First-party data and Third-party data at a glance
First-party data is information a business collects directly: transactions, account details, email engagement, app and website behaviour, support history, survey answers and preferences. The business knows where it came from, what people were told and what they agreed to, and can keep it accurate because it sees the customer's behaviour first-hand. It powers RemarketingAdvertising to previous visitors or engagers., customer-list audiences, personalisation, lifetime value analysis and measurement. A customer data platform (CDPUnified first-party customer profiles.) is one common place to unify it.
Third-party data is assembled by organisations that do not have a direct relationship with the people described. Data brokers and some platforms combine sources such as cookies, device identifiers, public records, purchase panels and partner feeds into audience segments sold to many buyers. Historically, data management platforms (DMPPlatform that organises audience data for targeting.) were the main way advertisers bought and activated these segments. The appeal is reach: audiences the business could not build itself.
The ground has shifted. Browsers have restricted third-party cookies and cross-site identifiers, mobile operating systems have added consent prompts for tracking, and regulators in many regions have scrutinised how data is collected and resold. The policies keep changing, so plans should rest on data the business controls. Privacy law and consent mechanics belong to the privacy and consent guide; this page gives no legal advice.
Compare table
First-party data vs Third-party data
| Row label | First-party data | Third-party data |
|---|---|---|
| Source | Collected directly from the business's own customers and visitors | Collected by other organisations and aggregated for sale or licence |
| Consent and lawful basis | Established by the business at collection, with records it controls | Depends on the supplier's collection practices, which the buyer must verify |
| Accuracy | High for behaviour the business observes; depends on data hygiene | Variable; often modelled or inferred, and quickly dated |
| Scale | Limited to people who have interacted with the business | Large, covering people who have never heard of the business |
| Cost | Cost of collection, systems and governance; no per-segment fee | Licence or per-use fees, plus integration |
| Durability under privacy changes | High, especially where collected with consent and stored server-side | Low; affected by browser, operating system and regulatory restrictions |
| Activation options | Email, customer lists on ad platforms, onsite personalisation, analytics | Programmatic segments, enrichment of profiles, prospecting audiences |
| Measurement use | Links marketing to real outcomes such as orders and retention | Limited; rarely connects to the business's own outcomes |
| Governance | Under the business's own policies and controls | Dependent on supplier contracts, audits and documentation |
The rows on consent and durability explain why the balance has moved. Data the business cannot vouch for is a liability as well as an asset.
When to choose First-party data
First-party data should be the foundation of almost every targeting and measurement plan. It describes people who have already engaged, so it is accurate and relevant: a list of recent purchasers, people who abandoned a basket, subscribers who opened the last few emails. It can be activated across channels, from email to customer-list audiences on ad platforms; the custom audiences guide shows how one platform uses it, and others document similar options.
It is also the basis of honest measurement. Orders, leads and renewals recorded in the business's own systems are the outcomes that matter, and only first-party data connects marketing activity to them. Server-side collection through the business's own endpoints keeps more of that data under its control.
Building it takes deliberate effort. Give people a reason to share information, ask only for what you will use, record consent and preferences properly, and keep the data in systems that talk to each other. The data flow and integrations guide covers how data should move between tools. The UK ICO's direct marketing guidance and the GDPR give the general direction on consent and transparency for marketing use. For customer lists uploaded to ad platforms, the guide to customer lists across ad platforms turns that direction into one consented, refreshed register.
When to choose Third-party data
Third-party data still has uses, but they are narrower than they were. The clearest is reach: prospecting to people who have never interacted with the business, where first-party data has nothing to offer. Some segments, such as broad demographic or contextual groups, can help a new brand or a new market find a starting audience.
The second use is enrichment: adding attributes such as company size or industry to business records the business already holds, where suppliers can document their sources. Business-to-business marketing often relies on this kind of firmographic data.
Use it with checks. Ask suppliers how the data was collected, what people were told, how often it is refreshed and how accuracy is measured. Test segments against a control before scaling spend, because modelled attributes are often less accurate than they appear. Prefer contextual targetingChoosing where ads appear from the content of the page, app or video being viewed rather than from data about the viewer. or platform audiences built on the platform's own first-party data where they meet the need; the guide to planning contextual targeting in DV360 sets out how to test context against a bought or platform segment on the same outcome. And keep a record of what was bought and how it was used, so the business can answer questions about it later.
Decision table
| Situation | Choose | Why |
|---|---|---|
| Remarketing to visitors and customers | First-party | Accurate, consented and controlled |
| Measuring marketing against orders, leads or retention | First-party | Only own data connects to real outcomes |
| Prospecting to people who have never engaged | Third-party or contextual, with checks | First-party data cannot reach them |
| Enriching business records with firmographic attributes | Third-party, from documented sources | Adds attributes the business cannot collect itself |
| Supplier cannot explain source or consent basis | Avoid | The risk falls on the buyer |
| Building audiences for personalisation | First-party | Reflects actual behaviour and preferences |
| Unifying data from many tools | First-party in a CDP or warehouse | Governance and identity resolution under one roof |
Building a plan around each
A durable data plan starts with an inventory of first-party sources: where data is collected, what consent covers, where it is stored and which tools use it. Gaps usually appear quickly, such as sign-ups without recorded consent, customer records split across tools, or site behaviour never linked to purchases. Close them in order of value, starting with the data that feeds the largest decisions.
Next, decide where the unified data lives. A CRM may be enough for a sales-led business; a customer data platform suits businesses with many tools and channels, a choice the CDP versus CRM comparison explores. Either way, define ownership, retention and access rules.
Then set rules for any third-party data: approved suppliers, documented sources, a test before scale, and a review date. Finally, plan for change. Assume that identifiers will keep becoming scarcer and that regulators will keep scrutinising resale, and favour approaches that would survive both: consented first-party collection, contextual targeting and aggregate measurement.
A pet food retailer rebalances its data
Barkley Pet Supplies spent £18,000 a month on prospecting audiences built from purchased interest segments, and found that only 22% of buyers from those audiences ordered a second time. It audits its own data and finds 65,000 customers, but consent records for only 40,000. It runs a preference-centre campaign, rebuilds its email consent flow and uploads consented customer lists to two ad platforms to build similar-audience prospecting. Within two quarters it cuts purchased segments to a single tested supplier at £4,000 a month, and repeat purchase among new customers from first-party-based audiences reaches 31%.
Common mistakes
- Buying segments without asking where they came from, then discovering the consent basis cannot be shown.
- Collecting first-party data without recording consent, which makes it hard to use later.
- Treating third-party segments as accurate because they are labelled precisely.
- Storing customer data in disconnected tools, so it cannot be activated or measured.
- Asking for more data than you use, adding risk without value.
- Assuming platform policies will stay the same, and building plans on identifiers that may disappear.
- Treating this comparison as legal advice. Lawful basisOne of the six grounds the GDPR accepts for processing personal data, such as consent, contract or legitimate interests, chosen for each purpose. and consent need proper legal review.
