Geofencing draws a virtual line around a real place and acts when a phone crosses it. A store sends an offer as a shopper walks in. A fleet manager gets an alert when a truck leaves a depot. For loyalty programs, it ties an offer to the moment a member is near the store, which makes consent and restraint as important as the tech.
What is geofencing?
Geofencing uses location data to set a virtual boundary around a real place. When a device enters or leaves, an action fires. TechTarget lists the inputs as GPS, RFID, Wi-Fi or cell data (TechTarget). The boundary itself is the geofence.
Android’s guide for app makers defines a geofence by three numbers: a place’s latitude and longitude plus a radius. Together they draw a circle around it (Android). Apple calls the same thing condition monitoring, also known as geofencing. Its own Reminders app uses it to fire a reminder when the user arrives at or leaves a place (Apple).
A fence can be as small as one store or as large as a city. TechTarget notes that marketers also draw polygons on a map when a circle does not fit the site (TechTarget).
How does geofencing work on a phone?
The phone’s operating system does the watching, so the app does not need to stay open. On iOS, the system monitors the fences and wakes the app when one is crossed (Apple Developer). On Android, the app hands its fences to Location Services, which sends an alert on entry, exit or a stay inside (Android Developers).
Both platforms put hard limits on what one app can do. Those limits shape every geofencing plan, and few guides mention them.
| Limit | iOS (Core Location) | Android (Location Services) |
|---|---|---|
| Fences per app | No more than 20 at once | Up to 100 per app, per device user |
| Fence size | Apple’s own sample uses a 200-meter radius | Set the minimum radius at 100 to 150 meters for best results |
| Location accuracy | Not stated in the guide | Usually 20 to 50 meters with Wi-Fi. Several hundred meters or more without it |
| Alert timing | The system wakes the app when a condition changes | Usually under 2 minutes. Up to 6 minutes if the phone has sat still |
| Permission | Location access the user grants in settings | Fine location, plus background location on Android 10 and later |
Sources for the table: Apple’s guide on monitoring regions (Apple Developer) and Android’s geofencing guide (Android Developers). Android also offers a dwell trigger, which fires only when a user stays inside a fence for a set time. Google recommends it to cut alert spam from people who drive past.
What is geofencing used for?
Marketing is the use most people meet. A shopper near a store gets a push message or an in-app offer (TechTarget). Fleet teams get alerts when vehicles enter or leave a set area, such as a customer site (Geotab). Security teams limit access to some systems to approved places, which blocks a stolen login used from elsewhere (Huntress).
The term also has a legal meaning. A geofence warrant asks a tech company for data on every device inside an area during a time window. In August 2024 the Fifth Circuit Court of Appeals held that geofence warrants are general warrants categorically prohibited by the Fourth Amendment (United States v. Smith). The court noted that Google had been the main recipient of such warrants.
Is geofencing legal?
Geofencing is legal in the US, but how a brand collects and uses the location data is regulated. Three rules show where the lines sit.
- Consent has to cover every use. In January 2024 the FTC announced a settlement with data firm InMarket Media. The order became final in May 2024. It bars InMarket from selling, sharing or licensing precise location data (FTC final order). The FTC said InMarket’s rewards app, CheckPoints, asked for location to give reward points. It did not tell users the data would also feed targeted ads (FTC).
- Keep data only as long as the purpose needs. The same FTC case faulted InMarket for keeping location data for five years (FTC).
- Some places are off limits. Washington’s My Health My Data Act makes it unlawful to put a geofence around an in-person health care site to track visitors or send them ads (RCW 19.373.080). The law sets that fence at 2,000 feet or less from the site (RCW 19.373.010).
How do loyalty programs use geofencing?
Loyalty apps hold the two things geofencing needs: an installed app and a known member. Starbucks is a clear case. Its privacy notice says location sharing lets it tailor the experience in and around its stores. It also uses location to tailor messages to time and place patterns, and to cap how often a member sees an offer (Starbucks Privacy Notice). The same notice says Starbucks uses Bluetooth in or near stores to sense a device, when settings allow.
The common loyalty uses follow from that:
- A welcome message or open order when a member arrives at a store.
- A bonus offer near a new or slow location, sent through targeted offers.
- Points or a badge for a visit, which gives the program a signal of store traffic.
- A switch to an in-store view of the app, such as the member’s barcode.
Design choices decide whether members welcome the messages. A dwell trigger skips people who only drive past. A fence around a mall or an airport needs a larger radius than one around a corner store. The location prompt should name the loyalty use, and any other use needs its own yes.
The 20-fence cap on iOS forces a hard choice for big chains. Apple tells developers to set priorities within that limit (Apple Developer). A chain with 3,000 stores has to register only the stores nearest each member, then swap them as the member moves.
How do you measure a geofencing campaign?
Measure it against members who crossed the fence and got nothing. Many members who walk into a store were going to buy anyway. A control group shows how many extra visits the message caused.
For illustration, a coffee program sends an arrival offer to members who enter a fence. Over four weeks, 50,000 members cross a fence. The program holds back 10% of them, or 5,000, as a control group. Of the 45,000 who get the offer, 9% buy within an hour. Of the control group, 7% buy anyway. The offer drove 2 extra purchases per 100 members, or 900 in total. Say each purchase earns a $1 reward. Then 4,050 buyers cost $4,050 in rewards. Only 900 of those purchases were extra. That works out to about $4.50 per extra purchase, the number to judge the offer by.
Frequency caps protect the result. A member who gets an offer on every walk past a store learns to ignore it. Starbucks names limiting how often a member sees an offer as one use of its location data (Starbucks Privacy Notice). Location is also part of a program’s first-party data, so the consent rules above apply to it.
Frequently asked questions
How accurate is geofencing?
Geofencing on a phone is usually accurate to 20 to 50 meters when Wi-Fi is available, says Google. Without Wi-Fi, such as on a rural road, the error can grow to several hundred meters or more. That is why Google sets the best minimum radius at 100 to 150 meters (Android Developers).
What is the difference between geofencing and geotargeting?
Geofencing reacts when a device crosses a boundary, while geotargeting picks an audience by where it is. A geofence fires a message on entry or exit (TechTarget). Geotargeting serves ads to people in an area, such as a ZIP code or city, whether or not they move. Loyalty apps lean on geofencing because the trigger is a real visit.
Do customers have to opt in to geofencing?
Yes, in practice, since a phone will not share location without the user’s say. Android apps need fine and background location access to use geofencing (Android). The FTC’s InMarket order also shows that consent must cover every use of the data, ads included (FTC).
Does geofencing drain the phone battery?
Geofencing uses some power, and an app can tune it. Google says a higher responsiveness value cuts power use by checking less often. A larger radius around places where people spend hours, such as home, also cuts how often the app checks (Android Developers).
How many geofences can one app use?
An iOS app can monitor 20 regions at once, and an Android app can hold 100 per device user. Apple caps each app at 20 conditions so all apps get a share (Apple Developer). Android sets its limit at 100 (Android Developers). Chains with more stores load the nearest ones and swap them as members move.
