Geofencing is sold as a simple idea: draw a boundary on a map, advertise to people who entered it. The drawing is trivial. Everything that determines whether it works is in the data underneath, where the location came from, how precise it is, how recently it was collected and whether the person consented to it being used this way. Vendors differ enormously on all four and rarely volunteer the differences.
Precision is the question that separates vendors
Location derived from a device's satellite positioning while an app was open is accurate to metres; location inferred from an address network can be accurate to a city block or to a postal district. The second is sold under the same name and cannot reliably tell somebody inside a store from somebody driving past. Ask what the source is, what the stated accuracy is, and what proportion of the audience comes from each.
Small boundaries are where the claims break down
Fencing a stadium or a large retail park is plausible. Fencing a single competitor's front door is where the data's precision runs out, and it is exactly what the sales pitch tends to describe. If the proposal involves boundaries measured in tens of metres, ask how many devices the vendor expects to see there in a week; the honest answer is often too few to matter.
Consent and sensitive locations are a real constraint
Location data is regulated differently across jurisdictions, some states impose specific obligations, and advertising based on presence at a health facility, a place of worship or a school carries both legal and reputational risk. Ask where consent comes from, whether sensitive categories are excluded, and get it in writing rather than as a reassurance on a call.
Measure with something other than the vendor's own attribution
Store-visit attribution is modelled, and the model belongs to the party being paid. Use a holdout: comparable areas where the campaign does not run, and compare actual sales or bookings. It is less impressive than a dashboard and it is the only way to know whether anything happened that would not have happened anyway.
Consider whether ordinary targeting does the same job
For most local businesses, targeting by postal area and radius in the ordinary advertising platforms reaches broadly the same people at a lower price and with better measurement. Geofencing earns its premium for genuinely event-driven cases, a conference, a fixture, a trade show, where being at a place at a time really is the signal. Ask a vendor to explain why the ordinary version would not work for you.
Questions people ask about geofencing company
How accurate is geofencing?
It depends entirely on the data source. Satellite positioning from an open app is accurate to metres; network-derived location can be accurate to a block or a district. Both are sold under the same word, so ask which you are buying.
Can we target a competitor's premises?
You can draw the boundary. Whether enough devices are seen there with enough precision to matter is a different question, and for a single small building the honest answer is often no.
Is it legal?
Generally, with real constraints: consent requirements vary by jurisdiction and advertising based on presence at health, religious or educational sites carries specific risk. Ask where consent comes from and which categories are excluded, in writing.
How do we know it worked?
Run a holdout in comparable areas and compare actual sales or bookings. Vendor-modelled store visits are calculated by the party being paid, which does not make them wrong but does make them unsuitable as the only evidence.