How does IP geolocation actually work?
IP geolocation works by matching your address against databases built from ISP registration records, network-latency measurements, and crowdsourced Wi-Fi and GPS signals — not by tracking your device directly. That combination is why it nails your country almost every time, gets your city right most of the time, and never pinpoints your street. Location is just one piece of what a bare address exposes — see our rundown of what your IP reveals for the rest, like your ISP and network type.
How the databases are built
Three main ingredients feed a geolocation database. The first is registration data: regional registries publish which organization or ISP each address block was allocated to, and providers usually declare a serving city or region for it — though blocks get reassigned and reshuffled between regional pools more often than that public data gets updated, which is where a lot of drift creeps in. The second is latency measurement: providers run probes from many servers at known physical locations and time how long it takes a signal to reach a given address block; consistent round-trip patterns let them infer a plausible physical distance and narrow the estimate. The third is crowdsourcing: apps and services with location permission occasionally report “I’m at these GPS coordinates and I currently see this public IP,” and enough of those reports stitched together anchor mobile and Wi-Fi address ranges to real places far more precisely than registration records alone ever could.
Why “approximate” is the honest word for it
None of those three methods locates you — they locate the address block your connection currently borrows from, which an ISP typically allocates per regional hub rather than per neighborhood. The dot our map shows is the calculated center of the area that block is documented against, and we label it “approximate” on purpose, because that’s exactly what it is: treat it as “somewhere in this metro area,” never as a home address. Two neighbors on the same ISP block can show noticeably different coordinates, and an address that moved to a new customer last month may still carry the previous city in a database that hasn’t caught up yet.
Why mobile and VPN throw it off even further
Mobile carriers route traffic through a limited number of regional gateways, so a phone’s public IP can geolocate to the gateway’s city rather than wherever the phone actually is — sometimes hundreds of kilometers off, even while sitting still at home on Wi-Fi-off cellular data. VPNs and proxies push this further on purpose: as covered in our guide on VPNs and proxies, they deliberately replace your real IP with one belonging to their own server, so any geolocation lookup reports the VPN exit location instead of yours — which, for most people using one, is the entire point.
Accuracy expectations by network type
Wired residential and fiber connections geolocate to the correct country essentially always and to the correct city the large majority of the time, since the underlying address blocks change hands infrequently and the registration data stays fairly fresh. Mobile connections nail the country too, but city-level accuracy varies widely by carrier and can land you in the wrong metro area entirely, especially outside major cities where gateways are sparser. VPN and datacenter addresses geolocate accurately to the server’s location by definition — it just isn’t yours, which is the whole point of using one. Corporate and campus networks often geolocate to company headquarters or a regional office rather than the specific building you’re sitting in, since the ISP account and its address block are frequently registered centrally rather than per branch. Satellite internet is its own case: providers often route traffic through a small number of ground stations, so an entire region can geolocate to wherever the nearest station happens to be. In every case, treat the result as a reasonable starting estimate, not a verified fact.