Third-party cookies have allowed advertisers to track users across websites and attribute conversions to specific ads. As browsers and platforms remove this tracking, the data available for attribution is shrinking, and existing models are becoming less accurate.
Google delayed the full deprecation of third-party cookies in Chrome multiple times, but the direction of travel is clear. Safari and Firefox already block third-party cookies by default. iOS privacy changes have already significantly reduced cross-app tracking. Attribution as it worked in 2018 is no longer possible in most environments.
What third-party cookies were doing for attribution
Third-party cookies allowed ad networks to recognise the same user across different websites. When someone saw an ad on one site and converted on another, the cookie linked those two events. This was the foundation of cross-site display and retargeting attribution.
Without third-party cookies, this cross-site user matching is broken for users who have not consented to tracking or who are in environments that block cookies. The same user appears as two anonymous visitors: one who saw the ad and one who converted.
What still works after cookies
- First-party cookie attribution: tracking a user's journey within your own website during a single session or across sessions on your domain still works
- Logged-in user tracking: if users are logged into your platform, you can match their ad exposure to their purchase using your own first-party ID
- Google's Privacy Sandbox technologies: designed to enable some level of interest-based advertising and attribution without cross-site tracking
- Server-side tagging: moving tracking from the browser to your server reduces dependency on browser cookies and improves data reliability
- Hashed email matching: sending hashed email addresses to platforms like Meta and Google allows matching with users who are logged into those platforms
The modelling gap
When tracking data is missing, platforms fill in the gaps with modelled data. Meta's Conversions API attempts to recover lost signal by matching server-side conversion events to users. Google models conversion behaviour for users where direct tracking is unavailable.
Modelled data is better than no data, but it introduces assumptions. The accuracy of modelled attribution depends on how well the platform can infer missing conversions from available signals. As tracking gaps widen, modelling becomes less reliable.
In a world with less user-level tracking data, the relative value of marketing mix modelling and incrementality testing increases. These methods do not rely on individual user tracking: they work from aggregate business outcomes, which remain measurable regardless of cookie availability.
Building a cookieless measurement stack
A robust measurement stack for a cookieless world combines several approaches. First-party data becomes more valuable: collect email addresses, build logged-in user journeys, and use customer match to reconnect first-party data with platform targeting.
Marketing mix modelling provides channel-level budget guidance based on aggregate revenue data rather than user-level tracking. Incrementality testing provides experimental validation that does not depend on cookies or cross-site tracking at all.
What to do now
Audit your current attribution setup and identify where it depends on third-party cookies or cross-site tracking. Implement server-side event tracking using Meta's Conversions API and Google's Enhanced Conversions where possible. Invest in first-party data collection. And start treating aggregate measurement methods, MMM and incrementality, as your primary measurement framework rather than a secondary check.
Are third-party cookies already gone from Chrome?
As of 2025, Chrome still supports third-party cookies but has given users the option to limit tracking, and Google has introduced user choice mechanisms. Full deprecation has been delayed repeatedly. Regardless of Chrome's timeline, Safari and Firefox already block third-party cookies, meaning a large portion of your audience is already in a cookieless state.
Will Privacy Sandbox replace third-party cookies effectively?
Privacy Sandbox technologies, including the Topics API and Attribution Reporting API, are designed to enable interest-based advertising and conversion measurement without individual cross-site tracking. They provide aggregate and probabilistic data rather than user-level records. Most advertisers and measurement providers are still evaluating how well these technologies meet their needs in practice.
Should I invest in a CDP to solve the cookieless problem?
A customer data platform (CDP) can help you consolidate and activate your first-party data, which becomes more valuable as third-party tracking diminishes. But a CDP is a data management tool, not a measurement solution. It helps you target known customers more effectively but does not solve the attribution challenge for users you cannot identify.