Why two IP geolocation databases show different cities

Investigate conflicting IP city labels using database dates, accuracy radius, app behavior and an agreed acceptance rule.

A fenced communications tower overlooks a city stretching along both banks of a river.

Treat location as an estimate with a date

An IP-geolocation record represents a publisher's best estimate of where an address is used. Data coverage, update cycles, address transfers and carrier routing can influence the result. Some databases provide an accuracy radius, which is more informative than a map pin viewed in isolation. A city name with a wide radius is a weak basis for a neighborhood-level promise.

Reproduce the discrepancy

Record the address privately, both database names, the observation time and the result your own site displayed. Identify whether the site also used an account address, browser permission or a manually selected store. If your site matched the intended regional experience, a discrepancy in another database may be interesting without being the cause of a user-facing failure.

Suppose two services identify the same country but different nearby cities. If the application changes behavior only at a national boundary, the disagreement may have little operational consequence for that particular buying decision. MaxMind IP geolocation accuracy comparison.

Choose an escalation that fits the evidence

A network operator may publish a geofeed, and a database provider may offer a correction channel. Keep those processes separate from a request for an immediate replacement address. In an XtremeProxy trial, agree which location signal controls acceptance before the first run. The agreement gives both teams a practical basis for deciding whether to rerun a check, seek an update or select a different location.

Two databases disagree about adjacent cities

A fictional assignment is labeled City A by one location database and City B by another. Both agree on country and the storefront correctly serves the expected national catalog. For this project the commercial decision is satisfied, while a finer city claim remains unresolved. Capture the database editions, observation time and the service's matching rule. If a local delivery requirement emerges, test the retailer's approved postcode model separately instead of averaging two inferred city names.

Keep an observation with the database edition

Location estimates can change after network reallocation or a database refresh. Record the lookup source and date, the level of precision the application requires, and the decision actually made. If two sources agree on country but differ on city, report the agreement at country level and the unresolved city separately.

A result displayed as a point on a map can look precise even when the underlying record describes a whole region. Retain the database’s granularity and any supplied accuracy radius. Converting the point to the nearest street address would introduce a claim the original data never made.

Make the destination decision visible

A shop consults database X and serves market A. A public lookup consults database Y and labels the route market B. Check the effective source inside the owned application before changing the order's targeting requirement. The useful report includes both estimates and the actual customer-facing decision. Agreement with an unrelated lookup is a separate measurement.

Preserve the source database version behind an internal city decision. An application might still classify an address with last quarter’s data even though its vendor’s public lookup has moved the same prefix to a neighboring city. Compare the application’s installed version and its update schedule with the version used for the external observation. Labelling the timestamps and versions can explain the discrepancy without claiming that either database independently establishes a device’s physical location. Preserve that context in the incident record.

Give the data owner enough context to determine whether a lookup result has business impact. Suppose both sources name different suburbs, while the application groups them into the same delivery territory. Its observed customer decision may remain correct. A later policy that introduces different delivery cutoffs for the suburbs would require a more precise test and a more specific acceptance rule. Preserve the original evidence and scope so a success at the first level cannot be reused uncritically for the finer requirement.

The appropriate resolution depends on what the application needs. A national catalog can work correctly amid a city disagreement, while a city-specific portal needs finer investigation. Dealer locator territory assignments shows a separate business territory rule to consult. Country, city and ASN targeting in a proxy order helps put the right granularity into the next order and acceptance brief. Technical references include RFC 8805 on self-published IP geolocation feeds. Check XtremeProxy location coverage.

Frequently asked questions

Does a database disagreement justify correcting every customer record?

First identify which database and matching rule the application uses and whether the disagreement affects a customer decision. A correction request should include dated evidence through the relevant provider's process, without making claims about an individual's physical address.

Related articles