
A first party data strategy for publishers is a plan for collecting data readers give you on your own site, turning it into audience segments, and passing those segments to ad buyers so they bid more. It covers consent, collection, segmentation, activation through Ad Manager, Prebid or deals, and testing whether the extra bids are real.
Here's the problem. A growing share of your visitors reach the auction with no third-party cookie at all, so buyers treat them as strangers and bid low. Safari has blocked third-party cookies by default since 2020. Most guides answer this with "collect more data", then stop before the part that earns money: getting that data into the bid request. This guide covers the whole path, and it pairs well with our explainer on how third-party cookies work today.
What is a first party data strategy?
A first party data strategy is the set of decisions about what reader data you collect, why readers would share it, how you group it, and where you send it. Without the last step, it's just a database.
First-party data is anything you learn directly from your own readers on your own site or app. Page views, sections read, newsletter sign-ups, logins and survey answers all count. You own the relationship, so you decide how it's used, within consent.
First party vs third party data, and the types in between
The difference between first party vs third party data is who collected it and where. Your data comes from your site. Third-party data is bought or synced from companies that tracked people elsewhere.
| Type | Who collects it | Publisher example | How buyers see it |
|---|---|---|---|
| Zero party data | You, because the reader told you on purpose | Topics picked in a newsletter preference centre | Strong intent, usually small lists |
| First-party data | You, from behaviour on your site | Reads three recipe pages a week | Trusted when it's consented and fresh |
| Second-party data | A partner, shared with you directly | A sister site's audience in a joint deal | Depends on the partner |
| Third-party data | Data companies, across many sites | "In-market for cars" from a cookie pool | Common but fading, and patchy on Safari |
Zero party data is really a subset of first-party data. The label just stresses that the reader chose to share it, which makes it easier to defend under privacy law.
Why first party data for publishers still matters in 2026
First party data for publishers matters because buyers pay more for impressions they can understand, and fewer impressions now arrive with a cookie that explains them. That's true even though Chrome kept third-party cookies.
Let's get the facts straight, because page one gets them wrong. In October 2025, Google said Chrome will keep offering users third-party cookie choice and retired most Privacy Sandbox APIs, including Topics and Protected Audience. So Chrome isn't "killing cookies". Safari is a different story. WebKit announced in March 2020 that cookies for cross-site resources are blocked by default.
What does that mean for you? Open your analytics and check the browser split. If a third of your traffic is Safari on iPhone, a third of your impressions already sell without cookie data. Add readers who decline consent in the UK and EU, and the gap grows. Your own data is the main signal you can add to those impressions.
Our view. Don't build a first-party data plan out of fear of a cookie deadline that isn't coming. Build it because Safari and no-consent traffic is already selling cheap, and that's money you can measure.
How to build a first party data strategy for publishers in 6 steps
Build your first party data strategy for publishers in this order: consent, audit, collection, segments, activation, testing. Skipping ahead is the most common reason the data never earns anything.
- Fix consent before you collect anything new.
- Audit the data you already have.
- Collect more with a clear reason to share.
- Group readers into a few segments buyers understand.
- Send the segments to buyers through the channels your setup supports.
- Test whether the segments raise revenue, and keep only what pays.
1. Fix consent first
Consent comes first because data collected without it can't be used safely, and buyers know that. In the UK, the ICO's cookie guidance says consent "must be actively and clearly given". It also says you can't set non-essential cookies before the user has consented. Prebid's docs add a warning that sending first-party user data may need special consent in some regions.
When we review a site, we load it in a private window and watch the network tab. If ad or analytics tags fire before the consent banner is answered, that gets fixed before anything else. Your consent platform should block those tags until the reader chooses.
2. Audit what you already collect
Most publishers already hold more first-party data than they think. List every place readers leave a trace:
- Analytics: sections, topics, visit frequency, device and country.
- Your CMS: the category and tags on every article.
- Email: newsletter lists and the topics people signed up for.
- Accounts: logins, comments, saved articles or paid members.
- Forms: contests, surveys, calculators and quizzes.
Note who owns each source and whether it was collected with consent. Data you can't legally use goes in the bin, not in a segment.
3. Collect more with a clear reason to share
Readers share data when they get something back. That trade is the core of how to collect first party data without annoying people. Good swaps for a content site include:
- A newsletter split by topic, so the sign-up itself tells you what they like.
- A free account that saves recipes, reading lists or calculator results.
- A short preference centre: "Which of these topics do you want more of?"
- Progressive profiling: ask one extra question per visit, not ten on day one.
- Gated extras, such as a downloadable checklist, while normal articles stay open.
Watch out for login walls on your main articles. They collect emails, but they also cut page views, and page views are what you sell. Test a wall on one section before you roll it out.
4. Build a few segments buyers understand
A segment is a group of readers who share a trait, such as "read three or more travel articles in 30 days". Start with three to five segments tied to your biggest content areas. Buyers can't use 200 tiny segments, and small segments may never reach serving thresholds.
Use standard labels where you can. IAB Tech Lab's Audience Taxonomy gives every common audience a shared ID, and version 1.1 is the one Prebid's docs point to. A buyer's DSP reads "IAB audience 1.1, segment X" without needing your private naming.
5. Send segments to buyers
Activation is the step that turns data into bids. The right channel depends on what you run: Ad Manager 360, plain Ad Manager, AdSense, or Prebid. The next section compares them.
6. Test and keep only what pays
Run a split test before you call it a win. Send segments for half your traffic and none for the other half, for two to four weeks, then compare revenue per thousand sessions on the same pages. Check Safari separately, because that's where the data should matter most. If there's no lift, change the segments or the channel, not the dates on a report.
Where to send first-party data so buyers can use it
First-party data reaches buyers through six main routes. Each has limits, and several are Ad Manager 360 only, so pick by your setup, not by the vendor pitch.
| Channel | What it does | Who can use it | Limits to know |
|---|---|---|---|
| Ad Manager first-party audience segments | Groups visitors by behaviour for targeting your own line items and deals | Ad Manager 360 | Ads serve to a segment after 100 members, and up to 150 for smaller lists; you set membership expiry in days |
| Publisher provided identifier (PPID) | Your own hashed user ID, used for frequency capping, segments and targeting | Ad Manager 360 | 22 to 150 characters; must be hashed or encrypted, never raw personal data; for non-Google demand only outside the EEA, Switzerland, UK and US states with comprehensive privacy laws |
| Publisher provided signals (PPS) | Passes your audience and contextual data to programmatic buyers in standard segments | Ad Manager 360 (Beta, on by default) | Uses standard industry segments such as the IAB audience taxonomy |
| Secure signals | Shares obfuscated signals with bidders through Authorized Buyers, Open Bidding and SDK Bidding | Ad Manager publishers | Signals must be hashed or encrypted and are shared only on your explicit instruction |
| Prebid first party data | Adds site and user data to every header bidding request | Any Prebid.js setup | User data may need consent; segments go in ortb2.user.data with a segtax ID |
| Curated Audiences and direct deals | Packages your segments into deals buyers target by deal ID | Any publisher with buyers or an SSP that supports it | Needs enough volume per segment to be worth a deal |
Each row is from the official page: Google's help on first-party audience segments, publisher provided identifiers, publisher provided signals and secure signals.
Prebid first party data in practice
If you run header bidding, Prebid is the fastest route because every bidder reads the same fields. Prebid's first party data docs say standard OpenRTB fields go in ortb2.site or ortb2.user. Your own extra fields go in ortb2.site.ext.data, and audience segments go in ortb2.user.data with a segtax value. Segtax 4 means IAB Audience Taxonomy 1.1.
pbjs.setConfig({
ortb2: {
site: {
ext: { data: { section: "travel", pagetype: "article" } }
},
user: {
data: [{
name: "example.com",
ext: { segtax: 4 }, // IAB Audience Taxonomy 1.1
segment: [{ id: "TAXONOMY_ID" }] // IDs from the taxonomy file
}]
}
}
});
Want to send data to only some bidders? Use pbjs.setBidderConfig() with the bidder list. Then open your browser's network tab, find a bid request (our guide to what bidstream data reveals walks through each block) and confirm the segments are really in it. We find setups where the config exists but runs after the auction starts, so no bidder ever sees it. Our Prebid.js setup guide covers where this code sits in the page.
Curated Audiences and deals
Curated Audiences is IAB Tech Lab's spec for sellers to label audiences with standard taxonomy IDs. It was called Seller Defined Audiences until IAB Tech Lab renamed it in December 2024. The idea is simple. You define the segment, you keep the raw data, and buyers target the label. Pair it with a private marketplace deal when one segment has enough volume for a buyer to care.
First party data monetization: where the money comes from
In a first party data strategy for publishers, first party data monetization comes from three places: higher open-auction bids on impressions you can describe, deals built on your segments, and reader revenue such as sponsored newsletters. Most sites see the first one before the others.
- Open auction. Segments and context in the bid request give DSPs a reason to bid on Safari and no-consent traffic. Requests buyers bid on also keep getting sent to them, which is how SSP traffic shaping works. This works at any size.
- Deals. A buyer who wants "UK parents reading about school costs" can target your segment through a deal ID. That needs steady volume.
- Your own products. Topic newsletters can carry sponsorships sold on audience size and interest, with no ad tech in between.
The size of the lift varies by niche, country and buyer, so your own split test is the number to trust. If you want help reading those numbers, our reporting and yield service does exactly that.
Context helps too. When a reader has no ID and no consent, the page itself is still a signal. Our guide to contextual advertising shows how to send page topics alongside your audience data.
First-party cookies, IDs and where they fit
First-party cookies are cookies set by your own domain, and they're how most first-party data gets linked to a returning reader. Safari's default blocking applies to cross-site cookies, not to cookies your own domain sets.
A first-party cookie can hold your own hashed user ID. That's what you pass as a PPID in Ad Manager 360. Shared IDs work the same way but are readable by many bidders at once. If you're weighing those, read our guide to universal IDs for publishers before adding vendor scripts. More IDs mean more page weight, and not every ID pays its way.
Should you buy a data platform? Usually not at first. Ad Manager segments, PPS and Prebid cover most needs for a single-site publisher. Our DMP for publishers guide explains when the extra tool is worth it.
Mistakes we see in publisher first-party data setups
The same handful of mistakes come up in almost every review. Check yours against this list:
- Tags fire before consent. Everything built on that data is at risk.
- Too many segments. Fifty segments of 40 people each never reach a serving threshold.
- Raw emails in a PPID. Google requires the ID to be hashed or encrypted and never raw personal data.
- Config that never reaches the bid. Prebid data set after
requestBids()is too late. - Old segments. A reader who read about mortgages a year ago isn't in market today. Set an expiry.
- No test. Without a split, you can't tell data lift from a seasonal Q4 bump.
Fix those six and you're ahead of most sites your size. None of them needs a new vendor.
Frequently asked questions
Do I still need first-party data if Chrome keeps third-party cookies?
Yes. Chrome kept cookie choice in October 2025, but Safari blocks third-party cookies by default and many readers decline consent. Your own data is the main signal for those impressions.
What's the difference between first-party cookies and third-party cookies?
A first-party cookie is set by the site you're on, such as your own domain. A third-party cookie is set by another domain, often an ad-tech company, and Safari blocks those by default.
Can I use first-party data without consent?
Not for ads in the UK and EU. The ICO says you can't set non-essential cookies before a user consents. Prebid also warns that first-party user data may need special consent in some regions.
Do I need Ad Manager 360 for a first party data strategy?
No. Audience segments, PPID and publisher provided signals are 360 features, but Prebid first party data, contextual signals and direct deals work on any setup.
How many readers do I need before segments are useful?
In Ad Manager 360, ads serve to a first-party segment once it has 100 members, and small lists may need up to 150. For deals, buyers usually want far more, so start with your largest topics.
What is zero party data?
Zero party data is information a reader deliberately gives you, such as topics picked in a newsletter preference centre. It's part of your first-party data, with clearer intent behind it.
Next step
A first party data strategy for publishers only pays once the data is in the bid request and tested against a control. Start with consent and three to five segments, then send them through Prebid or Ad Manager. If you'd like us to check your header bidding setup and data flow, see our header bidding service or ask for a free site review.
More guides in this topic
- Programmatic Advertising for Publishers in Plain English
- How an Ad Auction Works for Website Publishers
- What Is a Supply Side Platform? A Publisher Guide
- DSP vs SSP Explained From the Publisher's Side
- Programmatic Display Advertising Explained for Publishers
- Prebid.js Setup Guide for Publishers, Step by Step
- Programmatic Media Buying Explained for Publishers
- DMP for Publishers and When You Actually Need One
- Contextual Advertising Explained for Website Publishers
- What Are Third Party Cookies? A Guide for Publishers
- Walled Garden Advertising Explained for Publishers
- Real Time Bidding Explained for Website Publishers
- Ad Network vs Ad Exchange for Publishers Explained
- What Is Bid Shading? A Plain Guide for Publishers
- Privacy Sandbox for Publishers After Google Retired It
- What Is a Universal ID? A Guide for Publishers
- What Is Cookie Syncing? A Plain Guide for Publishers
- Bidstream Data Explained for Publishers (2026 Guide)
- Traffic Shaping in Programmatic Advertising for Publishers


