Partners Login
Contact UsStart here: the deadline disappeared, the problem did not
Most articles on this subject were written for a deadline that no longer exists. Google abandoned its plan to deprecate third-party cookies in Chrome in July 2024, confirmed in April 2025 that it would not show users a choice prompt either, and on 17 October 2025 retired most of the Privacy Sandbox — Topics, Protected Audience, Attribution Reporting and others. Only CHIPS, FedCM and Private State Tokens survive.
That changes the question. It is no longer “what replaces the cookie before Chrome switches it off?” It is “how do I reach and measure the audiences that were never addressable by cookies in the first place?” — Safari, Firefox, mobile apps and connected TV, which together account for more than 40% of web and mobile traffic.
Here is what is actually available, what each option genuinely covers, and where each one stops.
1. First-party data
What it is: data you collect directly — sign-ups, purchases, logged-in behaviour, newsletter subscribers.
Where it works: everywhere you own the relationship. It is the strongest asset most advertisers and publishers have, and the one least exposed to platform decisions.
Where it stops: it only describes people who already identified themselves to you. It cannot help you reach prospects who have not, and most publishers see a single-digit percentage of traffic authenticate. First-party data is a foundation, not a substitute for addressability at scale.
2. Universal and alternative IDs
What it is: shared identifiers, usually built from a hashed email or a login, that multiple ad tech vendors can read. UID2 and similar initiatives are the best-known examples.
Where it works: logged-in environments — retail media, subscription publishers, large authenticated platforms. Where authentication is high, coverage is genuinely good.
Where it stops: at the login wall. On open-web inventory, where the majority of users never sign in, match rates fall away sharply. These IDs also inherit the cookie’s original shape, which means they do not solve in-app or CTV addressability. A publisher relying on them alone will still have a large unaddressable remainder.
3. Contextual advertising
What it is: targeting the content of the page rather than the history of the person reading it.
Where it works: it is privacy-safe by construction, needs no identifier, and has improved considerably with better content classification. For upper-funnel brand work it is a legitimate answer.
Where it stops: it cannot retarget, cap frequency across sites, or attribute a conversion to an exposure. It tells you about the page, never about the person, so performance campaigns that depend on sequencing or measurement cannot be run on it.
4. Google Privacy Sandbox
What it is: Google’s browser-level replacement programme, built between 2020 and 2025.
Current status: largely retired. Google shut down most of the advertising APIs on 17 October 2025, and Chrome has continued removing what remained through 2026. CHIPS, FedCM and Private State Tokens survive, but they are not an audience addressability solution.
What to do about it: if your cookieless plan still has a Privacy Sandbox workstream in it, that plan needs revisiting. This is the single most common out-of-date assumption we see.
5. Data clean rooms
What it is: a controlled environment where two parties match data without either seeing the other’s raw records.
Where it works: measurement, overlap analysis and planning between large partners who both hold substantial first-party data.
Where it stops: it is an analysis and collaboration layer, not a way to make an anonymous bid request addressable in real time. Clean rooms answer “what happened?” across two datasets; they do not answer “who is this impression?” during an auction.
6. Identity resolution
What it is: rather than retrieving a stored identifier, identity resolution infers who a user is by clustering signals into an identity graph, then exposes that resolved identity to buyers in real time.
Where it works: this is the only approach on this list that addresses all four problem environments at once — cookieless browsers, in-app and CTV — without requiring the user to log in. Intent IQ’s graph is built from over 25 billion digital signals processed daily and more than 15 billion email open and log-in events a month, mapped across four connected layers: household, person, device and profile. See how the IIQ ID works.
What to check before buying: accuracy and freshness are the two questions that separate vendors. Intent IQ matches at 92%–97% accuracy, measured deterministically and verified daily, and no entity in the graph exists for longer than 48 hours before being refreshed. A graph that is accurate but stale produces confident mistakes.
Watch: How Trusted Identity Drives Revenue for Publishers & Results for Advertisers
Why identity accuracy shows up differently on the buy side and the sell side — and what each gains when it improves.
Which to choose, by role
If you are a publisher
Your problem is inventory that reaches the auction unidentified, attracts fewer bidders, and clears below its value. First-party data helps where users authenticate; contextual helps for brand budgets; identity resolution addresses the remainder, which is usually the largest part.
Delivered as Prebid EIDs inside the bid request, resolved identity restores competition in the auction. Published results: Publift recorded a 20.7% revenue uplift in the US and up to 70% in top deployments. Fandom saw a 55–140% increase in revenue from cookieless users, with 26% higher CPMs and 58% higher fill rates. Media Square doubled RPM on ID-less inventory in France while remaining fully GDPR compliant. Daily Mail reported higher bid rates, fill, CPMs and win rates across open exchange and PMP.
More detail: Bid Enhancement.
If you are an advertiser or agency
Your problem is an audience you cannot see. Fandom found cookieless traffic contributing under 20% of attributable programmatic revenue, even though 79% of that audience was on iOS — where US users carry an average household income around 40% higher than Android users. The audience was valuable and present; the measurement simply could not find it.
Identity resolution lets you activate your existing first-party, third-party and onboarded data against those users through your current SSP and DSP partners, and attribute across CTV, Safari, Firefox and Edge. In a controlled mobile test, BidMachine measured a 35%+ performance increase on Android and 85%+ on iOS against requests using MAIDs alone.
More detail: Cookieless Audience Activation.
The honest summary
There is no single replacement for the third-party cookie, and anyone selling one is overstating their case. Most organisations end up running two or three of these together: first-party data as the foundation, contextual for upper-funnel reach, and identity resolution to make the unauthenticated majority addressable and measurable.
What has changed since 2024 is that waiting is no longer a strategy. The platform-led replacement was cancelled. Whatever share of your traffic was unaddressable then is still unaddressable now, and nothing is arriving to fix it.
For the wider picture, read our guide to cookieless advertising.
FAQs
What is the best alternative to third-party cookies?
There is no single replacement. Most advertisers and publishers combine first-party data, contextual targeting and identity resolution. Identity resolution is the only one of the three that restores addressability across cookieless browsers, mobile apps and CTV without requiring users to log in, which is why it usually carries the largest share of the gap.
Is Google Privacy Sandbox still an option?
Largely no. Google retired most Privacy Sandbox APIs — including Topics, Protected Audience and Attribution Reporting — on 17 October 2025, and Chrome has continued removing the remainder through 2026. Only CHIPS, FedCM and Private State Tokens survive, and none of them is an audience addressability solution.
Do universal IDs like UID2 solve the problem?
Only where users log in. Identifiers built on hashed emails or authentication work well in retail media and subscription environments, but match rates fall sharply on open-web inventory where most users never sign in. They also do not address in-app or CTV, which never used cookies.
Can contextual advertising replace cookies?
For upper-funnel brand campaigns, often yes. But contextual cannot retarget, cap frequency across sites, or attribute a conversion to an exposure, because it describes the page rather than the person. It is a complement to identity, not a substitute.
Which alternative works on connected TV?
Only identity resolution. CTV has no cookies and no shared identifier, so the constraint is attribution rather than delivery. An identity graph that connects household, person, device and profile layers allows a CTV exposure to be tied to an outcome on another screen.
Are these alternatives privacy compliant?
They can be, and compliance should be your first question of any vendor. Intent IQ resolves identity without personally identifiable information, and deployments are designed to support GDPR, CCPA and GPP. The Media Square validation in France was conducted in full GDPR compliance.
How do I evaluate an identity vendor?
Ask how accuracy is measured and how often the graph is refreshed, which environments are covered, and whether coverage depends on logins. Our five considerations for evaluating an identity resolution tool sets out the full checklist.