How do I create a lookalike audience for meta ads?

Creating a lookalike audience for Meta Ads is one of the most powerful ways to scale your campaigns by reaching new people who are similar to your best existing customers. As an ad agency that has built its reputation on scaling profitable Facebook campaigns, we at Sagum have refined this process into a repeatable system. Here’s how to do it right.

Step 1: Define Your Source Audience

The foundation of any great lookalike audience is the source. This is the seed group of people you want Meta to model after. The quality of this source is everything-garbage in, garbage out. Your source should be a Custom Audience built from your highest-value data. The best sources are typically:

• Customer lists (from your CRM or email list): These are people who have already bought from you. Prioritize lists of your best customers (e.g., high lifetime value, repeat purchasers).
• Website traffic (via the Meta Pixel): Create a custom audience of people who visited key pages, such as a product page or checkout, within the last 30 to 180 days.
• App activity: If you have an app, use events like “purchase” or “add to cart” as your source.
• Engagement audiences: People who have interacted with your Instagram or Facebook content (e.g., watched 50% of a video, liked a post). These are less powerful than purchase-based sources but can work well for top-of-funnel prospecting.

Pro tip: A minimum of 1,000 people from the same country in your source audience is required for Meta to generate a lookalike. Aim for 1,000 to 50,000 for the best balance of scale and accuracy.

Step 2: Choose the Right Lookalike Size

When you create a lookalike audience, Meta asks for a percentage size (1% to 10%). This determines the breadth of the audience relative to the country’s total population. Here’s how we advise clients to think about it:

• 1% lookalike: The most similar to your source audience. It’s highly targeted, with the strongest likelihood to convert, but the smallest reach. Use this for bottom-of-funnel campaigns or when your source is extremely high quality (e.g., high-LTV customers).
• 3% to 5% lookalike: A good middle ground. It balances similarity with scale. We often start here for prospecting campaigns because it’s still quite relevant but allows you to reach more people.
• 10% lookalike: The broadest and least similar. This is essentially a scaled-down version of a broad audience. Useful for top-of-funnel awareness or when you need massive reach, but expect lower conversion rates.

Start with a 1% lookalike from your best customer list. If you’re seeing great results and need more volume, layer in a 3% lookalike. Never use a 10% lookalike as your sole audience unless you have a very small market.

Step 3: Create the Lookalike in Meta Ads Manager

Here’s the technical process, which we’ve executed hundreds of times:

1. Go to Meta Ads Manager and open the Audiences tool (under the “All Tools” menu).
2. Click the Create Audience dropdown and select Lookalike Audience.
3. In the Source field, choose your pre-built custom audience (e.g., “Website Purchases – Last 30 Days”). If you haven’t created it yet, you’ll need to do that first under “Custom Audiences.”
4. Select the Target Country (or countries) where you want the lookalike. Keep it to one country per lookalike for best performance-multi-country lookalikes dilute relevance.
5. Set your Audience Size (1% to 10%). We recommend starting with 1% or 3%.
6. Click Create Audience. Meta will then process the data, which can take anywhere from 6 to 24 hours to fully populate. You can use the audience immediately, but it may be incomplete until then.

Step 4: Optimize Your Campaign for Lookalike Performance

Creating the audience is only half the battle. To get the most out of lookalikes, you need to integrate them into a well-structured campaign. Follow these best practices from our playbook:

• Use lookalikes in dedicated ad sets: Don’t mix a lookalike with other targeting (like interest-based) in the same ad set. Give the lookalike its own ad set with a clear budget so you can measure its performance directly.
• Match your creative to the audience: A 1% lookalike may respond to a more direct-response offer, while a 5% lookalike might need more top-of-funnel brand messaging. Test different creative within each lookalike ad set.
• Layer with exclusions: Always exclude your existing customers or converters from lookalike ad sets. You don’t want to waste budget showing ads to people who have already bought. Create a custom audience of all converters and exclude it.
• Refresh your source regularly: Audiences decay over time. Update your source custom audience (e.g., “Website Purchases – Last 30 Days”) every month to keep the lookalike fresh. Meta will automatically refresh the lookalike if the source updates.

Common Pitfalls to Avoid

We’ve seen many advertisers fail with lookalikes because they overlook these details:

• Using a weak source: Don’t use a video-view custom audience as your primary source. It’s too broad. Stick to purchase-based or high-intent signals.
• Ignoring size: A 1% lookalike might be too small for your budget. If you’re spending $1,000/day, you may need a 3% or 5% lookalike to avoid over-delivering to the same people.
• Not testing multiple lookalikes: Create 3 to 5 lookalike audiences from different sources (e.g., one from email list, one from website purchases, one from add-to-cart). Let them run in separate ad sets and see which performs best.

Final thought from our experience: Lookalike audiences are a proven way to scale, but they aren’t a set-it-and-forget-it tactic. The most successful clients we work with treat lookalikes as living assets-constantly testing new sources, adjusting sizes, and refining creative. When done right, they become one of the most reliable engines for profitable growth in Meta Ads.