A core audience is a Meta ads audience you build from people’s characteristics rather than from your own customer data: location, age, gender, language and detailed targeting options such as interests and behaviours. When you save one so it can be reused across ad sets, Meta calls it a saved audience.
How a core audience works
You define the audience in the ad set or in the Audiences area of Ads Manager. A Brighton dog groomer might choose people within ten miles of BN1, aged 25 to 65, with an interest in dogs or pet care. Meta estimates how many accounts fit and shows ads to people inside that boundary.
Core audiences sit alongside the two other audience types. A custom audience is built from people who already know you, such as site visitors or a customer list. A lookalike audience finds people who resemble a source audience. Core audiences need no data from you, which is why they are usually the first thing a new advertiser sets up.
How the role has changed
Meta has steadily moved away from treating your targeting choices as strict limits. With Advantage+ audience switched on, your age range and interests become suggestions the system may go beyond if it finds better prospects. Location, minimum age and exclusions usually stay as firm controls. At the time of writing (October 2026), whether a setting is a limit or a suggestion depends on the campaign type and which options you choose, so check the ad set rather than assuming.
Some advertisers also face restrictions. Ads declared in a special ad category, such as employment, housing or financial products, get fewer targeting options, and Meta sets which limits apply by category and country.
Why it matters
For a local business, the location part of a core audience is often the most valuable setting in the whole account. A plumber in Croydon gains nothing from impressions in Newcastle, and drawing the location targeting around the area you actually serve protects the budget from the start.
Interest targeting is less dependable than it once was. Meta has removed many sensitive interests over the years, and interest labels are inferred from behaviour, so they are often broad or wrong. For many accounts with good conversion tracking, a wide core audience with few interests now performs as well as a carefully layered one, because the delivery system finds the buyers itself.
Common mistakes
- Stacking interests until the audience is tiny. Very narrow audiences push up costs and run out of fresh people quickly.
- Drawing the area too wide. A whole region when you serve a handful of postcodes spends money on people who can never become customers.
- Assuming interests are precise. “Interested in yoga” includes people who once liked a yoga video.
- Never revisiting saved audiences. Interests get renamed or removed, and an old saved audience may no longer be what you think.
How to act on it
Start with what must be true about every customer: where they are and how old they must be to buy. Set those firmly. Then test a broad version with no interests against one or two interest-led versions over a few weeks, judging them on cost per conversion rather than click costs.
If you serve a defined area, check the location map in Ads Manager before launch and exclude areas you do not cover. I set up and test audiences this way for local clients through my Facebook ads for local businesses.