What Makes an Address Valid? Postal Standards Explained
Most people use "valid address" to mean something like "this looks right." Postal systems mean something much more specific. Two addresses can look identical in structure yet sit worlds apart in terms of whether a letter will actually arrive.
At a Glance
- Format validity checks structure only: does the string have a house number, street, city, state, and ZIP in the right shape.
- Deliverability is a harder standard: the address must correspond to a real location in a carrier's delivery records.
- USPS exposes this distinction directly through its Address Information API, which returns a "DPV Confirmation" flag of Y or N alongside the standardized address.
- A delivery point is the specific two-digit code, combined with ZIP+4, that pins an address to one mailbox; Delivery Point Validation is the certified check for it.
- Synthetic addresses are built to pass format checks and fail deliverability checks, which is intentional, not a defect.
- International equivalents (Royal Mail's PAF, Canada Post's National Address Register) enforce the same distinction under different names.
Syntactic Validity vs. Deliverability
A syntactically valid address follows the correct format for its country. In the United States, that typically means a house number, a street name, a city, a two-letter state abbreviation, and a five-digit ZIP code. The address 742 Evergreen Terrace, Springfield, OR 97477 ticks every box: it has a recognizable structure, a real state code, and a ZIP that exists.
Deliverability is a harder standard. A deliverable address must actually correspond to a physical location in postal carrier records. The USPS Delivery Sequence File, for example, lists every address a carrier is authorized to deliver to. If your address string doesn't appear there, the post office won't send a truck.
This distinction isn't just a developer abstraction. USPS builds it directly into its own tooling. The USPS Address Information API documentation describes an Address Validation endpoint that "corrects errors in street addresses, including abbreviations and missing information, and supplies ZIP Codes and ZIP Codes + 4," and returns a DPV Confirmation field indicating whether the address "was considered deliverable or undeliverable." A response of "Y" means the address matched a real delivery point. A response of "N" means the format may be perfectly correct, but no carrier delivers there.
These two concepts are genuinely distinct. Format can be right while the underlying record is missing, the street was renamed, or the building was demolished. Postal validation software catches this distinction; simple pattern matching does not.
The Role of Postal Reference Data
Postal agencies maintain large, frequently updated databases that form the backbone of address validation. The USPS publishes several:
- Delivery Sequence File (DSF2): Every deliverable address in the country, updated weekly.
- ZIP+4 database: Extends basic ZIP codes to a narrower carrier route segment, often identifying a single building or side of a block.
- CASS (Coding Accuracy Support System): A certification standard for address-matching software, ensuring tools that standardize addresses meet USPS accuracy benchmarks.
Other countries have equivalents. Royal Mail in the UK publishes the Postcode Address File (PAF). Canada Post maintains the National Address Register. These databases cost money to license and require regular refresh cycles to stay current, which is why address validation is a paid service rather than something you can implement with a regular expression.
A critical detail: the USPS standardizes addresses before matching. "123 Main St." and "123 Main Street" are the same thing to CASS-certified software. The standardization step converts abbreviations, corrects common misspellings, and reorders components before the lookup runs.
What a Delivery Point Actually Is
The term "delivery point" gets used loosely, but it has a precise technical meaning. According to the Wikipedia entry on the subject, a delivery point is "a specific set of digits between 00 and 99 assigned to every address" that, combined with the ZIP+4, uniquely identifies one deliverable location, down to a specific suite or apartment. Delivery Point Validation, often shortened to DPV, is the certified USPS program that checks whether a given delivery point digit and ZIP+4 combination is real.
This is a useful mental model for the whole validity question. A format check confirms the shape of the string. A ZIP+4 lookup narrows the location to a small geographic segment. A DPV check confirms that one specific mailbox exists at that segment. Each layer catches something the previous one missed, and each layer costs more to run than the one before it, which is why most consumer-facing forms only implement the first layer.
Why Synthetic Addresses Are Format-Valid but Not Deliverable
Tools like this one generate addresses that are realistic in structure but entirely fictional. A generated address such as 4819 Maplewood Circle, Denver, CO 80203 might have a real ZIP code for Denver's Capitol Hill neighborhood, a plausible street name pattern, and a credible house number range. Every formatting rule is satisfied.
It will not, however, appear in the DSF2. No carrier is routed to deliver there. No apartment complex sits at that corner. Run it through the DPV Confirmation field described above and it will come back "N." The address is synthetic: format-valid, never deliverable.
This is the intended behavior. Synthetic addresses are built for software testing, form validation, and placeholder data. They let developers exercise address fields, test UI flows, and check that downstream systems handle input correctly, without touching anyone's real location data. Using a real person's address for testing creates privacy exposure; using a fictional address with correct structure avoids that entirely.
The distinction matters especially for address validation vs. verification: format validation catches structural errors, while verification confirms existence in postal records. A synthetic address passes the first test and fails the second, by design.
What Validators Actually Check
Address validation software generally works in layers, and each layer catches different problems.
Format Parsing
The first pass checks whether the input can be broken into recognizable components. Does it have a house number? Is the state code two letters? Does the ZIP match the expected pattern for the country? This catches obvious garbage like hello, world but won't catch a real-looking fictional address.
Standardization
Before looking anything up, good validators normalize the address. Street suffixes get abbreviated or expanded consistently. Directional prefixes (North, South, NW) are standardized. Secondary unit designators like Apt, Suite, and Unit are normalized. USPS standardization rules define the exact transformations for American addresses.
Geocoding and Record Lookup
The standardized address gets compared against reference data. A match confirms the address is deliverable. Partial matches indicate a likely typo or an address that's close but not exact. No match at all suggests the address is fictional, demolished, or simply not in the database yet (new construction sometimes lags several months).
Delivery Point Validation
At the highest confidence level, validators check the specific delivery point code, a unique identifier that maps to a single mailbox or suite within a building. This level of verification confirms not just that a street exists, but that the specific unit number is real and active.
| Address | Format Valid | Deliverable |
|---|---|---|
1600 Pennsylvania Ave NW, Washington, DC 20500 | Yes | Yes |
4819 Maplewood Circle, Denver, CO 80203 (synthetic) | Yes | No |
999 Fake St, Nowhere, ZZ 00000 | Partial (bad state) | No |
Springfield | No | No |
Address Formats Across Borders
The United States format is not universal. Address formats vary significantly by country. Japan writes addresses in descending geographic order, from prefecture to ward to block to building. Germany places the house number after the street name, not before. The UK uses a postcode at the end, but the postcode covers a much smaller area than an American ZIP code, sometimes just a single building.
Validating an international address requires country-specific rules and country-specific reference data. A tool built on USPS data alone cannot verify a German address, even if the format looks correct to an American eye. Synthetic address generators that support multiple countries apply each country's formatting rules, producing output that's locally plausible but globally fictional.
What This Means for Building or Buying a Validator
If you're deciding how much validation your application actually needs, it helps to separate the question into three tiers rather than treating "address validation" as one feature to turn on or off.
- Just checking the form isn't empty or malformed: a client-side regex or format check is enough, and it costs nothing per lookup.
- Reducing failed deliveries and returned mail: you need a standardization and match pass against reference data like DSF2 or the equivalent in your target country, which typically means a paid API.
- Guaranteeing an address is currently active and receiving mail: you need delivery-point-level validation, the most expensive and most accurate tier, usually reserved for high-value transactions like account verification or shipping insurance claims.
Most consumer signup forms only need the first tier. E-commerce checkout flows benefit from the second. Financial services and anything tied to identity verification often require the third.
Frequently Asked Questions
Can a format-valid address ever become deliverable?
Occasionally, yes. New construction sometimes produces addresses that match synthetic outputs by coincidence, especially for common street name patterns in growing suburbs. This is statistically rare and not something to rely on. For any testing scenario where real deliverability matters, use a real address you control, not a generated one.
Does ZIP code validation confirm deliverability?
No. ZIP codes cover geographic areas that contain thousands of addresses. A ZIP code matching a real city confirms the number is in the right range, nothing more. The house number and street still need to match a carrier record for the address to be deliverable.
Why do some forms accept synthetic addresses?
Many online forms perform only format validation: they check that the ZIP is five digits, the state is two letters, and the required fields are filled. This is cheap and fast. Full verification against postal databases requires an API call and a licensing agreement, so many businesses skip it and accept format-valid input without confirming deliverability.
Are synthetic addresses legal to use?
For software testing, form development, and UI prototyping, yes. The intended use of synthetic address generators is to avoid collecting real personal data. Problems arise if synthetic addresses are used to misrepresent identity or location in a context where accuracy is legally required, such as financial forms, government filings, or contracts. For those use cases, only real, verifiable addresses are appropriate.
What does a DPV "N" response actually tell me?
It tells you the standardized street address and ZIP+4 don't match any known delivery point in the carrier's records right now. It doesn't necessarily mean the address is fake. It can also mean the address is brand new and hasn't been added to the database yet, or that a typo in the unit number is close enough to pass format checks but not close enough to match. Treat "N" as "needs a human look," not automatically "invalid."