Geo-Fencing
Geo-fencing draws a virtual boundary around a physical area and triggers a marketing or app action when a device enters, dwells in, or exits that boundary.
Key takeaways
- A fence combines a shape with a trigger rule: entry, exit, or dwell for a set time.
- Dwell rules filter passers-by that a simple entry trigger would wrongly catch.
- Operating systems cap registered fences and throttle wake-ups to protect battery.
- An installed app with background location permission is a hard prerequisite.
- Indoor and underground positioning degrades, so triggers can fire late or not at all.
In depth
A fence is a shape plus a rule. The shape is a centre point and radius, or a drawn polygon following a building outline. The rule names the transition that fires: entry, exit, or dwell for a set number of minutes. An app or SDK registers the fence with the operating system, which watches location in the background and wakes the app when a boundary is crossed. Because the check is event-driven rather than continuous, the trigger arrives with a short delay rather than the instant someone steps across.
Radius trades accuracy against reliability. A tight fence is precise but misses people whose position is resolved imprecisely, while a wide one catches passers-by who never intended to visit. Dwell rules filter most of that noise, since requiring several minutes inside the shape separates a visitor from someone driving past. The operating system also caps how many fences an app may register and throttles wake-ups to protect battery, so a hundred small fences is not a workable design and clustered locations should be merged into fewer, larger shapes.
Deployment is nearly always about a physical moment: a venue, a store, a campus, a job site. In B2B the common case is an event, where entering a conference hall triggers an invitation to a short assessment and the fence records which session or hall the person was in. The location plus the quiz answers together let a rep on the floor pick who to approach next, since the fence supplies presence and the quiz supplies the reason to talk.
The technique needs a mobile app with granted background location permission, which is the constraint most plans miss. Without an installed app there is no fence, and web-based approximations only know where someone was when a page loaded. Consent has to be specific and revocable, and battery impact makes users withdraw it. Fences also fail indoors and underground where satellite positioning degrades, so a trade-show hall or basement venue may trigger late, early, or not at all.
Example in practice
How to measure it
Separate the fence from the message. Trigger count tells you the fence fired; open rate on the resulting notification tells you the moment and copy were right. Divide completions by triggers to get a per-fence completion rate, and compare fences against each other rather than against your general campaign benchmarks, because a venue fence and a store fence attract very different intent.
Watch the cost of the mechanism as well as its yield. Track permission revocations and app uninstalls in the days after a heavy fencing campaign, since a spike means the trigger frequency exceeded what users tolerate. On the value side, match triggered contacts against later meetings or orders and calculate the share of triggers that produced a real conversation, which is the only number that justifies the engineering and consent overhead.
Common mistakes
The frequent design error is firing on entry with no dwell requirement. A fence around a venue on a busy street triggers for every passing car, the notification arrives while someone is driving, and the campaign reports high reach with no engagement. Add a dwell condition of several minutes and cap how often one device can trigger, or a single person walking in and out repeatedly generates a stream of duplicate messages.
The second error is treating the fence as the whole campaign. A push notification with no reason to open it performs badly regardless of timing, and being nearby is not by itself a reason to buy. Pair the trigger with something the person can only use in that place and moment, such as a short assessment tied to the event they are attending, and prepare the follow-up before the day, because the value of the trigger decays within hours.
Frequently asked questions
What are good use cases for geo-fencing?
Events and trade shows, retail foot traffic, and competitive-location campaigns are classic uses. In B2B it works well for triggering a quick qualification quiz or booth invitation when prospects arrive at a venue.
How is geo-fencing different from geo-targeting?
Geo-targeting filters delivery by where someone appears to be when a request happens. Geo-fencing watches for a boundary crossing in real time and fires an action on entry, exit or dwell. One decides who is eligible to see something; the other decides when to interrupt. Geo-fencing needs an app with background location, while geo-targeting works on the open web.
Do I need an app to run geo-fencing?
For true background fences, yes. The operating system only monitors boundaries for an installed app that has been granted background location permission. Without one you can approximate with location-aware ads or a page that checks position when it loads, but neither will wake anything when a person arrives somewhere, which is the defining behaviour of a fence.
What consent do I need for geo-fencing?
Explicit, informed and revocable permission for background location, granted separately from general app use. Explain what is tracked, why, and how to turn it off, and honour withdrawal immediately. Local privacy law may add requirements, including limits on retention and on inferring sensitive attributes from places visited, such as clinics or places of worship.
What radius should I use?
Match it to the shape of the place, not to a round number. A single building is served by a tight polygon or a small radius plus a dwell rule; a campus or industrial park needs a wider one. Test by checking what fraction of triggers came from people who actually stayed, then tighten or widen until passers-by stop dominating the count.
Does geo-fencing work indoors?
Less reliably. Satellite positioning degrades inside large buildings and underground, so the device falls back on Wi-Fi and cell signals, which are coarser. Triggers may arrive minutes late, fire on the way out, or not fire at all. For indoor precision, venue Wi-Fi, beacons or a scannable code at the entrance are more dependable than a geographic fence.
What are realistic B2B use cases?
Events dominate: conferences, trade shows and customer sites, where presence is short-lived and highly relevant. Field service teams also use fences for arrival and departure logging. Competitor-location fencing is technically possible but often reads as intrusive, and it produces contacts whose presence says nothing about their authority to buy.