Facebook Ad Location Targeting (2027 Guide)

The six levels of Facebook location targeting

Location lives in the ad set, inside the Audience section you fill out when you build a Facebook ad from scratch, and it is one of the few hard boundaries Meta still lets you draw by hand. Everything else in that panel (interests, behaviours, lookalikes) is increasingly a suggestion the delivery system can look past. Location is not. If you set a country, the ad does not deliver outside it. So the first job is knowing the levels available, because each one answers a different business question, from "the whole nation" down to "the three streets around my shop".

There are six, and they nest inside each other:

The rule that trips people up is the radius column. A radius only draws a circle around a point or, in a limited way, around a named city. Regions, postal codes, and countries target the whole named area as it is, so there is no slider to widen or shrink them. If you want a circle, you need a pin or a city marker, not a postal code.

Two levels deserve a note. A DMA (Designated Market Area) is a US-only construct: Nielsen divides the country into 210 non-overlapping media markets, each a cluster of counties that share dominant local TV stations, and Meta lets you target them directly. That matters because a metro's real economic footprint often spills across state and county lines that a city marker misses. And the "region group" option at the top is broader than a single country: you can select a continent or a trading bloc in one click, which is how a brand that ships across a whole region avoids adding twenty countries by hand.

Radius targeting: how the pin drop and the slider really work

Radius targeting is the tool local advertisers reach for most, and it is also the one most often misused. Here is the mechanism. When you choose Drop Pin, you place a point on the map, Meta reads its coordinates, and it draws a circle you size with a slider or by typing a number. In the United States the radius runs from a 1-mile minimum to a 50-mile maximum, and in markets that display metric units it runs from 1 km to about 80 km (the metric equivalent of that 50-mile cap). A freshly dropped pin defaults to a 1-mile radius, which is deliberately tight, so widening it is a conscious choice rather than an accident.

A named city behaves differently from a pin, and the difference confuses people. When you select a city, Ads Manager starts you at a radius roughly the size of that city and will not let you shrink the circle below the city's own footprint. That is why you cannot select "New York" and then target a single mile of it. To go tighter than a whole city, you abandon the city marker, drop a pin on the exact street address, and set a small radius from there. Pin for precision, city marker for coverage.

Three limits shape what a radius can actually reach:

• It stops at national borders. A pin placed near a frontier draws a circle that is clipped at the boundary. Coverage never crosses into a neighbouring country, so a 50-mile radius on a border town covers far fewer people than the same radius inland.
• It needs a viable audience. A tight radius in a low-density area can leave too few people to spend against efficiently. When the pool is small, delivery gets expensive or the system leans on the "recently in" rule to fill it.
• The slider is not surgical. The circle catches everyone inside it, not a hand-drawn service boundary. If your catchment is an odd shape (one side of a river, a single suburb of a sprawling metro), a circle will always over-reach on some edge. Postal codes or exclusions do that job better than a bigger or smaller radius.

A worked example makes the scale concrete. A single-location dental clinic drops a pin on the practice and sets a 10-mile radius. In a metro of three million that circle might surface an eligible audience of a few hundred thousand people, which is far more than one clinic's ad budget can ever reach, so the binding constraint is spend, not audience size. The mistake would be widening the radius to 30 or 40 miles "to be safe": that only pours budget into neighbourhoods the patient will never drive from. For a fixed-location business, a tighter circle is almost always the more profitable one.

"Living in or recently in": why you cannot isolate residents anymore

This is the single most important change in Facebook location targeting, and most guides written before it still describe a menu that no longer exists. For years, the location box carried a presence dropdown with four choices, and picking the right one was a core local-targeting skill:

• Everyone in this location: residents and visitors, the broadest option.
• People who live in this location: residents only, inferred from where Meta places their home.
• People recently in this location: anyone whose most recent device location was the area, which pulled in visitors and travellers.
• People travelling in this location: people in the area but a long way from where they live, the tourist setting.

On May 1, 2023, Meta quietly removed that dropdown. Every location set now uses one combined setting: "living in or recently in this location." There is no menu to change, and the three other options are gone. Search Engine Land documented the rollout, and Meta shipped it without an announcement, which is why so many advertisers discovered it only when their local results shifted.

The practical effect is blunt: you can no longer buy residents on their own, and you can no longer buy tourists on their own. A plumber or other home-services trade who used to select "people who live in this location" now unavoidably pays to reach recent visitors who have no use for a local tradesperson. A hotel that used to select "people travelling in this location" now cannot isolate travellers from the locals who live down the road. Both businesses lost a lever, and the only response is to work the other controls harder: a tighter radius, sharper exclusions, and creative that self-selects (a message that only a resident, or only a visitor, would respond to).

Why your ads reach people outside your area

This is the perennial question, and after the 2023 change the honest answer has several parts. Understanding them stops you chasing a setting that does not exist.

The "recently in" rule is doing exactly what it says. Your set now includes anyone whose recent location was your area, not just people who live there. Someone who visited last week, drove through, or connected from a device that pinged your city can qualify. This is the biggest single source of "why am I reaching outsiders", and it is not a bug you can switch off.

Location is inferred, not declared. Meta places people using device GPS, IP address, Wi-Fi and connection data, and profile signals, none of which is perfect. A phone on a work VPN, a stale home city on a profile, or an IP that resolves to the wrong suburb all put a person in the wrong bucket. Expect a small percentage of delivery to land just outside your intended edge for this reason alone.

A radius is a circle, not your service map. If your true catchment is irregular, the circle will always spill over the parts that stick out. That is geometry, not a targeting fault.

A thin audience forces the system to widen. When your area is small and your budget wants more results than the pool can supply, Meta leans harder on the "recently in" population and the edges of your radius to find deliverable people. A slightly bigger, cleaner audience often reaches fewer irrelevant people than a razor-thin one that the system has to stretch.

It is usually not Advantage+. A common misconception is that automation is ignoring your location. It is not. Advantage+ audience treats interests and custom audiences as suggestions, but location, age, and gender remain hard limits it must obey. If your ad is delivering outside your country, that is a setup error somewhere in the location box, not the algorithm overriding you. Inside your set area, though, the recent-visitor rule still applies, because that is part of how a location is defined rather than part of the audience automation.

The fix is rarely one switch. Tighten the radius to your real catchment, add exclusions for the zones you never want (see below), and let your creative carry some of the filtering: an offer that names the neighbourhood, or a message that only makes sense to a local, quietly discourages the wrong clicks even when the wrong people are technically eligible.

Location exclusions: carving out where you do not want to spend

Exclusions are the most underused control in the whole panel, and they solve problems the include list cannot. In the location box, flip the toggle from Include to Exclude, then add any level (a country, a region, a city, a postal code, or a pinned radius) that you want removed from delivery. The logic is strict and useful: exclusions win. Anyone inside an excluded area is dropped even if a broader included area covers them, so you can target a whole metro and cut a specific district out of the middle of it.

A few patterns where exclusions earn their place:

• You cannot serve everywhere you can reach. A courier that delivers to a city but not to one gated island, or a service that skips a congestion zone, excludes exactly that area rather than trying to draw a doughnut with the radius slider.
• Cutting a low-value pocket from a metro buy. If one district in your city converts badly (wrong demographics, no parking, out of your price band), exclude it and let the budget concentrate where it works.
• Suppressing an area you already own. A brand running a separate campaign in one region can exclude that region from the national campaign so the two do not compete in the same auction and inflate each other's costs.
• Keeping clear of a location you should not advertise near. Some regulated categories need to avoid certain areas; an exclusion is the clean way to guarantee it.

One caution: every exclusion shrinks your audience, and stacking too many can leave a set too thin to deliver. Exclude with intent, then check the estimated audience size before you publish.

Bulk locations, DMAs, and multi-market setups

When you outgrow a single city, the question becomes how to cover many places without building a separate campaign for each. Meta supports bulk entry: instead of adding locations one at a time, you can paste a list of cities, regions, or postal codes into the location box in a single action, which turns a fifty-postcode delivery zone from an afternoon of clicking into one paste. You can also mix levels in the same set, for example a whole country as the base with a handful of cities excluded, so a national campaign that skips two markets does not need a workaround.

DMAs are the right tool when your real market ignores administrative lines. Because a US metro's economic footprint often crosses county and state borders, targeting the DMA (one of Nielsen's 210 markets) captures the whole trading area in one selection, where a city marker would clip the suburbs that actually shop with you. This is a US-only option; outside the US, regions and postal-code lists do the equivalent job.

How you group locations changes how you learn from the data. A single set that blends very different markets (a dense city and a rural county, or two countries with different costs and languages) hides which area is carrying the results. If two markets behave differently enough that you would act on the difference, split them into separate ad sets so each reports on its own. If they behave alike and you only care about the combined result, keep them together to give the delivery system a healthier pool to optimise across. The choice is a reporting decision as much as a targeting one.

A decision guide: local service business vs national brand

The same controls point in almost opposite directions depending on what you sell and where you serve. Two profiles cover most cases.

If you are a local service business (a clinic, a tradesperson, a restaurant, a single store), start narrow and widen only if you must. Drop a pin on the address, set the radius to the distance customers genuinely travel, and add postal-code exclusions for any pocket inside that circle you do not serve. Accept that the "recently in" rule will include some visitors, and let your offer speak to locals so the wrong clicks self-filter. The failure mode here is a radius set "to be safe" at 40 miles when customers come from 8, which quietly burns budget on people who will never walk in.

If you are a national brand (ecommerce that ships country-wide, a service with no physical location), do the opposite: select the whole country as a single location, leave the radius off entirely, and let the delivery system find buyers across all of it. Slicing one country into dozens of city sets is the classic own goal, because it fragments the audience, and each thin set struggles to gather the roughly 50 optimisation events per week Meta's learning phase needs to stabilise. Split geography only when a region genuinely differs (a separate language, a different price, a warehouse that only ships regionally) or when you want a clean read on one market. Otherwise, one broad location plus strong creative beats a mosaic of tiny ones.

Cost sits underneath both decisions. Media is not cheap: Gupta Media's tracker put the blended Meta CPM near $8.19 across 2025, and where you point that spend matters. CPM varies widely from one market to the next, and the most competitive, saturated markets command the highest rates, so a broad national buy in an expensive market pays a premium a local buy in a cheaper one never sees. That is one more reason a local business should keep its circle tight and a national brand should watch which markets it lets a broad set drift into.

Getting location targeting right before you launch

Location is the control the delivery system will not overrule, so it rewards a few minutes of care before you publish. Run this quick check on every ad set:

• Right level for the job. Pin and radius for a fixed location, country for national reach, DMA or postal codes for anything in between. Do not use a country when you mean a city, or a city when you mean a street.
• Radius sized to real travel, not comfort. Match it to how far customers actually come. Tighten first, widen only if the audience is too thin to deliver.
• Exclusions in place. Cut the zones you cannot serve or do not want, and remember exclusions override includes.
• Audience size sane. Read the estimate. Too small and delivery gets expensive and spills; too broad and a local budget spreads thin.
• Spill accepted, not fought. Since May 2023 you cannot exclude recent visitors, and inferred location is imperfect, so treat a small outside-the-edge percentage as normal rather than chasing a setting that no longer exists.

Location decides where your ad is allowed to appear; the creative decides whether it works once it gets there. A tight, correct radius that pairs with the wrong message still wastes money, and the fastest way to improve a local campaign is usually the offer and the hook, not another tweak to the map. That is the loop worth running on repeat: research what is working in your market, generate and edit the creative on-brand, launch it to the exact area you have defined, then read the result and refine. A platform like AdPlay.ai keeps that loop in one place, but the discipline holds with any workflow. Get the geography right once, then spend your energy on the ad itself.