Verification vs. Correction
A common misunderstanding about what CDYNE Postal Address Verification service does has to do with the difference between verification and correction. Verification simply says something is correct or it isn’t. We can easily verify 2 + 2 = 4, likewise 3 + 3 = 4 fails verification. Verification makes no attempt to fix what is false. It is a simple statement that something is correct or it isn’t.
Correction on the other hand takes a statement and tries to make it true and that is a much more complicated process. Take the above example: 3 + 3 = 4; we can correct that to 3 + 1 = 4 OR 3 + 3 = 6. Both corrections are true but which statement was originally intended is unclear. To properly correct we need more information.
CDYNE PAV is certified by the United States Postal Service under a program called CASS (Coding Accuracy Support System). CASS rules require that our software take an input address and verify it exists in the USPS databases. Of course there is more to it than that. There is standardization, limited spelling correction, and certain inexact matches are allowed, but ultimately following CASS rules is about verification not correction.
This is best summed up by an excerpt from the USPS Address Matching Guidelines documentation:
“…the main goal of all address-matching software
should be to make THE MATCH — not just any match”
When an attempt is made to match an input to the database, if a match cannot be found, or it is unclear which address the input should match, the input fails verification. It is also important to point out that failing verification does not mean an address cannot be delivered, but mailing out something without a verified address may take longer, it might go to the wrong location, it might be returned undeliverable, or it might be lost in the system.
Addresses are Complicated…
Addresses are a mess. One of the greatest contributing factors to this is that the USPS has no control over city, street names, or street numbering. City borders move, streets are renamed, and numberings can change based on local and state government decisions.
Furthermore each municipality may create its own way of making addresses. For instance in Wisconsin the following addresses is common:
S90 W13970 Boxhorn
DrMuskego, WI 53150
In this case S90 W13970 are actually latitude and longitude coordinates.
To counter this USPS has developed its own collection of identifiers that it uses to expedite the movement of mail:
Finance numbers are typically assigned to a single post office and contain multiple zip codes. For the most part it isn’t necessary for anyone outside the USPS to know about finance numbers, but they are convenient for creating a collection of addresses to verify a single address against.
Zip codes are usually a collection of routes used to deliver mail, but may also be a collection of PO boxes, a unique number for a single business, etc. It is important to recognize that while zip codes are often thought about as regions they really are a group of places mail can be delivered. As such they can overlap or even be discontinuous.
Plus 4 values further subdivide a zip code into areas like a city block.
Delivery point number, which is only added after an address has been fully verified, signifies a particular location that receives mail.
The USPS has created and assigned these numbers for one reason only: to make delivering a piece of mail as easy as possible.
Address verification is the process of taking the input address, finding the best match in the USPS database, and returning the address found in the database.
Street Information Broken Down
Typical street information can be broken down into parts. While most addresses follow this format there are many that don’t and have to be handled separately. Consider the following prototype:
1234 N Main Street S Apt 101
Primary number (1234) – Typically the primary number is the first element and contains sequential alphanumeric information. Most often the input is checked to be in a range of values for verification. This value is present in almost all addresses and cannot be corrected, it’s either wrong or its right.
Pre-Directional (N) – The pre-directional can be one of the 4 cardinal directions or one of the 4 ordinals. This value isn’t always present and correction for it follows a very narrow set of rules.
Street Name (Main) – The street name is present in almost all addresses. It often contains more than one word and can be the most loosely matched too.
Suffix (Street) – The suffix is the “type” of street: Avenue, Road, Parkway, etc. Like the pre and post directionals it can only be corrected in a limited fashion, must match a set of values, and is abbreviated. While you will usually see a suffix it is sometimes missing.
Post-Directional (S) –The post-directional can be one of the 4 cardinal directions or one of the 4 ordinals. This value isn’t always present and correction for it follows a very narrow set of rules.
Secondary Abbreviation (Apt) – The secondary abbreviation is typically one of a collection of possible values: Apt (Apartment), Ofc (Office), Flr (Floor), Ste (Suite), etc. If all other address information matches this will be corrected to the database value if possible.
Secondary Number (101) – The secondary like the primary number contains alphanumeric information that is typically part of a sequence. It also cannot be corrected.
All address verification starts with the City, State, and Zip Code data; commonly referred to as the last line. From this information a list of finance numbers is built, and the finance numbers are separated into 8 categories:
The table above shows the priority of address matches in each finance number. For instance a match found in a Finance number that matches the City, State and ZIP in the input address would have higher priority than a match that is in a finance number that only matches the City and State of the input address. It’s important to note that a street address must match at least two pieces of the last line data, the only exception is when a zip 5 is the only piece of last line data provided.
Next a list of street names is pulled for all of the relevant finance numbers. There can be tens of thousands of street names and all of these must be compared to the input street name which can be ambiguous itself. A list of all possible addresses is generated from possible matching street names.
An attempt is then made to match the input address to one of the addresses in the list. If a single match is found it is returned, if multiple matches are found some tie breaking is attempted but if no single match can be determined the address fails verification.
Just because an address fails verification does not mean it is undeliverable. The post office may still attempt to deliver the piece of mail but it will most likely take longer to deliver, if it is delivered at all.
Click here for more information about PAV, or get a free trial key to test the API for free.