Analytics

First-party data

Data a business collected directly from its own customers.

CRM records, purchase history, site behaviour, email lists. Gathered by the business itself with a lawful basis, as opposed to bought from a broker. As third-party cookies recede it has become the main input for audience building and measurement.

Key takeaways

  • Uploading a customer list to an ad platform is a disclosure to a third party, whatever the interface makes it feel like.
  • Consent obtained for a newsletter does not automatically cover advertising. The purpose has to have been stated.
  • Hashing limits what the platform learns. It does not make the upload lawful on its own.
  • The obligation sits with the client as controller, and the agency is usually a processor. Both need that in writing.
  • Settle ownership and deletion before anything is uploaded, not at the end of the relationship.

Why it became the centre of everything

For most of the last two decades, targeting and measurement ran on identifiers the business did not own: third-party cookies, device identifiers, data bought from brokers. That layer has been dismantled by a combination of regulation, browser policy and platform decisions, and it is not coming back.

What is left is what a business collected itself, from people who dealt with it directly. That is a smaller dataset and a better one: it describes actual customers rather than inferred segments, and it does not depend on a technology anybody can switch off.

The practical uses are concrete. Matched audiences for targeting and suppression, lookalike modelling built from real customers, offline conversion import so the ad platform learns which leads were worth having, and measurement that survives the loss of client-side tracking.

Which makes it the client most durable marketing asset, and the reason the governance around it matters more than the tactics that use it.

Uploading a list is a disclosure, not a setting

The interfaces make this feel like a configuration step. Choose a file, map the columns, name the audience. It is not a configuration step: it is sending personal data about identified people to a third party, and it needs a lawful basis like any other disclosure.

The specific trap is purpose. Consent collected for a newsletter covers sending a newsletter. It does not automatically cover passing the same address to an advertising platform to build a matched audience, unless the privacy notice the person actually read said so.

Legitimate interest is sometimes argued as the basis instead, and sometimes correctly, but it requires a balancing assessment rather than an assumption, and in the EU the ePrivacy rules sit alongside GDPR rather than being replaced by it.

The point for an agency is narrower than the law: it is that the question has to be asked before the upload, of someone at the client who can answer it. An agency that uploads because the account manager sent a spreadsheet has made a decision that was not theirs to make.

What hashing does and does not do

Platforms require identifiers to be hashed before upload, usually SHA-256, and the interfaces present this as the privacy protection. It is a real protection and it is narrower than it appears.

What it does is prevent the platform reading the plaintext addresses of people who are not already its users. What it does not do is prevent matching, since the platform hashes its own user list the same way and compares. Matching is the entire point, so a mechanism that prevented it would be useless.

It also does not change the legal character of the transfer. Hashed personal data is still personal data where it can be matched back to an individual, which is precisely the situation here. The GDPR treats pseudonymisation as a safeguard, not as an exemption.

So hashing belongs in the conversation as a reason the transfer is proportionate, not as a reason it does not need a basis. The distinction matters because the first is defensible and the second is not.

What to settle before touching it

Three things, agreed in writing, before any list moves anywhere. Who is the controller and who is the processor, which for an agency is almost always the client and the agency respectively. What the agency may do with the data. And what happens to it on the day the relationship ends.

That third one is the one nobody raises at kickoff and everybody needs at the end. Audiences built inside the client ad accounts stay with the client. Lists sitting in an agency tool, a shared drive or somebody laptop do not, and an agency holding a former client customer list with no agreement about it is a problem waiting for a lawyer.

A data processing agreement covers most of this and exists for exactly this purpose. It is not exotic paperwork; it is the normal instrument, and it is far easier to sign in week one than to negotiate during an offboarding.

The other habit worth having is to build audiences inside the client own accounts rather than the agency. It is better for the client, it removes the offboarding question entirely, and it costs nothing except the ability to reuse the work somewhere it should not be reused anyway.

What it means for an agency

Agree in writing what the agency may do with it and what happens on the day the relationship ends, before anything is uploaded anywhere. An agency holding a client customer list with no agreement is a problem waiting for a lawyer.

Where it goes wrong

  • Treating the upload as a configuration step rather than a disclosure of personal data to a third party.
  • Relying on consent collected for a different purpose, most often a newsletter signup.
  • Citing hashing as the reason no lawful basis is needed, which confuses a safeguard with an exemption.
  • Building audiences inside agency accounts, which turns offboarding into a negotiation about data.

In short

First-party data is the one marketing asset a client genuinely owns, which is what makes it valuable and what makes the paperwork around it worth doing properly. Building audiences inside the client own accounts, with an agreement signed before anything moves, removes almost every version of this problem in advance.

Frequently asked questions

Can I upload a client customer list to Google or Meta?

Only with a lawful basis that covers advertising, and covered by the privacy notice those customers actually saw. Consent given for a newsletter does not extend to passing the address to an ad platform unless the notice said so. The obligation sits with the client as controller, which means the question has to be answered by someone at the client before the upload rather than assumed by whoever has the file.

Does hashing make a customer list upload compliant?

No. Hashing stops the platform reading the addresses of people who are not already its users, which is a real protection, but matching still happens because the platform hashes its own list the same way. Hashed data that can be matched back to an individual is still personal data, and GDPR treats pseudonymisation as a safeguard rather than an exemption.

Who owns the audiences an agency builds?

Settle it in writing before building any. Audiences created inside the client own ad accounts stay with the client, which is the clean arrangement. Lists held in agency tools or shared drives do not, and an agency still holding a former client customer data with no agreement covering it is exposed. A data processing agreement signed in week one costs nothing; negotiating during an offboarding costs a great deal.

Why has first-party data become so important?

Because the layer it replaced has been dismantled. Third-party cookies, device identifiers and broker data have been eroded by regulation, browser policy and platform decisions, and are not returning. What remains is what a business collected directly from people who dealt with it, which is a smaller dataset describing real customers rather than inferred segments, and it does not depend on a technology somebody else can switch off.

Related terms

Sources

Last reviewed

Keplent keeps track of these for you.

It holds what each client agreed to, watches what their campaigns actually did, and tells you in plain sentences what needs you this week.

Join the waitlist