Lookalike audiences are a powerful tool for prospecting, designed to help you find new customers who share characteristics with your existing ones. By taking data from a “seed audience”, a group of users you already know, Meta’s algorithms can identify a new, broader audience with similar demographics and behaviours. This allows you to expand your reach to a relevant audience without manually defining every trait of your ideal customer. A mailing list is noted as a very effective source for building a seed audience.

Lookalike audiences are a core component of a modern audience prospecting strategy. The key rule for prospecting is to test different audience types against each other, specifically interest, lookalike, and broad audiences, to see what works best for your business.

How do lookalikes work?

Lookalike audiences function by using a “seed audience” as a foundation. Meta’s algorithm looks at the demographic and interest trends of this seed audience. It then finds a new group of similar people to target. In essence, the algorithm replicates the characteristics of your seed audience into a new one.

For example, if you create a lookalike audience from a list of users who have visited your website, Meta will find users who share similar traits to those visitors. The platform’s powerful algorithms have access to a wealth of data, allowing them to find these similar users with high accuracy.

The percentage you choose for a Lookalike audience is a key factor in its size and similarity to your seed audience. While the general rule for Meta audiences is “the bigger the better”, a balance between size and similarity often works best for lookalike audiences. A good strategy is to trial different percentages, such as 3%, 6%, and 8%, to find the right balance for the best Cost Per Acquisition (CPA).

What’s a seed audience?

A seed audience is the original group of people used to create a Lookalike audience. The quality and relevance of this seed audience are crucial, as Meta’s algorithm analyses its characteristics to find new, similar users. A seed audience can be made from either First Party or Third Party data.

First-Party Data This is data you have collected directly from your customers or website visitors.

• Customer Lists: You can upload a list of your existing customers directly into Meta’s Ads Manager. A mailing list is an excellent source for creating a seed audience as it contains engaged users who have opted-in to receive communication from your brand.

Third-Party Data This is data collected by an outside source, in this case, Meta itself.

• Website Pixel Data: If you have a Meta pixel installed on your website, you can create a seed audience from people who have visited your site or taken a specific action, such as viewing a product or making a purchase.
• Engagement Data: While not explicitly detailed as a seed source for Lookalikes in the provided text, the concept of using engagement data from ads or other social interactions could be considered a form of third-party data that can be used to create custom audiences, which can then serve as a seed.

Understanding lookalike percentages

The size and similarity of a Lookalike audience are determined by a percentage you set. This percentage ranges from 0% to 10%.

• A 0% lookalike audience is the smallest and most similar to your seed audience.
• A 10% audience is the largest and least similar to your seed audience.

While the general rule for Meta audiences is “the bigger the better” , a balance between size and similarity often works best for lookalike audiences. A good strategy is to test different percentages, such as 3%, 6%, and 8%, to find the right balance for the best cost per acquisition (CPA). For example, a 3% lookalike audience will be larger and less similar to the seed audience than a 1% audience, while a 5% lookalike audience might have an estimated size of 2,300,000-2,700,000 users.

How to set up a lookalike audience

Setting up a lookalike audience is a straightforward process within Meta’s Ads Manager. It starts with having a high-quality “seed audience.” This can be a customer list you upload, or data from your Meta pixel based on website visitors who have completed a specific action. Meta will then analyse the demographic and interest trends of your seed audience and find new users who share similar characteristics to target with your ads.

To create a Lookalike audience:

1. Navigate to the Audiences in your Meta Ads Manager.
2. Choose “Create Audience” and then select “Lookalike Audience”.
3. Select your seed audience from the available sources, such as a pixel, custom audience based on website visitors or a customer file.
4. Specify the audience location (e.g., United Kingdom, Ireland, United States) and the desired percentage range (e.g., 1%, 3%, 6%, 8%). Don’t worry, this isn’t targeting users. It’s just setting up the range of your audience. So, if you’re based in Northern Ireland, select United Kingdom.
5. Create the audience. Once it’s ready, it will be available for use in your ad sets. The availability status will be listed as “Ready” in the audience table.

By leveraging Lookalike audiences, you can effectively scale your campaigns by reaching new people who are highly likely to be interested in your products or services, based on the success of your existing customer base.

Tips when using lookalikes

In a nutshell, Lookalikes take your best data and turn it into your best growth tool, if you feed the algorithm with high-quality signals and test your percentages properly. They’re especially powerful when paired with a strong creative that speaks directly to the intent or problem your seed audience represents.

• Quality over quantity with your seed audience. A list of 1,000 high-value customers will usually outperform 10,000 random newsletter sign-ups.

• Avoid overlapping lookalikes. If you’re testing 3% and 6% audiences, exclude the 3% from the 6% to keep them clean and avoid bidding against yourself.

• Keep your seed audience fresh. If your customer list is months old, update it regularly, or use dynamic audiences like website traffic or email segments synced via Klaviyo.

• Stack lookalikes smartly. You can combine lookalikes from multiple seed sources in one ad set, as long as they’re aligned in intent (e.g. multiple customer types).