and more.

Meta then searches its entire user base to find people who match those characteristics most closely. The resulting audience is people who look like your best customers, even though they have never interacted with your brand.

The algorithm is processing data from billions of users and billions of data points. It is significantly more sophisticated than any manual interest or demographic targeting you could configure yourself. Lookalike audiences were among Meta's most powerful DTC targeting tools from 2015-2021, and still have significant value in specific contexts today.

Creating a Lookalike Audience

In Meta Ads Manager:

• Go to Audiences in the Business Suite
• Click "Create Audience" then "Lookalike Audience"
• Select your source audience (custom audience)
• Select target country/region
• Select lookalike percentage (1-10%)
• Click "Create Audience"

Processing time: 1-6 hours for Meta to build the lookalike. The audience remains active and updates as Meta's algorithm continues processing. Minimum source audience size: Meta requires at least 100 people in a source custom audience to create a lookalike, but recommends 1,000-5,000 for best results. More source data creates more accurate lookalikes.

Lookalike Percentage Sizes

When you create a lookalike, you select a percentage (1% to 10%) that determines how closely the lookalike matches your source audience.

General rule: 1-2% lookalikes are most similar and typically perform best. 3-5% offers more reach with somewhat lower similarity. 5-10% provides broad reach but weaker targeting signal.

For DTC brands with limited source audience size, testing 1% vs 2-3% stacked lookalikes is worthwhile. Some brands find wider lookalikes outperform tighter ones when the source audience is small or imperfect.

Best Source Audiences for DTC

The quality of your lookalike is directly determined by the quality of your source audience. Higher-value source audiences create better lookalikes.

Ranked by Typical Performance (Highest to Lowest)

1. High-value purchasers: Your top 20-30% of customers by LTV. This creates a lookalike of your best customers, not just average buyers. 2. All purchasers: Your complete customer list. The most common and reliably strong lookalike for DTC. 3. Initiated checkout audience (Pixel): People who started checkout but did not complete. High-intent users who closely resemble buyers. 4. Add-to-cart audience (Pixel): Somewhat broader than checkout initiators, still high-intent. 5. 30-day site visitors: Much broader intent signal; useful when customer list is small. 6. Email subscribers: Strong if your list is engaged and built from buyers/highly qualified leads. 7. Video viewers (50%+ of your best video): Users who watched half of your best-converting ad. Good creative-based signal. Avoid as source audiences: Broad page engagers (fans of your page who have never bought), follower lists with no purchase signal, purchased or unverified email lists.

Lookalike vs Broad Targeting in 2026

This is the most important context for understanding lookalike audiences today.

Meta's Advantage+ Audience (broad targeting) has significantly closed the gap with lookalike targeting since 2022. Meta's machine learning is now so effective that giving the algorithm open access to find buyers (broad targeting with no audience restrictions) often matches or outperforms lookalike targeting in Advantage+ Shopping Campaigns.

Why? Because Meta's algorithm already has all the data underlying your lookalike. When you create a 1% purchaser lookalike, you are essentially telling Meta to use data it already has. In Advantage+ Shopping, Meta uses that data automatically without you specifying it.

The practical implication:

• In standard manual campaigns: lookalikes are still valuable targeting signals
• In Advantage+ Shopping Campaigns: adding lookalike suggestions rarely improves performance significantly, and can restrict the algorithm

MHI Media's 2025 testing across 30+ DTC clients found that ASC campaigns with no audience suggestions outperformed ASC campaigns with lookalike suggestions in 68% of cases. Broad targeting via ASC is now the default recommendation.

Where lookalikes retain clear value:

• Standard manual campaigns outside of ASC
• Testing specific audience hypotheses
• International expansion where the algorithm has limited local data
• Niche products where broad audience would reach too many irrelevant users

Performance Data

Lookalike vs Broad Targeting Comparison (Manual Campaigns, DTC, 2025)

Lookalikes still edge out broad targeting in manual campaigns. In ASC campaigns, the gap reverses.

Source Audience Quality Impact

The data confirms: source audience quality is the primary lookalike performance driver.

When to Use Lookalikes

Standard manual campaigns: Use 1% purchaser lookalikes as your primary cold audience targeting. Stack with 2-3% for additional reach. New markets/geographies: When your brand enters a new country where the algorithm has limited pixel data, a purchaser lookalike from your primary market can seed initial targeting. Niche products: For products with narrow demographic fit, a tight purchaser lookalike constrains the algorithm to find genuinely relevant buyers faster than broad targeting. Testing creative with controlled audiences: When you want to isolate creative performance without audience variation, a consistent lookalike audience provides a stable testing environment.

Common Mistakes

Using too broad a source audience: An audience of 100,000 general page engagers creates a weak lookalike. Quality over quantity in source audiences. Small source audience: Below 500 people, lookalike accuracy is significantly reduced. Build your customer list before relying heavily on lookalikes. Over-using lookalikes in ASC: Adding 5 lookalike suggestions to an ASC campaign limits the algorithm's flexibility. Use zero or one suggestion in ASC. Ignoring geographic accuracy: Lookalikes are country-specific. A US purchaser lookalike does not perform in the UK without creating a separate UK lookalike using UK or similar purchaser data. Running lookalike and retargeting in the same ad set: Your 1% purchaser lookalike may overlap with your existing customer base. Exclude purchasers from your lookalike campaigns to keep prospecting and retargeting separate.

Key Takeaways

• Lookalike audiences find new users similar to your best customers using Meta's machine learning
• 1% purchaser lookalikes are the most effective source for DTC brands
• In Advantage+ Shopping Campaigns, broad targeting often outperforms lookalike suggestions
• Source audience quality is the primary performance driver; use purchasers, not general engagers
• Lookalikes retain strong value in standard manual campaigns and new market expansion

FAQ

How many people do I need in my source audience?

Meta requires a minimum of 100 people but recommends 1,000-5,000 for good accuracy. The more purchasers you have in your source audience, the more data Meta has to build an accurate lookalike. Below 500 purchasers, consider combining purchasers with initiated checkout users to increase source audience size.

Do lookalike audiences update automatically?

Yes. Meta's lookalike audiences refresh every 3-7 days based on updates to your source custom audience. If you upload a new customer list, the lookalike will update within the next refresh cycle.

Should I use a 1% or 3-5% lookalike?

Start with 1% for highest accuracy. Test 2-3% if you need more reach or if your 1% lookalike is saturating quickly (high frequency). Some DTC brands find stacking 1%, 2%, and 3% lookalikes in separate ad sets works well for scaling.

Are lookalike audiences still effective after iOS 14?

Lookalike audiences have been affected by iOS 14 because they are built from pixel data, which became less complete after the ATT framework. However, they remain valuable, especially when built from server-side event data via the Conversion API. CAPI-based lookalikes are more accurate than pixel-only lookalikes in a post-iOS 14 environment.

Can I create a lookalike from a competitor's customers?

No. You cannot create a lookalike from a competitor's customer data. You can only use custom audiences from your own data: your pixel events, your customer list, your app events.

---

MHI Media builds lookalike audience strategies for DTC brands scaling Meta spend. Get a free audit.