Geofencing 101: What It Is, What It Isn't, and How to Set It Up

A geofence is a virtual boundary drawn on a map. When a tracked asset crosses it, the platform fires an event, log, notification, alert, or automated workflow.

That's the whole concept. The complexity comes from what you do with it.

After watching dozens of operators set up their first geofence configuration, from Lake Tahoe rental marinas to ATV outfitters in Moab and motorhome fleets in Quebec, we've seen the same mistakes repeated. This post is the version of geofencing 101 we wish every new operator had before their first configuration call.

The two types of geofences

There are really only two functional categories, regardless of what platform you use:

• Inclusion zones. "The asset should stay inside this area." Triggers when the asset exits. Used for keeping rentals in approved water, keeping shop vehicles in a service radius, etc.
• Exclusion zones. "The asset should never enter this area." Triggers when the asset enters. Used for hazard zones, restricted waters, no-go properties, off-limits trails.

A lot of platforms layer extra labels on top of these, "safe zones," "alert zones," "no-go zones", but functionally everything reduces to one of those two categories. Knowing that simplifies setup considerably.

Shape choice: circle vs polygon

For 90% of cases, a circle is fine. Center it on a point of interest (your shop, your dock, the bad rock), set a radius, done.

Polygons matter when the area has a complex shape, a lake with multiple coves, a trail network that follows valleys, a marina with adjacent municipal water. We see operators try to draw polygons with 40+ points for shapes that a 5-point polygon would handle equally well. Don't.

Practical guidelines:

• Use circles for ranges and points. "Within 500m of the dock" is a circle.
• Use polygons for shorelines and trail networks. "The rentable area of Lake Powell" is a polygon.
• Keep polygons under 12 points. More resolution rarely improves outcomes; less resolution makes editing easier.
• Don't draw at maximum precision. A 50m fuzzy fence works better than a 5m tight one, because GPS itself has variance.

What ping rate does to your geofence behaviour

This is the part operators most often miss. A geofence event only fires when a ping is received and processed. If your tracker pings every 60 seconds, your fence is effectively "60-second-precision." An asset can cross a tight fence and re-cross before you ever see an event.

Two implications:

• Tight fences need fast pings. A 50m radius fence with 60s pings will miss fast-moving crossings. Either widen the fence or speed up the pings.
• Wide fences are forgiving. A 500m radius works fine even at 60s pings because the asset can't cross and re-cross within a single ping interval at realistic speeds.

For marine and off-road operators using 10-second active pings (what we recommend), 50–200m fences work well. For platforms with 60s+ ping rates, widen everything by a factor of 3–5x.

Hysteresis: avoiding the bouncing alert

A renter sits on the boundary of a fence. They drift in, drift out, drift in. Every drift fires an alert. By lunch you've ignored 47 notifications.

This is solved by hysteresis, defining separate entry and exit thresholds with a buffer between them. The asset must be inside fence A to count as "in," and outside fence B (slightly larger) to count as "out." The gap between A and B is the buffer.

Practical setup:

• 200m circle for "asset is inside this area"
• 250m circle (concentric) for "asset has clearly left"
• Alert only fires when crossing the outer boundary, not the inner

Most modern platforms, ours included, handle this automatically with a buffer parameter you set per fence. The default is usually 10–20% of the fence radius. Bigger buffer = fewer false alerts, slightly less responsiveness.

The minimum-viable geofence configuration

For a new rental operator (boat, jet ski, ATV, RV, anything), we recommend exactly this on day one:

1. Primary inclusion fence. A generous polygon covering the rentable area. Buffer: 15%. Alert on exit, soft severity.
2. One critical exclusion fence. A specific danger area or off-limits zone. Hard alert, phone the operator immediately.
3. Dock perimeter fence. A small circle around the home base. Used to determine "trip start" and "trip end," not as an alert.

That's three fences. Total config time: about 15 minutes. Tune the rest after 2–4 weeks of real data.

Common mistakes we see

A short list of patterns that show up repeatedly:

• Overlapping fences with conflicting rules. Asset enters fence A (allowed) and fence B (denied) simultaneously. Most platforms resolve this with priority levels, but operators rarely configure those, leading to ambiguous events.
• Time-based fences without timezone handling. "No movement between 10pm and 6am", defined in what timezone? An RV in Arizona vs Saskatchewan vs Atlantic Canada needs different handling.
• Fence boundaries on natural drift lines. Shoreline fences drawn exactly at the waterline produce constant false positives as the unit drifts. Pull the fence 20–50m inland.
• Fences with no notification routing. Triggering events that nobody sees. The alert system needs a destination, operator phone, dashboard, email, webhook, defined per fence.
• Same fence for all units. A jet ski's "no-go" zone is different from a fishing boat's. Apply fences per unit class, not fleet-wide.

Notification routing: who gets what

A good operator has tiered alerting:

• Soft notifications (dashboard chip, email digest) for routine boundary events. Renter drifts out of zone briefly, comes back. No human action required.
• Standard alerts (in-app push) for sustained boundary breaches. Renter has been outside the fence for 90+ seconds. Operator should check.
• Hard alerts (phone call + SMS) for critical events. Off-hours theft, hazard-zone entry, impact in restricted area. Operator must respond.

The mistake we see most is putting everything at "hard alert" severity. Within a week, the operator has tuned them all out and starts missing the genuine emergencies. Tiered alerting is non-negotiable.

Where geofencing breaks down

Geofences don't help with:

• Threats that happen inside the fence. A renter who damages a unit inside the rental area still damages it. Geofences are spatial, not behavioural.
• Tracker failures. If the device is dead, removed, or jammed, no fence fires.
• Stolen-but-stationary units. A unit hidden in a barn five blocks away may never cross your fence if your fence is drawn too narrowly.
• Multiple units moving together. A "convoy" of stolen units moving as a group still triggers individual alerts, which can overwhelm the operator at the worst moment.

For each of these, you need complementary controls: behavioural alerts (speed, impact), tamper detection, broad inclusion fences with wide buffers, and grouped alerting for fleet-wide events.

Beyond basic fences: the next layer

After 4–8 weeks of running a basic setup, operators tend to add:

• Speed-bounded zones. A fence where speed above X knots is an alert, separate from the boundary itself.
• Time-bound fences. Active only during certain hours (e.g. shop hours).
• Unit-class fences. Different fences applied to different unit categories.
• Sequence rules. "Unit exited dock area, didn't return within X hours", a stateful pattern, not just a boundary.

All of these compose on top of basic geofencing. They're worth adding once basic setup is solid, not before.

FAQ

How many geofences can I configure?

On our platform there's no practical limit per operator. We recommend keeping it under 25 per fleet to avoid alert overload.

Can I import geofences from another platform?

Yes, we support GeoJSON import. Most platforms can export to GeoJSON; if yours doesn't, we have a CSV converter.

Do geofences work offline?

If the tracker is buffering positions offline, geofence events are evaluated when the device uploads the buffered data. You'll get a delayed alert, not a missed one.

Can renters see the fences in their rental app?

Optional, per-fleet setting. Some operators show inclusion zones to renters as "your rental area" overlays. Most don't show exclusion zones for security reasons.

What's the practical accuracy limit of a geofence?

With 10-second pings and clear-sky GPS, you can rely on ~3–5m accuracy. With heavy tree cover or urban canyon conditions, 10–15m. Don't draw fences tighter than your worst-case accuracy.

Ready to set yours up

If you're new to geofencing and want help configuring your first three fences, reach out via the rental section. For most fleets, it takes 30–60 minutes including walkthrough. We work with operators across the US and Canada, coast to coast.