Service Objects has been validating addresses via real-time API interfaces for over 20 years, giving us a level of expertise that few companies can match. We constantly see challenging addresses that cover a wide range of issues, including extra data, missing data, misspelled data, incorrect city/state/postal codes and more.
This blog contains examples of some of the most challenging address scenarios we have seen over the years. Many of these have been changed from the originals, since they potentially could be considered personally identifiable data points, especially in the case of private residences.
Examples of interesting addresses we see
This is certainly not an exhaustive list of scenarios we see and can fix, but just a cross section of some of the most interesting ones. Let’s take a look at some of our favorites. Addresses in the United States primarily have a combination of the following elements:
House Number, Pre Directional, Street Name, Suffix, Post Directional, SUD 1,
SUD 2, PMB
City, State Zip Code
(SUD = secondary unit designator covering APT, Floor, Building and the like).
A legitimately clean address by itself can be very complex to validate if it has a lot of these elements. For example, an address like:
123 NE Main St SW Bldg 23, Suite 55
Anytown, CA 93101-1234
could be a legitimate address, but it is still tough to deal with programmatically due to all of the elements, especially with multiple SUD values. What about something like:
4000 Kirkwood St Georges Rd
At first glance it appears that the address provider messed up and submitted two addresses: Kirkwood Street and Georges Road. However, this one is actually valid and should be read as:
Kirkwood Saint Georges Street.
In hindsight it seems obvious, but programmatically these sorts of things can wreak havoc. Primary house numbers do not always have to be clean numbers. Addresses such as:
27W183 Main St
0-143 Main St
123 ½ Main Street
among many others, are all potentially valid house numbers. While not technically appropriate, it is not uncommon for us to see numbered street names fully spelled out – for example:
123 Second Street
instead of the more proper
123 2nd
This means you might have to take into account different possible permutations of the same street name. Finally, there are the not-uncommon street alias addresses, which confuse our clients more than they do us. For example:
4750 US Highway 78
is also known as:
4750 Stone Mountain Hwy
These are always fun ones to explain since the output address could look very different from the input one – suggesting a bad change. Add to all of this, addresses do not always come in clean, requiring data correction/standardization from our Address Validation services. For all of the work that Service Objects does to clean/fix/standardize messy addresses, far and away the most important thing to us is that we do not return an address that is changed to an incorrect address. While it is not always possible to be perfect, our clients rely on us to return valid, accurate results that they would feel comfortable sending valuable shipments to.
Common address mistakes
We see lots of challenging addresses coming through our system, and we do our best to fix them whenever possible. It is dangerous to fix everything, so it is a delicate balance of when to fix and when to fail. We do also offer other options such as Address Detective, which can use other data points to correct the address. Again, the goal is to make as many changes as possible without potentially changing the address into a completely new address. Here are some of the most common situations with addresses that need to be fixed. One of the main ways that addresses come in incorrect is via spelling errors. Take this address:
400 Hichcook Rd
Santa Barbara, CA
which is actually supposed to be:
400 Hitchcock Way
Santa Barbara, CA
Our service is able to compare this address against close streets and determine that Hichcook Rd is not correct and Hitchcock Way is the intended choice. We also see addresses that include extra information that is not relevant to an address, which can add lots of challenges to an address parser trying to figure out how to identify the address components. For example, addresses such as:
Historic Old Mission 123 Main St and C/O John Smith
123 Main St makes it more challenging to identify the actual address. In rare cases, we also see a house primary number represented as the spelled-out version of the number. For example:
One Battleship Rd
Wilmington, NC
is a valid address, but many parsers will consider it invalid because programmatically it appears to be missing the house number. We can generally clean up all of these types of cases; however, there comes a point where there might be too many extra elements and any changes might be too dangerous. Here as well, our goal is to always return an accurate address when corrected.
Our latest update to handling tough addresses
Service Objects recently put out a new build that is even better at handling these tough to fix type addresses. In some cases, we find addresses that contain combinations of the above issues. For example, the address:
Kaiser Permanente
One Kaisser Place
(Ordway Bldg), Fl 26
contains a spelled-out building number, a misspelled address and multiple extra points of information. The correct address here is:
1 Kaiser Plaza
FL 26
We are able to break apart this address and safely reform it into a valid address. Addresses with missing spaces can also cause challenges, such as this example:
63350Orchardclb dr#xxx.
These can be very challenging programmatically, but we are able to convert this to the correct address of:
6335 Orchard Club Dr STE XXX.
Missing spaces affect parsing, and the abbreviation of Club also adds an extra challenge.
We are constantly working to improve our Address Validation tools, and will soon be working on a new version of our DOTS Address Validation API that will allow more customizability, some additional dataset capabilities and much more. Please feel free to contact us if you are interested in our Address Validation offerings!
