If you've ever shaded in a ZIP code on a map — for a sales territory, a school district analysis, or a real estate heatmap — you've quietly committed to an idea that isn't quite true: that ZIP codes are geographic regions with defined borders. They aren't. ZIP codes are mail delivery routes, and treating them as polygons leads to problems that range from minor annoyances to multimillion-dollar legal disputes.
Here's what ZIP codes actually are, what they aren't, and where the difference shows up in practice.
The U.S. Postal Service defines a ZIP code as a collection of mail delivery routes that share a sectional center facility (SCF) and a destination post office. That's it. A ZIP code is not a shape. It's a set of address points — individual mailboxes — that happen to be served by the same letter carriers and grouped under the same number for sorting purposes.
Some ZIPs cover dense geographic areas. Some are buildings (like the 12345 General Electric facility in Schenectady, NY). Some serve rural delivery routes that snake across hundreds of square miles of farmland with no real "boundary" at all — just a list of mailboxes on a sequence of roads. A few cross state lines. A few exist purely for PO Boxes and have no street geography whatsoever.
Software, marketers, and journalists almost universally treat ZIP codes as polygons. The reason is practical: people need geographic data, and the U.S. Census Bureau publishes a dataset called ZCTAs (ZIP Code Tabulation Areas) that approximates ZIP codes as polygons for statistical purposes. ZCTAs are derived by taking the ZIP code most commonly associated with the addresses inside each Census block and assigning that ZIP to the block.
ZCTAs are useful — they're how every ZIP-code map you've ever seen is built — but they're not the same thing as USPS ZIP codes. The Census Bureau's own documentation explicitly warns:
"ZCTAs do not always have the same boundaries as the ZIP codes used by USPS. ZIP code boundaries change frequently, and ZCTAs are not designed to match exactly."
Auto and home insurance carriers use ZIP codes as a primary input for pricing. Two houses on opposite sides of the same street can have wildly different insurance premiums if a ZIP boundary runs down the middle of the road — a phenomenon documented in numerous regulatory complaints. Some state insurance departments (notably California's) require carriers to use more granular geocoding rather than ZIP-only models, precisely because the routes don't align with risk in any meaningful way.
Older e-commerce systems used ZIP codes to determine the local sales tax rate. This is now widely understood to be unreliable. A single ZIP can span multiple tax jurisdictions; a ZIP centered in a city can include unincorporated county pockets with different rates. The 2018 South Dakota v. Wayfair Supreme Court ruling pushed many e-commerce platforms to adopt rooftop-accuracy address resolution rather than ZIP-based tax lookup, because the legal risk of charging the wrong rate is now significant.
Public health researchers commonly aggregate disease incidence, income data, or social outcomes by ZIP code — often because that's the granularity at which the source data is reported. Because ZIP codes vary enormously in population (from 1 person to over 100,000) and area (from a single building to thousands of square miles), comparing rates across ZIPs without normalization produces misleading results. The "ZIP code lottery" phenomenon — where life expectancy varies dramatically across ZIPs in the same city — is real, but the analytical method that produces those headlines often has methodological weaknesses tied to ZIP geography.
Real estate listings often imply that a ZIP code corresponds to a school district. It almost never does. School district boundaries are set by school boards and state education departments using their own geography that has nothing to do with USPS routing. The same goes for voting precincts, congressional districts, and city limits — none of which align reliably with ZIP boundaries.
One of the strangest consequences of ZIP geography: a ZIP code's "preferred city name" can be different from the city you actually live in. USPS assigns a single preferred city name to each ZIP, based on the post office that serves it. If you live in an unincorporated area or a smaller community within the service area of a larger town's post office, your mail might be addressed with the larger town's name even though you've never been a resident there. This leads to homeowners insisting they live in a city they technically don't, or paying higher taxes because tax software thinks they live in a wealthier nearby town.
For most professional applications, you should not use ZIP codes as geographic units. Better alternatives:
For consumer-facing displays and search interfaces, ZIPs are still useful — people understand them, they're memorable, and they're "close enough" for finding roughly where something is. The trouble starts when ZIP-based estimates feed into automated decisions: insurance rates, tax calculations, lending models, or research conclusions.
USPS itself, in Publication 28: Postal Addressing Standards, makes this explicit: "ZIP codes do not represent fixed geographic boundaries. They identify the mail delivery routes assigned to a particular address." That is the authoritative answer, straight from the agency that owns the system. Anyone using ZIP codes as polygons is making a useful approximation — but it's still an approximation, and one that occasionally generates bad outcomes.
The next time you see a colorful ZIP-code map of incomes, health outcomes, or political affiliations, remember: that map is showing you a delivery route, dressed up as geography.