- Products
- Solutions
- Developers
- Support
- About
- Login
- Blog
Use case · Returned mail
Catch undeliverable addresses before the mail run
Stop paying to print and mail addresses that come back. Validate, correct, and standardize every US address with one SQL call in your Snowflake account. Fixable addresses get corrected. Anything that can't be confirmed as deliverable gets flagged before the file goes out.
USPS CASS certified
25 years in production
Sub-100ms latency
99.999% uptime
SOC 2 Type II
Zero data retention
worksheet.sql
SQL
-- Validate your member file where it lives CALL APP_CODE.AV3_GET_BEST_MATCHES( 'SELECT CONSUMER_ROW_ID, BUSINESSNAME, ADDRESS, '' '' AS ADDRESS2, CITY, STATE, POSTALCODE FROM MY_SCHEMA.MEMBERS', 'MY_SCHEMA.AV3_RESULTS' ); -- → validated rows land back in your table ADDRESS DPV IsResidential 200 N Main St 1 False 520 W Taylor St 1 True
Standardized member addresses with DPV & RDI — no files to move.
Trusted by data teams at brands that can't afford bad records
The problem
A bad address costs you before you find out it's bad.
A missing apartment number. A ZIP that doesn't match the street. A misspelt street name. None of these stop the piece from printing. It prints, it gets postage, it ships, and it comes back. Then you pay for the correction and the second attempt.
Duplicates do the same thing quietly. The same household stored under two spellings is two records, two pieces, two postage charges.
The usual fix adds its own cost. Export the list, send it to an outside service, wait, re-import the cleaned file, reconcile what changed. Then do it again for the next run. Every cycle burns days against a drop date and leaves another copy of your customer data outside your environment.
- Undeliverable mailThe piece prints, gets postage, ships, and comes back — then you pay for the correction and the second attempt.
- Duplicate recordsThe same household stored under two spellings is two records, two pieces, two postage charges.
- Wrong or missing unit or ZIPA missing apartment number or a ZIP that doesn't match the street never stops the piece from printing.
- The export-and-reimport cycleEvery run burns days against a drop date and leaves another copy of your customer data outside your environment.
How it works
One SQL call. Validated data back in your table.
1
Install the native app
Get the app from the Snowflake Marketplace and grant the external access integration. The app runs inside your Snowflake account. You control the access and can revoke it at any time.
2
Select the addresses to validate
Pass a SELECT query with the addresses you want to validate and specify where you want the results saved. AV3 processes the records directly from your Snowflake data.
3
Use the validated output
AV3 creates a timestamped results table in your schema, based on the table name you supply in the call, with the full validation output. Your corrected and standardized address data is ready for the next step in your workflow — no batch files, CSV exports, or re-imports.
worksheet.sqlSQL
-- Validate a SELECT of your address records -- and save the results to an output table CALL APP_CODE.AV3_GET_BEST_MATCHES( 'SELECT CONSUMER_ROW_ID, BUSINESSNAME, ADDRESS, '' '' AS ADDRESS2, CITY, STATE, POSTALCODE FROM MY_SCHEMA.MY_TABLE', 'MY_SCHEMA.AV3_RESULTS' );
Install, call, use — versus contact sales, get a script, wait.
What the output includes
A decision per record, not just pass or fail.
Most validation services return valid or invalid. That doesn't tell you what to do with the record. A missing unit number is fixable. An address USPS has no delivery point for is not. Those two need opposite handling, and the difference is in the output. AV3 returns:
- Standardized components — street, city, state, ZIP+4, formatted to the CASS standard.
- DPV Description — whether the address is a confirmed delivery point, and if not, what the postal data says about it.
- DPV Notes Description — what was found or missing at the delivery point, so a fixable address gets corrected and kept rather than dropped.
- Residential Delivery Indicator (RDI) — flags whether the address is residential or not.
- Corrections Description — what changed and why, so the record is auditable.
Every kept address comes back in one standard form, so records that differed only by spelling now match. Duplicates surface before the file goes out, not after two pieces arrive at the same door.
Before → afterCASS certified
Input · as captured
1200 oak av apt3, santa barbra ca
AV3 output · standardized
1200 Oak Ave Apt 3
Santa Barbara, CA 93101-8523
Santa Barbara, CA 93101-8523
Input · not deliverable
742 Nowhere Rd, Lompoc CA
Who it's for
For the engineer who needs to clean the mailing list.
Marketing's last mailer had 18% come back undeliverable
Validate and standardize the list before export. Missing apartment numbers, wrong ZIPs, and addresses that are not deliverable get caught while there's still time to fix them.
Mail ops is paying to print pieces that never arrive
Flag the non-confirmed addresses before the file goes to the printer, so suppression happens upstream of production instead of after the returns land.
Compliance needs proof the notices went to valid addresses
Statements, policy documents, and required notices carry a delivery obligation. AV3 is USPS CASS certified, so the validation record behind each address meets a recognized postal standard.
Duplicate addresses are hiding in different formats
Standardize address variations into a consistent format, making duplicates easier to identify and merge before mailing the same location multiple times.
Security & architecture
You control the connection. We retain nothing.
AV3 runs as a native app inside your Snowflake account. When you call AV3_GET_BEST_MATCHES, the app sends address data to the Service Objects validation API through a single external access integration that you explicitly grant at setup. The API validates the address, returns the result, and retains nothing. Your data stays private and secure because it never sits in a third-party system and nothing is kept on our side.
The external access integration is the only connection. It is fully visible in the app manifest before you install, you grant it yourself, and you can revoke it at any time. There are no batch files in transit, no data sitting in a third-party bucket, and no CSV exports leaving your environment.
USPS CASS certified
SOC 2 Type II
PCI-DSS
Your Snowflake accountAV3_GET_BEST_MATCHES runs on your data, in your warehouse
Single external access integrationYou grant it, you can revoke it — the only connection
Service Objects validation APIValidates & returns the result — stores nothing
Pricing
Transparent pricing. Pay through the Snowflake Marketplace.
Standard
$2,500 / year
100,000 validations
- Full API output
- Sub-100ms latency, 99.999% uptime
- 20-day trial — 1,000 records, full output
Higher volume
Enterprise
$7,500 / year
500,000 validations
- Full API output
- Sub-100ms latency, 99.999% uptime
- 20-day trial — 1,000 records, full output
Marketplace Capacity Drawdown (MCD) supported.
Use your existing committed Snowflake spend to purchase. No separate procurement process, no additional contract. If your team has committed Snowflake credits, this purchase can draw from that balance.
Unlike most address validation vendors, pricing is visible before you talk to anyone. No "request a quote," no sales call required to learn what it costs.
Why Service Objects
Enterprise reliability, proven since 2001.
The validation infrastructure behind AV3 has run in production for over two decades — trusted by Fortune 500 data teams and backed by billions of validated transactions.
25 years
In production, since 2001
8 Billion+
Validations delivered
99.999%
Uptime SLA, financially backed
SOC 2 Type II
Independently audited security
Fortune 500
Clients rely on our data
Run it on your next list. Free for 20 days.
1,000 address validations with full API output. Install from the Snowflake Marketplace and see how many addresses are actually deliverable. No credit card needed.
USPS CASS certified
Sub-100ms latency
Zero data retention
Not the person who installs things in Snowflake? Forward this — your data team can have it running in an afternoon.
Questions, answered
Frequently asked questions
Does it guarantee delivery?
No. It confirms whether the address is a valid delivery point and corrects it to CASS standards. You stop mailing addresses that are incorrect. What the carrier does after handoff is outside AV3's control.
How much of my returned mail will this remove?
The share caused by bad or non-confirmed addresses, which is the part you can control. It won't catch a move you have no record of, or a refusal at the door. Run the trial on a real list to get your own number.
Does it find duplicates?
It standardizes every address to one form, which makes duplicate records match. Merging them is a step you run on the standardized output.
What does AV3 validate?
US mailing addresses. It validates, corrects, and standardizes each one, and confirms whether mail can actually be delivered there. If you need email validation, phone validation, or international addresses, Service Objects has separate APIs for those.
Does my data leave Snowflake?
Address data is sent to the Service Objects validation API through a single external access integration that you grant at setup. The API validates the address and returns the result. No data is stored, logged, or retained by Service Objects. There are no batch files, no CSV exports, and no data sitting in a third-party system.
Is the app CASS certified?
Yes. AV3 is USPS CASS certified, validated against official postal data that is updated monthly. Service Objects is also SOC 2 Type II and PCI-DSS certified.
What does the trial include?
A free 20-day trial with 1,000 address validations. The trial provides the full API output. At the end of 20 days or 1,000 records (whichever comes first), the trial stops. No automatic charges or conversions.
Can I pay using my existing Snowflake commitment?
Yes. AV3 supports Marketplace Capacity Drawdown (MCD). If your organization has committed Snowflake spend, this purchase can draw from that balance with no separate procurement process or contract.
How is this different from exporting a CSV and using an external validation service?
No export, no file transfer, no re-import, and far fewer resources to run. You call it from SQL and results land in a Snowflake table. The traditional batch-file workflow is eliminated entirely.