A custom audience is Meta's way of letting you reach, or exclude, people who already know you. Unlike interest targeting, it starts from your own data or your own content, which is why it powers retargeting, buyer exclusions and lookalike audiences.
The main types
- Website: people who visited pages or triggered events such as ViewContent, AddToCart or Purchase, as reported by the Meta Pixel and Conversions API. At the time of writing, these can look back up to 180 days.
- Customer list: emails, phone numbers and other identifiers you upload, which Meta matches to accounts.
- Engagement: people who watched your videos, interacted with your Instagram account or Facebook Page, opened a lead form, and similar actions.
- App and offline activity: actions in your app, or in-store purchases you send to Meta.
How to build a useful starter set
- Purchasers, last 180 days: to exclude from prospecting when you want new customers, and to seed lookalikes.
- Cart and checkout abandoners, last 7 to 14 days, excluding purchasers: your core retargeting pool.
- Video viewers and Instagram engagers: warm audiences for storytelling and proof.
- Full customer list with a value column: for value-based lookalikes.
Name each audience clearly, with its source and window, so anyone working in the account knows what it holds.
Privacy, hashing and consent
When you upload a customer list in Ads Manager, identifiers are hashed before they leave your browser: converted into irreversible codes with SHA-256, so Meta can match people without receiving the raw details. If you send lists or events through the API, you normalise the data (for example, lowercase and trimmed emails) and hash it yourself. Meta's customer list terms explain how hashed data is used and deleted, so read them before your first upload.
Hashing protects data in transit. It isn't consent. You still need a lawful basis to use customers' details for advertising, for example under India's Digital Personal Data Protection Act, 2023, or GDPR in the UK and EU. In practice that means a clear privacy notice, consent where the law requires it and honouring opt-outs promptly. Some categories, such as health, face extra Meta restrictions on the data they can share and the audiences they can build.
Common mistakes
- Windows that are too long. A visitor from five months ago is barely warmer than a stranger.
- Not excluding buyers from abandoner and prospecting campaigns, so you pay to re-sell to people who just bought.
- Stale lists. An uploaded list is a snapshot. Refresh it regularly or connect your store so it updates.
- Single-identifier lists. Adding phone numbers, names and locations alongside email improves match rates, the same logic behind event match quality.
For the tracking that feeds website audiences, see the Meta Pixel and Conversions API setup checklist.
Frequently asked questions
What is the difference between a custom audience and a lookalike audience?
A custom audience contains people who already interacted with you, such as visitors, customers or engagers. A lookalike audience contains new people Meta finds who resemble a custom audience you choose as the source. You retarget with the first and prospect with the second.
Is it legal to upload customer emails to Meta?
It can be, if you collected the data lawfully and have a valid basis to use it for advertising under the laws that apply to you, such as India's DPDP Act or GDPR. Meta hashes the data before matching, but hashing doesn't replace consent or a clear privacy notice. Take legal advice if you're unsure.
Why is my customer list match rate low?
Usually because the list has only one identifier, such as email, or the data is out of date or badly formatted. Add phone numbers with country codes, names and locations where you have them, formatted the way Meta's uploader asks. Personal email addresses are more likely to match than work ones, because most people sign up to Facebook and Instagram with a personal address.
