How It Works

For every record in a job, Address IQ takes the raw input address, validates it against official USPS postal reference data, and produces a standardized, deliverable result along with codes that describe what happened. Internally, the engine handles parsing, normalization, matching, and code assignment — you supply the input and configuration; Address IQ returns corrected, coded output.

Address IQ relies on official postal reference data - think of it like a phone book. It looks up something it knows (the input address) to discover something it doesn't (the correct, complete, deliverable form). Along the way it can correct misspelled street and city names, just as you might notice a name is spelled differently when you look it up.

When the software evaluates candidate matches for an address line, it scores each one for confidence and selects the best. To be chosen, a candidate must score highly and stand clearly apart from the alternatives - so ambiguous addresses are left unassigned rather than guessed at.

Every processed record carries a code that tells you what happened, which makes results easy to audit and route:

• Status codes describe how an assigned address differed from the input - for example, which components were corrected or added.
• Error codes explain why an address could not be assigned - such as a street not found, an invalid range, or a last line that doesn't agree.

Address IQ reads from files or database tables and can write to either. It is flexible about layout: input can be discrete fields (house number, street, city, and so on) or free-form multiline blocks, and output can be returned in whichever shape best serves the next step.

Output is available in four broad forms:

• Standardized data laid out the same way as your input fields.
• Raw data parsed into discrete components.
• Standardized data broken into discrete components.
• Additional codes and components generated during processing.

Address IQ uses multi-threaded processing so it can move through large files quickly, and it's built for production batch workloads. Real-world throughput depends on your hardware - available CPU cores, memory, and disk speed - as well as job configuration and how many jobs run at once.