How to Vet an Audience Data Vendor: Sourcing to Suppression
The surest way to tell whether an audience data vendor is trustworthy is to ask it to show its work: where each record came from, how opt-outs and deletion requests reach your file, how often the data is refreshed, and what the contract lets you do with it. If you’re figuring out how to vet an audience data vendor, treat a vague or evasive answer to any of those questions as a reason to walk away, however attractive the offer looks.
That caution is well founded. When the FTC studied nine data brokers, it found a fundamental lack of transparency across the industry, with data often passing through multiple layers of brokers before it reached the company using it. By the time records land in your CRM or a custom audience, they can be several steps removed from the person who first shared them. If those records are stale or were never properly permissioned, your sender reputation, your inbox placement and your legal exposure absorb the cost.
This guide walks through the checks that matter, in the order you’ll meet them. It covers sourcing and proof of origin, how consent and opt-outs flow through to your file, refresh cadence and data decay, and hashed suppression and sample testing to protect your existing list before any match runs. It then turns to the contract and security terms worth insisting on, and closes with the red flags that should end the conversation.
The short answer: a good vendor can show its work
A data vendor worth hiring can explain, in writing, five things: where its records come from, how opt-outs and deletion requests flow through to the files it sells, how often the data is refreshed, how your suppression list is handled, and what the contract lets you do with the data. If a vendor can’t put those answers on paper, you don’t yet have enough to make a decision.
That burden falls on you as the buyer, and there’s a good reason. When the Federal Trade Commission studied nine data brokers, it found that the industry operates with a fundamental lack of transparency. The same study found that brokers collect and store billions of data elements covering nearly every U.S. consumer. It also found that data often passes through several layers of brokers before it reaches the company that finally uses it. Each extra layer makes it harder to know where a record started, whether consent came with it, and whether an opt-out ever caught up with it.
For direct-response marketers, affiliates and agencies, those gaps cause real problems. Records with an unclear origin can hurt your sender reputation, put stale contacts in your file and leave you holding data you can’t defend.
Treat the sections below as a call script. Each one covers one part of the vendor’s process:
- The question to ask
- What a credible answer sounds like
- What should make you pause
If you’d like to see how we approach custom audiences and data append ourselves, our data services overview walks through it.
This guide covers buyer questions only. It isn’t legal advice, and state rules vary, so check your specific obligations with counsel.
Sourcing: where did these records come from, and can they prove it?
Start with sourcing, because every later question depends on it. A trustworthy vendor can tell you where a record came from and when it was collected. That matters more than it might seem. When the FTC studied the industry, it found that consumer data often passes through multiple layers of data brokers who share with each other. The “source” a vendor names may itself be another reseller.
Sourcing questions to ask every vendor
Source categories named
Ask the vendor to name where records come from, such as signups, publisher partners or public records.
Record-level source and date
Each record should carry its own source and collection date, not a vague database-wide claim.
How many layers deep
Data often passes through several brokers. Ask whether the vendor collected the data itself or bought it from another broker.
Observed versus inferred fields
Ask which fields come straight from the source and which are modeled from combined online and offline data.
Supplier consent checks
Ask how the vendor confirms its own suppliers collected the data with proper consent.
Questions to ask
- Which source categories feed the file? Ask the vendor to name them. Common categories are first-party signups, publisher partners, public records and licensed compilers. A vague answer like “proprietary network” is not a source.
- Does each record keep its origin? Ask how every record carries its source and collection date through processing. If that lineage is lost when files are merged, nobody can trace a complaint or an opt-out back to where it started.
- Do you check your suppliers’ consent? Regulators expect oversight that reaches upstream. In a 2026 settlement, the FTC required Kochava and its subsidiary to run a sensitive location data program and report incidents when third parties share data in violation of contract terms. Ask what your vendor does when one of its own suppliers breaks the rules.
- Which fields are observed and which are inferred? The FTC also found that brokers combine online and offline data. Ask which attributes were collected directly and which were modeled or guessed. An inferred income band should not be treated like a confirmed postal address.
The sample-record test
Pick a handful of records from a sample file. For each one, ask the vendor to show where it came from, when it was collected and under what permission. A vendor that can’t answer “where did this record come from?” for a sample has not documented its sourcing. That gap will follow the data into your campaigns.
If you want to see how sourced records get matched to your existing file, here is how audience data append works.

Consent and opt-outs: how do deletion and opt-out requests reach your file?
A vendor should be able to explain, step by step, how a consumer’s opt-out or deletion request moves from their system to the records already sitting in your CRM. If they can’t, those requests probably stop at their door, and the records you bought keep circulating.
California gives you a concrete benchmark. Data brokers there register with the California Privacy Protection Agency (CalPrivacy) and set up an account on the Delete Request and Opt-out Platform (DROP). Under the California DROP regulations, registered brokers must access DROP at least once every 45 days, delete the personal information of matching consumers and direct their service providers and contractors to do the same. The agency lists the registration fee at $6,000.
Questions to ask
- Are you registered as a California data broker? A plain yes or no, with the registration you can check yourself.
- How often do you pull DROP lists? “At least every 45 days” is the legal floor, not a selling point.
- How do deletions and opt-outs reach files you already delivered to clients? Ask whether they send suppression updates, how often, and in what format.
- How was consent captured for any sensitive categories? Ask to see the actual consent language and when it was shown to the consumer.
That last question matters more than many buyers realize. The FTC’s 2026 order against Kochava bars the company from selling sensitive location data without the consumer’s affirmative express consent. A vendor that can’t produce clear consent records for sensitive data is passing that risk on to you. If you also use visitor or lead identification tools, the same thinking applies, and our guide to privacy considerations for lead identification tools covers the overlap.
Keep the scope in mind. DROP obligations come from California law and cover California consumers’ data. Other states set different rules, so ask the vendor which state data broker registries it’s on and how it handles requests from each.
Refresh cadence: how fresh is the data and how is decay handled?
Ask two questions together: how often the vendor refreshes its data, and how long it keeps data before deleting it. A vendor that only talks about refreshing may still be holding old records with no end date. The FTC’s study of nine data brokers found that some stored consumer data indefinitely, which may create security risks. Get the vendor’s retention limits in writing alongside its refresh schedule.
Questions to put to the vendor
- Record-level dates: Can you show a “last verified” or “last seen” date on each record or field? One refresh date for the whole database tells you very little about the specific row you are paying for.
- Decay handling: How do you remove dead email addresses and households that have moved before you deliver a file? Do I get credit for records that bounce?
- Different schedules for different fields: Do B2B fields, where job changes are the main cause of decay, follow a different refresh schedule from consumer fields, where address moves are the main cause? If you are buying B2B and first-party data append, ask how often job titles and company emails are rechecked.
- Retention: How long do you keep data, and what triggers deletion?
Why this matters for your inbox placement
Stale or inferred email addresses raise the risk of bounces and complaints. Mailing too many of them can hurt your sender reputation across your whole program. Treat appended emails with the same caution as any new signup source. Introduce them slowly, watch bounce and complaint rates closely, and suppress problems quickly. To see where appended data fits next to your other acquisition channels, compare it with these signup sources ranked by deliverability risk.
Hashed suppression and sample testing: protect your list before you match
Before any vendor appends or matches against your data, settle two things. First, how your suppression and customer files will travel. Second, how you will check their accuracy yourself. A headline match rate tells you very little. A test you control tells you a lot.
Send hashed identifiers, not plain-text addresses
Ask whether the vendor accepts your suppression list and existing-customer file as hashed identifiers, such as normalized and hashed email addresses. If it does, plain-text addresses never have to leave your systems. That is a real improvement, but it is not a privacy shield. The FTC has stated plainly that hashes are not anonymous and can still be used to identify people. Treat a hashed file as personal data. Your contract should limit what the vendor may do with it and say how long they may keep it.
Get the normalization rules in writing
A hash changes completely if the input changes by a single character. “Jane@Example.com ” with a trailing space produces a different hash from “jane@example.com.” Ask the vendor to document every normalization step: lowercasing, trimming spaces, and any other rules they apply. Then apply the same steps on your side before you hash. Mismatched formatting quietly breaks matches. It can also let suppressed contacts slip back into a campaign.
Run a blind sample before you buy
Build a holdout file of records where you already know the correct answers. Send it to the vendor without saying what you know, then score what comes back:
- Accuracy: how many appended values match what you know to be true?
- False matches: did the vendor attach data to records that should not have matched?
- Coverage versus quality: a high fill rate filled with wrong answers is worse than a modest one you can trust.
Ask how each match is made
There are two broad approaches. A deterministic match joins records on an exact identifier, such as an email, postal address or phone number. A probabilistic match infers a link from patterns and similarities. Ask which method the vendor used for each record, and whether every appended field carries a confidence level. That way you can decide which fields are safe to use when you build custom audiences for affiliates and direct-response senders.

Contract and security: permitted use, retention, resale and safeguards
Put the rules in writing before any file changes hands. Two state laws give you a solid baseline for what that contract and the vendor’s security setup should include.
Contract terms to require in writing
Limited, specific purposes
The agreement should name exactly what the data may be used for.
Opt-out compliance
The vendor should honor opt-out requests you forward and keep reasonable security.
Right to stop misuse
You should be able to take reasonable steps to stop and fix unauthorized use.
Notice if they can't comply
The vendor must tell you if it can no longer meet its privacy obligations.
Due diligence on record
Keep proof you checked the vendor; under California's rule, due diligence matters if the vendor later misuses data.
Under California’s CCPA regulations, a business that sells or shares personal information with a third party must have an agreement that spells out the limited and specific purposes the data may be used for. The same rule requires the third party to honor opt-outs and provide reasonable security. It also gives the business the right to stop and remediate unauthorized use, and it requires the third party to tell the business if it can no longer meet its obligations. Due diligence matters too. The checks you run on a vendor count toward whether you can claim you had no reason to believe your data was being misused.
Clauses to write in
- Permitted uses: name each allowed purpose. Anything not listed is off limits.
- No resale or co-mingling: the vendor may not resell your files or fold them into its own database.
- Retention and deletion deadlines: set how long the vendor may keep your data and when it must delete it, with written confirmation of deletion.
- Audit rights: you can verify that the vendor is following the contract.
- Indemnities: state who covers the cost if the vendor breaks the agreement.
Ask your own counsel how strict each clause should be for your level of risk. The same discipline applies in reverse if you license your own audience. Data monetization done with clear permitted uses protects both sides.
The security question to ask
Vermont’s data broker statute requires brokers to maintain a written, comprehensive information security program with administrative, technical and physical safeguards. The program must be reviewed at least once a year or after material business changes. Brokers must also supervise their service providers and hold them by contract to appropriate security measures.
That gives you a simple benchmark for any vendor: “Show us your written security program, and show us how you hold your own subcontractors to it.” A vendor that takes security seriously should be able to answer both parts.
Red flags that should end the conversation, and a sensible next step
If a vendor can’t or won’t answer the basic questions in this guide, you have your answer. Any one of these warning signs is reason enough to walk away:
- No answer on sourcing. The vendor won’t name its source categories or show source and collection-date fields at the record level. The FTC’s study of nine data brokers found that information often passes through multiple layers of brokers. That is why “it’s from our proprietary network” isn’t a real answer.
- Guaranteed match rates or guaranteed inbox results. Match rates depend on the quality and makeup of your own file. Inbox placement depends on your sending history, authentication and engagement. Nobody can honestly promise either number in advance.
- Resale or reuse of your data. Watch for contract language that lets the vendor add your customer or suppression files to its own database, use them to “enhance” its products, or share them with other clients. Your suppression file should only ever be used to suppress.
- No registration where it’s required. If the vendor clearly sells data about people it has no direct relationship with, check whether it appears on California’s data broker registry. Registered brokers must access the state’s DROP system at least every 45 days and delete matched consumers’ personal information. A missing listing raises a real question about whether deletion requests ever reach the records you’d be buying.
A vendor that treats these questions as routine and answers them in writing is usually one worth testing. A vendor that treats them as an inconvenience is telling you how it will handle problems later.
A sensible next step
Digital Bulldogs offers audience data append, custom audiences and identity resolution, along with deliverability work such as list hygiene and inbox placement monitoring. If you’d like a second opinion on a vendor’s answers, or help planning a hashed suppression and sample test before you commit, you can book a free consult. Bring the vendor’s responses and draft contract, and we’ll go through them with you.
Frequently asked questions
What questions should I ask a data vendor about sourcing?
Ask where each record came from, how it was collected, and how many companies handled it before it reached the vendor. The FTC’s study of nine data brokers found that personal data often passes through multiple layers of brokers and that online and offline data are routinely combined. A trustworthy vendor can trace a record back to where it was first collected and show the notice or consent that came with it.
Also ask whether the file contains sensitive categories, such as precise location, and what consent covers them. In its order against Kochava, the FTC banned selling sensitive location data without affirmative express consent. It also required incident reports when third parties share data in violation of contract terms. If a vendor can’t explain the consent basis for sensitive data, don’t buy from that vendor.
Is sending a hashed email list to a vendor anonymous?
No. The FTC has said directly that hashes are not anonymous and can still identify people. Data counts as anonymous only when it can never be linked back to a person. A hashed email exists so it can be matched, so it doesn’t meet that bar. Treat a hashed suppression or match file as personal information. Share it only under a written agreement, limit what the vendor may do with it, and don’t assume hashing makes it less sensitive.
How often do data brokers have to process deletion requests in California?
Registered data brokers must access California’s Delete Request and Opt-Out Platform, known as DROP, at least every 45 days. Each time, they must delete the personal information of matched consumers and tell their service providers and contractors to delete it too. If a vendor sells data about Californians, ask how it runs that cycle. Also ask how it makes sure deleted records don’t show up again in files it delivers to you.
What should a data vendor contract include?
California’s regulations list the terms required when personal information is sold or shared with a third party. They include:
- limited, specific purposes for using the data
- a commitment to honor opt-outs
- reasonable security
- the right to stop and fix unauthorized use
- notice if the third party can no longer meet its obligations
Treat those as the minimum, and use the same list to check any vendor contract you’re offered.
Beyond those terms, put retention limits, resale restrictions and rules on onward sharing in writing. The same regulation says due diligence counts toward whether a business had reason to believe its data was being misused. So keep a record of the questions you asked a vendor and the answers you got. If you’d like help reviewing a vendor before you sign, you can book a free consult with Digital Bulldogs.
Do data brokers have to keep a written security program?
In Vermont, yes. Data brokers there must maintain a written comprehensive information security program with administrative, technical and physical safeguards. They must review it at least once a year or after material changes, and supervise their service providers by contract. Even if Vermont’s law doesn’t cover your vendor, ask for a summary of its security program and how often it’s reviewed. A vendor that takes security seriously should be able to answer.
Sources
- FTC Recommends Congress Require the Data Broker Industry to be More Transparent and Give Consumers Greater Control Over Their Personal Information · Federal Trade Commission · 2014-05-27
- Data Broker Registration and Accessible Deletion Mechanism (text of regulations) · California Privacy Protection Agency (CalPrivacy)
- Cal. Code Regs. Tit. 11, § 7053 – Contract Requirements for Third Parties · Legal Information Institute (California Code of Regulations text)
- No, hashing still doesn't make your data anonymous · Federal Trade Commission, Office of Technology · 2024-07-24
- FTC to Ban Kochava and Subsidiary from Selling Sensitive Location Data to Settle Charges They Sold Location Data Linked to Millions of Mobile Devices · Federal Trade Commission · 2026-05-04
- 9 V.S.A. § 2447. Data broker duty to protect information; standards; technical requirements · Vermont General Assembly

