How to Build Lookalike Audiences That Actually Convert

Lookalike audiences are one of the most powerful prospecting tools in Meta's advertising platform. When built correctly, they let you find new customers who share characteristics with your best existing ones — at scale and at a cost that often beats every other prospecting method available.

But there's a significant difference between a lookalike audience that drives profitable growth and one that wastes your budget. This guide covers how to build the former.

How Lookalike Audiences Work

A lookalike audience starts with a "seed" audience — a group of people you define, such as your existing customers, website converters, or high-engagement users. Meta's algorithm analyses this seed audience across hundreds of data points (demographics, interests, behaviours, platform activity) and finds other users who share similar patterns.

The result is a new audience of people who've never interacted with your brand but who resemble your best prospects. You control the size via percentage: a 1% lookalike in the UK represents roughly the top 440,000 most similar users; a 5% lookalike represents about 2.2 million.

Choosing the Right Seed Audience

The seed audience is the single most important factor in lookalike performance. Garbage in, garbage out.

Best Seed Audiences (Ranked by Effectiveness)

• Top customers by LTV: Your highest lifetime value customers. This tells Meta to find people who don't just buy once — they buy repeatedly and at high value.
• Recent purchasers (last 90 days): Fresh data produces more relevant patterns than old data. A purchaser from last month is more representative of your current customer than one from two years ago.
• High-value purchasers: Customers whose orders exceeded a certain threshold (e.g., £100+). Filters out bargain hunters and discount-only buyers.
• All purchasers: Still effective but less targeted than the options above. Good for brands with smaller customer lists.
• Engaged email subscribers: People who open and click your emails regularly. High engagement indicates genuine interest.
• Video viewers (75%+ watched): People who watched most of a product video have shown strong intent.

Weak Seed Audiences to Avoid

• All website visitors: Too broad. Includes bounces, accidental clicks, and bots. The resulting lookalike lacks focus.
• Page likes / followers: Often accumulated through engagement campaigns or organic growth that may not represent your ideal customer.
• Very old customer data: If your customer list is primarily from 3+ years ago, those patterns may no longer be relevant.

Lookalike Percentage: How Wide to Go

The percentage controls the trade-off between similarity and reach:

1% Lookalike

• Highest similarity to your seed
• Smallest audience size (~440K in the UK)
• Best for: Conversion campaigns with limited budgets, initial testing
• Typical performance: Lowest CPA, highest conversion rate

1–3% Lookalike

• Good balance of similarity and reach
• Medium audience size (~880K–1.3M in the UK)
• Best for: Scaling campaigns that have proven product-market fit
• Typical performance: Slightly higher CPA than 1%, but significantly more reach

3–5% Lookalike

• Broader audience with weaker similarity signal
• Large audience size (~1.3M–2.2M in the UK)
• Best for: Awareness campaigns, creative testing with large budgets
• Typical performance: Higher CPA, but useful for scale and prospecting at the top of funnel

5–10% Lookalike

• Very broad — approaching general population characteristics
• Best used as: An Advantage+ audience suggestion rather than a hard targeting constraint
• Typical performance: Marginal improvement over fully open targeting

Advanced Lookalike Strategies

Stacking Lookalikes

Create multiple lookalike audiences from different seed sources and layer them:

• 1% LAL from purchasers AND 1% LAL from high-LTV customers AND 1% LAL from engaged email subscribers
• Users who appear in multiple lookalikes are especially strong prospects
• Meta's algorithm will naturally prioritise users who match multiple signals

Exclusion-Based Lookalike Laddering

For systematic scaling, create tiered campaigns:

• Campaign 1: Target 1% LAL, exclude nothing
• Campaign 2: Target 2% LAL, exclude 1% LAL
• Campaign 3: Target 3% LAL, exclude 2% LAL

This ensures no audience overlap between campaigns and lets you set different CPAs and creative strategies for each tier.

Value-Based Lookalikes

When uploading a customer list, include a "value" column representing each customer's total purchase value. Meta will weight higher-value customers more heavily in the lookalike algorithm, producing an audience more likely to become high-value customers themselves.

Combining Lookalikes with Interest Targeting

For highly niche products, narrow your lookalike with interest layers:

• Start with a 1–3% lookalike
• Add an interest or behaviour restriction that's core to your product
• This reduces reach but increases relevance for markets where the lookalike alone might be too broad

Common Lookalike Mistakes

• Refreshing too frequently: Updating your seed audience weekly doesn't give the algorithm time to learn. Refresh seed audiences monthly or quarterly.
• Too many overlapping lookalikes: Running five campaigns targeting similar lookalikes fragments your budget and increases auction competition against yourself. Use the audience overlap tool to check.
• Ignoring creative: A perfect lookalike audience still needs compelling creative to convert. Your ad is the bridge between audience match and action.
• Not excluding existing customers: Always exclude your customer list from prospecting lookalikes. Without exclusions, the algorithm may spend heavily on people who'd buy anyway.

Lookalikes in an Advantage+ World

With the rise of Advantage+ campaigns and broad targeting, some advertisers wonder if lookalikes are still relevant. The answer is yes — but their role has shifted.

Rather than serving as hard targeting constraints, lookalikes now work best as:

• Advantage+ audience suggestions — giving the algorithm a starting point
• Standalone prospecting in accounts with limited conversion data
• Testing tools to validate which customer segments drive the best new customers

Building high-converting lookalike audiences is part art, part science. If you want expert help optimising your prospecting strategy, our Meta Ads management team works with brands daily to build and refine lookalike strategies that drive profitable growth. Book a free consultation via our Calendly to discuss your approach.

Frequently Asked Questions