Key takeaways
- First-party data is what your business collects directly from its own customers and visitors: site and app events, orders, signups, CRM records and survey answers. It is the only data whose source, consent and definitions you control.
- The measurement problem is browser storage, not a cookie apocalypse. Safari blocks all third-party cookies and deletes cookies created in JavaScript after 7 days without user interaction, while Chrome kept users' third-party cookie choice in October 2025.
- The most useful first-party records for measurement are the boring ones: an order table that stores the campaign of the first landing, a user ID for logged-in visitors, lead stages from the CRM and a 'how did you hear about us' answer.
- Collect it with consent, keep only what a decision needs, and when you share it with ad platforms send it hashed, as Google Ads enhanced conversions does with SHA256.
- Write the source into the order record at the moment it happens. A channel report rebuilt later from cookies cannot recover what the browser already deleted.
First-party data is the information your business collects directly from its own visitors and customers: what they do on your site and app, what they buy, what they tell you when they sign up, and what your sales and support teams record. For measurement it matters more every year, because it is the one data source whose origin, consent and definitions you control, and the one that browser privacy settings cannot delete after the fact. This guide covers how first-party data differs from second- and third-party data, which records actually improve marketing measurement, and how to collect them with consent.
First-party data is information an organization collects directly from its own audience through channels it owns, such as its website, app, checkout, CRM, support desk and surveys.
What is first-party data, and how is it different from second- and third-party data?
First-party data is collected by you, from people who dealt with you. Second-party data is someone else’s first-party data that they share with you directly, under an agreement, such as a marketplace sharing order data with its sellers. Third-party data is assembled by a company that never had a direct relationship with the people in it, usually by aggregating many sources, and it is the hardest to trace back to a source and a consent.
| First-party | Second-party | Third-party | |
|---|---|---|---|
| Who collects it | You | A partner, from its own customers | A data company, from many sources |
| Typical examples | Orders, site events, signups, CRM stages, survey answers | A marketplace’s sales report for your products, a co-marketing partner’s list | Bought audience segments, interest and demographic profiles |
| Do you know how it was collected? | Yes | Only what the partner tells you | Rarely |
| Do you control the definitions? | Yes | No | No |
| Use for measurement | The base layer | A useful supplement for a channel you do not own | Targeting at best, not proof of results |
The distinction is about the relationship, not the technology. A cookie set by your own domain is first-party; the same kind of cookie set by an ad network’s domain while a visitor reads your page is third-party. An email address typed into your checkout is first-party data even when you later upload it, hashed, to an ad platform.
Why does first-party data matter more for measurement now?
Because browsers keep shortening how long any script can recognise a visitor, while your own records stay put. The common story, that “cookies are dying”, is half right and causes bad decisions, so it helps to be precise about what changed.
- Safari blocks third-party cookies outright. WebKit’s documentation states that Intelligent Tracking Prevention “by default blocks all third-party cookies” and that “there are no exceptions to this blocking” (WebKit).
- Safari also limits your own script-set cookies. The same page says ITP deletes all cookies created in JavaScript, plus localStorage, IndexedDB and other script-writeable storage, after 7 days of no user interaction. When Safari classifies the referring site as a tracker that decorates links, cookies created in JavaScript on the landing page are capped at 24 hours (WebKit).
- Disguising a third-party server as your subdomain does not escape it. Since November 2020, ITP caps cookies set in third-party CNAME-cloaked responses to 7 days, the same as cookies created in JavaScript (WebKit).
- Chrome did not drop third-party cookies. In October 2025 Google wrote that Chrome “will maintain our current approach to offering users third-party cookie choice”, and announced it would retire several Privacy Sandbox technologies, including the Attribution Reporting API, Protected Audience and Topics (Privacy Sandbox).
Put together: a Safari visitor who returns after eight days often looks like a new user to any tag that relies on a JavaScript cookie, cross-site identifiers are gone in Safari, and in Chrome they depend on the user’s choice. Add consent banners and ad blockers and the gap grows further. What stays complete is the order in your database, the account in your user table and the lead in your CRM. Those records are the anchor you reconcile every tag-based report against, as the conversion tracking guide does with ad platforms.
Which first-party data should you collect for marketing measurement?
Start with the records that answer “which spend produced which revenue”, not with everything you could store. Five types carry most of the value:
- Orders with their source attached. When an order is created, write into the order row the UTM source, medium and campaign of the visit that started the session, plus a stable transaction ID. A field saved at the moment of purchase survives even if the browser deletes every cookie the next day. Consistent tagging is what makes this field usable, which the UTM parameters guide covers.
- A user ID for logged-in visitors. Once someone signs in, your own account ID lets you join visits from a phone and a laptop into one person, which no cookie can do across devices. The same ID is what lets you add a customer’s repeat orders to their first one when you calculate cohort-based customer lifetime value.
- Events from your own site. Product views, add-to-cart, form submissions and signups, named and parameterised to a plan. The event tracking guide shows how to design one so the reports stay clean.
- CRM stages for leads. For a lead-gen business the form fill is not the money. Qualified and won stages, sent back with the original lead’s source, are what let you compare channels by cost per customer instead of cost per form.
- Self-reported attribution. One optional question at checkout or signup, “How did you hear about us?”, captures channels a tag never sees: a friend’s recommendation, a podcast, a Telegram channel post read on another device.
What to skip: fields you cannot name a decision for. Every extra personal field increases what you have to protect and explain, and adds nothing to a channel report.
How do you collect first-party data with consent?
Collect it in a way you could explain to the customer on the same screen. In practice that comes down to a short sequence:
- Ask before non-essential tracking runs. A consent banner decides which analytics and marketing tags may load. Size what it removes from your reports before you trust any trend; the guide to cookie consent and analytics shows how to measure your consent rate and correct for it per channel. If the banner and the tracker come from one tool, the choice is enforced and stamped on every record; AdPix cookie consent works that way and keeps an exportable consent log as proof.
- Store the source server-side at the moment it matters. Read UTM parameters on landing, keep them for the session, and write them to the order or lead record when it is created. This does not need a cookie that outlives the visit.
- Send purchases from your server, not only the browser. A server-side purchase event is not lost to an ad blocker or a closed tab, but only one path should report each order, or revenue doubles. The server-side tracking guide covers which events to move first and what a tagging server adds on top.
- Hash personal fields before sharing them. When you send customer data to an ad platform, send it hashed. Google Ads enhanced conversions, for example, uses SHA256 to hash first-party customer data such as email addresses before it is sent, to improve the accuracy of conversion measurement (Google Ads Help).
- Write down who can see what. A short internal note on which fields exist, why, how long you keep them and who has access is the minimum. Data protection law differs by country, so check the rules that apply to your customers; this guide is not legal advice.
Worked example: rebuilding a channel report from your own orders
The numbers here are hypothetical. Say an online shop records 1,200 paid orders in a month in its database. The analytics tool attributes revenue for 840 of them to a channel. The other 360 show as direct or unattributed.
The team makes two first-party changes. It saves the session’s UTM values into each order row at checkout, and it adds an optional “How did you hear about us?” question to the thank-you page. A month later, with 1,200 orders again:
| Channel | Orders the tag attributed | Orders from the order table (UTM saved at checkout) | Plus survey answers on untagged orders | Ad spend | Cost per order (tag) | Cost per order (order table + survey) |
|---|---|---|---|---|---|---|
| Google search ads | 310 | 360 | 375 | $9,000 | $29.03 | $24.00 |
| Instagram ads | 180 | 260 | 300 | $9,000 | $50.00 | $30.00 |
| 150 | 170 | 175 | $500 | $3.33 | $2.86 | |
| Organic and other | 200 | 230 | 270 | — | — | — |
| Unattributed | 360 | 180 | 80 | — | — | — |
The arithmetic for Instagram: $9,000 ÷ 180 = $50.00 per order on the tag’s count, and $9,000 ÷ 300 = $30.00 once the order table and the survey are counted. The tag was not wrong about the orders it saw; it missed buyers whose cookies were gone by the time they paid, and buyers who saw a post on their phone and bought on a laptop. A team reading only the tag-based report would have judged Instagram two-thirds more expensive per order than it was ($50.00 against $30.00) and might have cut it. For Instagram specifically, the Instagram ads tracking guide shows how to join tagged sessions, per-creator codes and the self-reported source into one order record, de-duplicated by order ID.
Two cautions keep this honest. Survey answers are self-reported and over-credit memorable channels, so use them to place unattributed orders, not to overwrite tagged ones. And neither column proves the ads caused the sales; for that you need a holdout, as the incrementality testing guide explains.
How do you use first-party data with ad platforms?
Use it in two directions, carefully. Outbound, you send hashed conversion data back so the platform measures and bids on real orders rather than only on what its pixel saw; enhanced conversions in Google Ads is one documented example (Google Ads Help). Inbound, you compare the platform’s claimed conversions with your order table and compute how much each one over- or under-claims, before you trust its cost per acquisition.
Keep one rule: the order table is the truth, and every platform is a view of it. If a platform reports 400 conversions for a month in which your database shows 300 orders from that channel, the ratio of 1.33 is a number to track, not a mystery to argue about. The attribution models guide explains why the two counts differ even when both systems work correctly.
Where does AdPix fit for first-party data?
AdPix is a web analytics platform that replaces Google Analytics 4: realtime, acquisition, engagement, conversion and e-commerce reports, plus attribution, lift testing and marketing mix modelling, with no data sampling and a free plan. For first-party data the relevant parts are where the data lives and who owns it. AdPix collects events with your site’s own tag, shows a user’s visits across devices and channels as one journey, hosts the data on infrastructure inside Iran and lets you export it at any time. A cookie consent banner is included on every plan, including Free, and API access for pulling the data into your own systems starts on Starter. If a rule says analytics data must sit on servers you run yourselves, the self-hosted route is Matomo On-Premise, and the Matomo vs GA4 comparison shows the server size and upkeep that ownership costs.
AdPix runs alongside Google Analytics without interfering, so you can compare both against your order table before deciding anything. If you are still on GA4 alone, the guide to exporting GA4 data shows how to keep a copy of your history, and the website tracking guide puts first-party data into a full tracking plan. The practical first step costs nothing: add a source column to your order table this week, and compare it with your analytics channel report at the end of the month.
UTM link builder
Tag campaign links so every click is credited to the right channel. The free plan covers 100,000 page views a month for each website, no credit card. Setup takes under five minutes.
Frequently asked questions
What is first-party data?
First-party data is information a business collects directly from its own audience through channels it owns, such as website and app events, purchases, account signups, CRM records, support conversations and surveys. Because the business collected it itself, it knows the source, the consent given and how every field is defined.
What are examples of first-party data?
Common examples are orders in your shop database, events your analytics tag records on your site, email addresses from newsletter signups, logged-in user IDs, lead stages in your CRM, answers to a post-purchase survey and phone or chat orders your team logs by hand.
What is the difference between first-, second- and third-party data?
First-party data you collect yourself from your own customers. Second-party data is another company's first-party data that it shares with you directly under an agreement. Third-party data is assembled by a company with no direct relationship to the people in it, usually by aggregating many sources, and is the least transparent about origin and consent.
Are first-party cookies affected by Safari's tracking prevention?
Yes, partly. WebKit's documentation says Safari's Intelligent Tracking Prevention deletes all cookies created in JavaScript, and other script-writeable storage, after 7 days without user interaction with the site, and caps cookies set in CNAME-cloaked third-party responses to 7 days. Cookies are first-party by domain, but a returning Safari visitor can still look new after a week.
Do I need first-party data if Chrome is keeping third-party cookies?
Yes. Google said in October 2025 that Chrome will keep its current approach of offering users a third-party cookie choice, but Safari already blocks third-party cookies by default and limits script-set cookies, and ad blockers and consent choices remove more. Your own order and customer records are the only measurement layer none of these settings can erase.
Sources and methodology
Browser behaviour (Safari's third-party cookie blocking, the 7-day limit on cookies created in JavaScript and on CNAME-cloaked cookies) comes from WebKit's documentation and blog in the sources; Chrome's position on third-party cookies and the retired Privacy Sandbox APIs come from Google's Privacy Sandbox post of 17 October 2025; enhanced conversions come from Google Ads Help. All were opened on 15 November 2026. AdPix capabilities and plan placement come from adpix.io/pricing (read on 13 November 2026). The store in the worked example and all its figures are hypothetical.
- WebKit: Tracking Prevention in WebKit
- WebKit: CNAME Cloaking and Bounce Tracking Defense (12 November 2020)
- Privacy Sandbox: Update on Plans for Privacy Sandbox Technologies (17 October 2025)
- Google Ads Help: About enhanced conversions
This article was drafted with AI assistance and reviewed by the AdPix team before publishing. Editorial policy