Automated Address Verification: How It Works and What It Catches

Automated address verification is the process of checking every address in your file against USPS data before you print anything. Software standardizes the format, confirms the delivery point actually exists, appends the ZIP+4 and carrier route, and flags the records that cannot be delivered. Across the consumer files we process, approximately 94% of records pass address validation (CASS/DPV) on a first pass, meaning they come back fully confirmed to the unit. The remaining 6% are the ones quietly costing you postage.

One definition to fix before anything else, because it drives every number below: "pass" here means fully confirmed, not merely mailable. The 6% that does not pass is a mix of records that will still deliver, records that will reach a building but not a person, and records that cannot deliver at all. Telling those apart is the entire job.

The reason this matters is that a bad address does not announce itself. It prints, it gets inserted, it gets postage applied, and it comes back weeks later as a return or simply disappears. You paid for all of it either way.

We have run mail out of a single Lakeland, Florida production facility for 35+ years, and address verification is the step customers most often assume their list already had. This page explains what the process checks, what it misses, and how to build automatic address screening into a workflow so bad records never reach the press.

Request a list evaluation and we will tell you what your file actually looks like before you commit to a drop.

What Automated Address Verification Actually Checks

Address verification is not spell-check. It is a sequence of distinct lookups against USPS reference data, and each one answers a different question.

Standardization rewrites the address into the exact USPS format. "123 North Main Street, Apartment 4" becomes "123 N MAIN ST APT 4". This matters because every downstream process, including presort software and the barcode, expects the standardized form.

CASS matching confirms the address exists in the USPS address database and appends the ZIP+4, delivery point digits, and carrier route. CASS stands for Coding Accuracy Support System, and it is the USPS certification that mailing software must hold to produce automation-qualified mail.

Delivery Point Validation (DPV) is the step that separates a real address from a plausible one. CASS can confirm that 100 through 200 Main Street is a valid range. DPV confirms that 147 Main Street is an address a carrier actually delivers to. An address can be perfectly formatted, sit on a real street, carry a valid ZIP+4, and still be a delivery point that does not exist.

Secondary address validation checks the apartment, suite, or unit number. This is where large multi-tenant buildings generate most of their failures, because a record with the right street and the wrong unit looks correct to a human reviewer.

The Four DPV Outcomes and What Each One Means for a Mail File

DPV does not return a simple yes or no. It returns one of four codes, and the code determines both what the record costs to mail and whether it will arrive.

The distinction people usually get wrong is what these codes do to pricing. Y, D, and S all confirm at the primary number, so all three receive a ZIP+4 and a delivery point barcode and all three remain eligible for presorted automation prices. The difference between them is delivery risk, not postage. A D or an S will reach the building and may not reach the right unit.

N is the code that actually fails. An address that cannot be confirmed cannot carry the barcode that automation pricing depends on, so those records either mail at a higher price or come out of the file.

The USPS Products That Repair What DPV Flags

A D or an S code identifies a problem. Three additional USPS datasets exist to fix it, and a verification process that stops at DPV leaves that value on the table.

• SuiteLink appends missing secondary information to business addresses by matching the company name against USPS records. This is the direct remedy for a D code on a commercial file.
• LACSLink converts addresses that were renumbered or renamed by a local government, most often rural routes converted to street-style addressing for emergency services. Without it, those records look invalid when they are simply out of date.
• DSF2 (Delivery Sequence File, second generation) adds delivery attributes that DPV does not return: the walk sequence used for carrier route sortation, seasonal and drop designations, and whether the address is residential or business.

The practical sequence is CASS and DPV first to establish whether the address is real and how it is flagged, then SuiteLink and LACSLink to repair what can be repaired, then DSF2 if you need walk sequence or delivery type. Skipping the repair layer means suppressing records that were recoverable.

Real-Time Verification and Batch Verification Solve Different Problems

Most discussions of automated address verification collapse two different jobs into one. They are not interchangeable, and choosing the wrong one is why some organizations verify constantly and still mail bad records.

Real-time verification runs at the point of entry. Someone types an address into a web form, a call-center screen, or a CRM record, and the system checks it before the record saves. This prevents bad data from entering the database in the first place.

Batch verification runs against a file you already have. You hand over a list of 40,000 records, the whole file is processed at once, and you get back standardized addresses, DPV codes, and a suppression list. This is what happens before a mailing.

Real-time verification is a data-hygiene investment. Batch verification is a production requirement. An address verified correctly at signup in 2023 can be perfectly valid and completely wrong today, because the person moved and the address did not.

What Address Verification Does Not Catch

Verification confirms that an address is real and deliverable. It says nothing about whether your intended recipient still lives there. Those are separate questions answered by separate processes, and conflating them is the most expensive assumption in list management.

• Movers. A verified address remains verified after the occupant leaves. Catching the move requires NCOA processing against the USPS change-of-address database.
• Vacants. A carrier-identified unoccupied address is a valid delivery point that nobody is reading mail at. DPV returns a vacant indicator alongside the confirmation code, so the flag is available the moment you verify, but it is a separate field from Y/D/S/N and it is easy to ignore.
• Commercial mail receiving agencies. Mail drops and mailbox stores validate cleanly but often signal a recipient you did not intend to reach. DPV returns a CMRA indicator too, in the same separate-field pattern as the vacant flag.
• Deceased and do-not-mail suppressions. These come from separate suppression files, not from verification.
• Duplicates. Two records for the same household both verify perfectly and both get mailed.

We see the duplicate problem constantly. A file assembled from three sources will verify at a high rate and still mail the same household three times, because verification checks addresses one record at a time and has no opinion about the file as a whole.

If you want a picture of your file before committing to a drop, the mailing list health check is the fastest way to see where the losses are.

International and Hard-to-Verify Addresses

Four categories break the rules above, and each one produces a failure that looks like bad data but is not.

International addresses fall outside CASS and DPV entirely. Both systems are USPS products covering domestic delivery points, so a Canadian or European address will not return a DPV code at all. International records need their own validation path and their own postage treatment, and they should be split out of the domestic file rather than left in it to fail.

Rural route and highway contract addresses verify normally but often carry formatting that looks wrong to people reviewing a file by eye. Resist the urge to hand-correct these. The standardized form the software returns is the form USPS expects.

Military addresses using APO, FPO, and DPO designations are domestic for postage purposes but follow their own addressing rules, with the unit and box number carrying the routing and the AA, AE, or AP state code standing in for a geographic state. They are carried in the standard USPS ZIP+4 and CASS data with special match handling rather than in a separate database, and they should never be treated as foreign addresses.

New construction is the most common false negative we see. A house completed three months ago can be a real, occupied delivery point that has not yet propagated into the reference data, so it returns as unverifiable. In a file covering a fast-growing market, a cluster of failures in one ZIP code usually means new development rather than bad data.

The practical rule is to segment rather than discard. Records that cannot be verified are not automatically undeliverable; they are records you cannot presort at automation prices. Whether they are worth mailing at a higher rate is a business decision, and it should be an explicit one.

Move Update and the 95-Day Rule

If you mail at presorted prices, address verification is not optional. USPS requires that addresses on presorted First-Class and Marketing Mail be updated against change-of-address data within 95 days of the mailing date. This is the Move Update standard, and it is a condition of the discounted rate, not a best practice.

The practical consequence is a calendar constraint. A list processed in January does not satisfy Move Update for a drop in June. Organizations that mail quarterly need to build reprocessing into the schedule rather than treating the list as a fixed asset. If the mechanics of the move database are new to you, what NCOA is and how it works covers that side in detail.

The standard itself is published in the USPS Domestic Mail Manual section 602.5, which sets out the acceptable Move Update methods and the 95-day window. Current prices for each class live separately in USPS Notice 123, the Price List. Those are two different documents and it is worth citing the right one for the right question.

Why Verification Changes What You Pay

Address verification is usually sold as a quality measure. It is also a pricing mechanism, and that is the more compelling argument for most budgets.

Presorted automation prices require a standardized, CASS-coded address that DPV can confirm and that carries a valid delivery point barcode. Records that cannot be confirmed have no barcode, so they either mail at a higher price or come out of the drop. The cleaner the file, the deeper it presorts and the lower the effective per-piece cost.

Marketing Mail letter presort Mixed pricing is $0.467 per piece per USPS Notice 123 effective July 12, 2026. Run the arithmetic on a modest failure rate and the case makes itself.

Take a 50,000-record file that nobody screens before the drop. If 6% of it comes back short of a full confirmation, that is 3,000 pieces going into the mail with known defects. At $0.467 per piece, the postage riding on those 3,000 pieces is $1,401, and every one of them was printed, inserted, and handled before it shipped. Production cost sits on top of that number, not inside it.

Treat $1,401 as a floor rather than the real exposure. It prices all 3,000 at the Mixed automation rate, which only the D and S records actually earn. The N-coded share of that group carries no barcode, so it either mails at a higher non-automation price or comes out of the file entirely, and either outcome costs more per record than the figure above.

The groups also fail differently, which is the point of screening. Some records in that 6% are fine and simply looked wrong, usually rural-route formatting or new construction the reference data has not caught up to. The D and S records will reach the building without reliably reaching the person. The N records mostly will not arrive at all. Verification is what lets you separate them before you pay to mail any of them, rather than after the returns come back.

That is one drop. An organization mailing quarterly repeats the exercise four times a year against a list that is decaying the entire time.

Building Automatic Address Screening Into a Workflow

Automatic address screening works when it is a step in a process rather than a task somebody remembers. The organizations that get this right treat it the same way they treat a backup: scheduled, unattended, and verified by exception.

The pattern below is the one we see hold up across customers who mail on a schedule.

1. Verify at entry. Every form, portal, and call-center screen validates the address before the record saves. This is the cheapest possible place to catch an error.
2. Re-verify on a schedule. Run the full database through batch verification quarterly, independent of any specific campaign.
3. Process before every drop. Run CASS, DPV, and NCOA against the actual mailing file, not the master database, within the Move Update window.
4. Suppress by rule, not by judgment. Decide in advance what happens to an N code, a vacant flag, and a duplicate, then apply it automatically.
5. Keep the audit trail. Retain the processing report for each drop. It is the evidence behind your postage claim.

Automatic address screening at this level is not a heavy lift once it is configured. The work is in deciding the rules, and the rules only need deciding once.

For most of our customers, steps 3 through 5 happen here as part of the job. The data services side of the operation runs the file, returns the report, and hands a clean list to production, which means the customer is not maintaining verification software to send mail four times a year.

How to Evaluate an Approach

Most organizations arrive at this topic asking which product to buy. The more useful first question is which of three models fits, because that decision constrains everything after it.

Licensing software in house makes sense when address volume is continuous, engineering capacity exists to maintain the integration, and the data cannot leave your environment for regulatory reasons. The cost is not the license; it is owning the CASS cycle, because certification requires the reference data be current and USPS reissues it monthly.

Using a service bureau makes sense when mailing is periodic rather than continuous. Somebody else carries the certification, the monthly data updates, and the presort software, and you pay per file or per thousand records. For an organization mailing quarterly, this is almost always cheaper than maintaining the capability.

A hybrid covers the common real case: lightweight real-time checking at the point of entry to keep the database clean, with full CASS, DPV, NCOA, and DSF2 processing performed on the mailing file before each drop.

Whichever model you pick, these are the questions worth asking, and none of them require naming a product.

• Is the process CASS certified, and how often is the underlying USPS data refreshed?
• Does it return the DPV code itself, or only a pass/fail? You cannot act on a verdict you cannot see.
• Does it include SuiteLink, LACSLink, and DSF2, or only CASS and DPV?
• Is NCOA processing included, and does the output document Move Update compliance for your postage claim?
• What does it return for a record it cannot confirm: a flag you can review, or a silent deletion?
• Does the output survive into the presort, or does the file get re-keyed somewhere downstream?

That last question catches more problems than the rest combined. A file can be verified perfectly and then lose its coding in a handoff between systems, which is why the answer matters more than the vendor.

How This Works in Production at MPA

We process address files across all 50 states from one facility, and the workflow is deliberately unglamorous. The file comes in, gets standardized and CASS coded, runs DPV, gets NCOA processed against the USPS move database, and gets deduplicated. What comes back is a mailable file, a suppression list, and a report showing exactly what happened to every record.

Across the consumer files we run, approximately 94% of records pass address validation (CASS/DPV), and after NCOA processing the deliverability rate on the mailable portion runs around 98.5%. Those are our own observed rates rather than an industry benchmark, but they are what a healthy file looks like. If your list is materially below them, the gap is usually age rather than sourcing.

Because we produce the mail as well as process the data, the verification step is not a handoff. The file that gets presorted is the file that came out of verification, and the barcode on the piece reflects the delivery point the software confirmed. That continuity is the practical argument for running data and mailing services under one roof: nobody has to reconcile two versions of the list.

Talk to us about your list and we will process a sample before you commit to a full drop.

Frequently Asked Questions

What is automated address verification?

Automated address verification is software-driven checking of a postal address against USPS reference data. It standardizes the format, confirms the delivery point exists through DPV, and appends the ZIP+4 and carrier route needed for presorted mail. It runs either in real time at data entry or in batch against an existing file.

Is CASS certification the same as address verification?

No. CASS is the USPS certification a software product holds, confirming it can standardize and code addresses accurately. Address verification is the process that software performs. A CASS-certified process is the requirement for automation pricing, but the certification belongs to the tool, not to your file.

Does address verification tell me if someone moved?

No, and this is the most common misunderstanding. Verification confirms the address is real and deliverable, not that your recipient is still there. Catching movers requires NCOA processing against the USPS change-of-address database, which is a separate step.

How often should a mailing list be verified?

Before every presorted drop, without exception, because USPS Move Update requires address data updated within 95 days of the mailing date. Independently of campaigns, quarterly batch verification of the full database keeps decay from compounding between mailings.

What does a DPV code of N mean?

N means USPS data does not confirm the address as written. The record carries no delivery point barcode, so it cannot presort at automation prices, and most N records should be suppressed rather than mailed. Review them before discarding when the file covers new construction, because a genuinely new address returns N until the USPS reference data catches up.

Can address verification fix an incomplete apartment number?

Not reliably. Verification will flag the record with a D code when the building is confirmed but the unit is missing, or an S code when the supplied unit is wrong. Correcting it requires sourcing the missing information; the software can identify the gap but cannot invent the unit.

Does verifying addresses actually lower postage costs?

Yes, in two ways. Verified, DPV-confirmed addresses qualify for presorted automation prices that unverified records cannot earn, and suppressing undeliverable records removes the postage and production spend those pieces would have consumed.

Does address verification work on international addresses?

Not through CASS and DPV, which are USPS systems covering domestic delivery points only. International records return no DPV code and need a separate validation path. The practical approach is to split them out of the domestic file so they do not distort your pass rate or interfere with presorting.

Where to Start

Address verification is the least visible step in a mailing and the one with the clearest return. If your list has not been processed recently, that is where the next drop is losing money.

We have handled mailing data for more than 700 lifetime business customers since 1989. Request a quote or send us a sample of your file and we will show you what verification finds.