Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Glossary term

Custom Audience

An ad audience built from the advertiser's own sources: customer lists, site visitors or engagers.
Paid Mediaplatform featureUpdated

Definition

Meta's Business Help Centre describes a custom audience as an ad targeting option that lets an advertiser find its existing audiences among people across Meta technologies. It is built from one of four broad source types:

  • Website and app activity. People whose browsers or devices fired the advertiser's pixel or app events, or whose events arrived through the Conversions API, filtered by event, page and a retention window.
  • Customer lists. Files of identifiers such as email addresses and phone numbers, uploaded by the advertiser and matched against Meta accounts.
  • Catalogue activity. People who viewed or acted on products from the advertiser's catalogue.
  • Engagement. People who interacted with the advertiser's pages, profiles, videos, lead forms and other content on Facebook and Instagram.

Each source has its own rules for what can be captured and how long people stay in the audience, and the custom audiences guide sets those out source by source.

In practice

Matching. A customer list is hashed before it leaves the advertiser's environment and compared with Meta's own hashed identifiers; only the matches join the audience. Not every person on a list has an account that can be matched, so the audience is smaller than the file, and adding further identifier columns for the same people improves the chance of a match.

Size. Small audiences are hard to deliver to efficiently and show a floor label rather than a count. Website audiences are limited by how many visitors the pixel and server events can actually record, which consent choices and browser restrictions reduce.

Consent. The advertiser must have the right to use the data for this purpose and must declare where a list came from. The hashing protects the data in transit; it does not remove the obligations that come with it.

Illustrative example — not a real result

A sofa retailer stops prospecting to last month's buyers

Haverleigh Sofa Company spends £12,000 a month on broad prospecting campaigns. Its order system shows that 640 of the 2,100 purchases attributed to those campaigns last quarter came from people who had already bought a sofa in the previous 60 days, many of them buying a matching footstool. The retailer uploads its purchaser list each week as a customer list audience, which matches 4,800 of its 7,000 recent buyers, and excludes it from prospecting. Last quarter's £36,000 bought 1,460 new customers, about £24.70 each; with the same spend the next quarter brings 1,680, about £21.40 each. The matched list is reused in a small separate campaign that offers accessories to recent buyers.

Targeting, exclusion and seeding

Custom audiences do three different jobs, and confusing them causes most setup mistakes.

Targeting reaches the people in the audience directly, which is the basis of Remarketing: showing ads to people who visited, engaged or bought. Retention windows and event filters decide how warm each audience is.

Exclusion keeps the audience out of a campaign. Excluding recent purchasers from prospecting, or people already in a nurture sequence from cold ads, stops spend on people the campaign was not meant for. As Meta's automated audience options widen targeting by default, exclusions are one of the few audience controls that still bind; how Advantage+ audiences treat custom audiences as suggestions rather than limits is explained in that guide.

Seeding uses the audience as the source for a Lookalike Audience, so the quality of the custom audience decides the quality of the lookalike.

For campaigns aimed at consideration, the engagement audiences a campaign builds are often its main output, a point the Meta consideration setup makes when it treats video viewers and profile visitors as the audience the next campaign reaches.

Equivalents on other platforms

Other platforms offer the same idea under other names: customer list matching and first-party audience lists in Google's tools, and first-party audiences in demand-side platforms. In Display & Video 360 the advertiser's own lists sit beside Google audiences and third-party segments, as the DV360 audiences guide describes. The mechanics differ, but the three jobs of reaching, excluding and seeding carry across.

Remarketing is the most common use of a custom audience. Lookalike Audience is built from one. The Conversions API feeds website audiences with server-side events that browser restrictions would otherwise lose. Segmentation is the planning discipline behind them: deciding which groups of known people deserve different messages, or none at all. The wider Meta targeting section covers how custom audiences fit with interest and automated targeting, and a Meta Ads account audit checklist includes checks that each audience is populating and that exclusions are in place.

Common confusions

  • Custom audience as remarketing. Remarketing is one use; exclusion and seeding are the others.
  • List size as audience size. Only matched people join, so the audience is smaller than the upload.
  • Hashing as consent. Hashing protects identifiers in transit; the right to use the data must exist anyway.
  • Custom audiences as hard limits everywhere. Under some automated audience settings they act as suggestions, while exclusions still apply.
  • Set and forget. Lists go stale and windows expire; customer lists need refreshing from current data.

More terms to look up

Where this term is used

Sources

  1. About Custom Audiences — Meta Business Help Centre

Related