How to Create Meta Custom Audiences and Lookalike Audiences

A custom audience is built from eligible data sources representing people who have a particular relationship with your business. A lookalike audience is an approach that helps you reach new people who resemble the source audience you select. Their purposes differ: one can build on an existing relationship, while the other can support new customer discovery. Neither guarantees you will reach everyone on a list or that similar people will buy.

Choose the purpose and data source

First answer one question: do you want to inform people who have already shown interest, or reach new people based on successful customers? For example, showing new content to people who viewed a product page without purchasing is different from a repeat customer campaign. Making this distinction early simplifies source selection and messaging.

Depending on the options available in your account, you may use website events, eligible customer lists, video views or Facebook and Instagram engagement. Check the connections and usage conditions for each source. Having access to an Instagram account does not mean every ad account can use its engagement data. Business asset connections and permissions matter.

Prepare the customer list

If you use a customer list, work only with records that are suitable for sharing and advertising use. Do not choose a purchased list, a list of unclear origin or one collected for another purpose simply because it is large. Leave out sensitive information and unnecessary columns. Review the required identifier formats in the current template.

1. Remove duplicate records for the same person.
2. Make fields such as email addresses and phone numbers consistent.
3. Separate invalid or outdated records.
4. Define the list by its business purpose, such as eligible existing customers.
5. Check field mapping during upload and correct any reported errors.

Because of matching, the number of rows in your file may differ from the audience available for advertising. Do not respond to low matching by collecting additional personal data. First inspect file formatting, source quality and the accuracy of existing records.

Create an engagement or website audience

Open the option to create a new custom audience in the audience tools within Ads Manager. Select the relevant source and define suitable conditions. For a website audience, verify that you selected the correct data source and that the event you want to use is actually being received. For engagement audiences, select the correct Page or Instagram account. Menu order can vary by account.

For an example application, you could assess people who viewed products in the past 30 days. Thirty days is not a universal recommendation: the purchase decision period and audience volume matter. An expensive service may have a long consideration period, while a frequently consumed product may have a shorter relationship window. Configure conditions so they do not contradict one another.

Make audience names descriptive by including the source, behavior and time period. For example, Web_ProductView_30Days can help your team understand what is being used. This naming approach is not a technical requirement. If you change the rules later, update the name or documentation too.

Choose a strong source for a lookalike audience

In the lookalike audience creation option, select an eligible source audience. Consider the geography where you want to reach new people and the size options shown on screen. Eligibility requirements and minimums for source audiences can change, so use the current interface as your reference. A smaller or larger similarity range is not by itself a quality guarantee.

The business value of the source audience matters. Actual purchasers and people who briefly watched a video do not provide the same signal. However, a very narrow, outdated customer group or one representing different products may also be unsuitable for your current offer. Examine the behavior represented by the source as well as its size.

Verify campaign usage and results

When using the prepared audience in an ad set, read how audience automation works. In some setups, a custom or lookalike audience may be treated as a suggestion, and delivery may expand beyond it. If a business purpose requires reaching only a specific group, verify whether the available controls support that requirement.

In an example test plan, you could prepare a product explanation for an existing engagement audience and an introduction to the main benefit for new people. Do not compare the groups with identical expectations, because their familiarity with the business differs. Track spend, completed outcomes and lead quality. Estimated size or creation status does not demonstrate that advertising generated sales.

Common mistakes and final checklist

Common mistakes include creating an audience without checking the source, using purchasers in an inappropriate campaign and treating a similarity percentage as a guaranteed success formula. A custom audience is not always a complete customer list. Data receipt, matching, eligibility and delivery are different stages. When you find an issue, identify the affected stage.

• Is the audience's purpose clear?
• Are the source and permissions correct?
• Have customer record usage conditions been checked?
• Are the event, behavior and time period selected correctly?
• Does the source audience reflect current business value?
• Are campaign suggestions and controls understood?
• Is there a reporting plan for evaluating outcome quality?