How Geofencing Advertising Works, Step by Step

Geofencing advertising shows ads to phones that have been inside a virtual boundary drawn around a real place, such as an event venue, a showroom or a trade show hall. When a phone reports a location inside that boundary, it can join an audience. Ads then reach that audience in apps and on websites, during the visit or for a set time afterward. This explainer is part of our paid advertising guide.

It walks through each link in the chain: where location comes from, how it reaches ad systems, how an audience is built, and where the method runs out of precision.

The basic idea

A geofence is a shape on a map. It can be a circle around a point, or a polygon traced around a building, a parking lot or an event site. The fence itself does nothing. It is a rule that says: any device seen inside this shape, during this time, belongs to this audience.

From there, geofencing works like any other audience in digital advertising. The ad platform keeps a list of devices that matched the rule. When one of those devices later opens an app or a website with an ad slot, the platform can bid to show your ad.

So geofencing really answers one question: who was here? It doesn't know why someone came, whether they bought anything, or who they are. That makes the choice of place, and of time, the core strategic decision. A fence around a trade show during show hours says far more about intent than a fence around a busy road.

Where the location comes from

Phones work out their location from several signals. Apple lists cellular, Wi-Fi, GPS and Bluetooth for iPhone Location Services. Google's ads help describes the same sources and how precise each one is:

Precise device location needs the user's permission. On iPhone, an app must ask before it uses Location Services, and people can allow it once, only while the app is open, or not at all. Android splits access into foreground and background, and into precise and approximate location. Each of those choices shrinks the pool of devices that can ever match a tight fence.

How location reaches the ad system

Most geofenced ads are bought programmatically. When an app has an ad slot to fill, it sends a bid request to ad buyers. The industry standard for that message is OpenRTB, published by IAB Tech Lab. Its Geo object carries the device's location, where it came from, how accurate it is and how old it is.

Here is a trimmed example of that object, with placeholder coordinates:

{
  "geo": {
    "lat": 40.0000,
    "lon": -100.0000,
    "type": 1,
    "accuracy": 20,
    "lastfix": 45
  }
}

`type` 1 means GPS or Location Services, 2 means IP address and 3 means user provided, such as registration data. `accuracy` is the device's own estimate in meters, and `lastfix` is how many seconds old the reading is. The specification also says a postal-code centroid should not be passed as a precise point. A buyer can use these fields to ignore readings that are too vague or too old for a small fence.

The device object in the same request can carry an advertising ID and a limit ad tracking flag. That ID is what lets a platform remember a device it saw inside the fence and recognize it later.

From fence to audience to ad

Put together, a geofencing campaign runs in five stages:

1. Draw the fence. Choose the place and trace its boundary.
2. Set the time. Decide which hours or dates count, such as show days only.
3. Collect matches. Devices reporting a location inside the fence, in that window, join the audience.
4. Serve ads. Show ads to the audience while they're there, or later, within a lookback window you choose.
5. Measure. Compare what the audience did against people who weren't reached.

The ads themselves run wherever the platform buys space: mobile apps, mobile websites and, with some partners, other devices linked to the same household. Pairing a fenced audience with streaming TV ads is one common follow-up. We cover that pairing in how geofencing pairs with streaming TV.

What geofencing can and can't do

Geofencing is good at reaching people based on somewhere they've actually been. It suits events, trade shows, your own locations and venues closely tied to what you sell. It is less suited to reaching everyone in a wide area. For that, ordinary location targeting by city, ZIP code or radius is simpler, which we compare in geofencing vs geotargeting.

It also has hard limits. A fence can only match devices that shared a precise location, so the audience is a sample of visitors, not all of them. Readings drift, especially indoors or in dense areas, so a fence drawn tightly around one shop may also catch the shop next door. And small targets may not serve at all. Google says a very small radius or location may show ads "intermittently or not at all", and its interface won't accept a radius under 1 km.

How you size the fence, set the window and choose how long people stay in the audience decides most of the outcome. That's covered in geofence radius and audience best practices.

Privacy and platform rules

Location data is personal, and regulators treat it that way. In its enforcement actions, the US Federal Trade Commission says location data is sensitive personal information. It lists places, such as medical facilities, places of worship and shelters, whose visit data businesses should in most cases not use at all. It also expects clear consent for each use.

Ad platforms add their own rules. Google requires minimum area and user counts for every location target. It also bans ZIP-code targeting for housing, employment and consumer finance ads in the US and Canada. Meta has its own policies for those categories, so check them too. Any partner you use should be able to explain where its location data comes from and how consent was collected.

If you'd rather have the fence, the audience and the follow-up ads handled for you, our in-house ads team plans and runs geofencing campaigns alongside the creative.