Challenge · Data
The cookie is gone and nothing single replaced it
Everyone waited for one replacement. Several partial ones arrived instead, which turns out to be workable, as long as you own a base of customer data to plug into them.
20 August 2026 · 6 minute read
There is no single successor to the third-party cookie. What exists is a set of overlapping industry identifiers, chiefly two called UID2 and RampID, both built on hashed email addresses collected with consent, alongside platform-owned graphs and contextual targeting. These are described as a patchwork, usually as a criticism, but a patchwork is a reasonable outcome: each identifier covers a different slice, and coverage overlaps. For a consumer brand the practical priority is not choosing an identifier. It is building the first-party base that feeds all of them, because every one of these approaches works better the more of your own customer data you can bring.
What actually replaced the cookie?
| Approach | Built on | Where it works |
|---|---|---|
| UID2 | Hashed email, with consent | Open web, CTV, increasingly retail media |
| RampID | Resolved identity graph | Where a data partnership already exists |
| Platform graphs | Logged-in users | Inside Meta, Google, LinkedIn |
| Contextual | The page, not the person | Everywhere, with no identity at all |
| Your own first-party data | Your customers | Everywhere, and it feeds all of the above |
The last row is the one that matters most and gets discussed least, because it is unglamorous. Every approach above works better with more first-party data, which makes it the highest-return investment regardless of which identifier wins.
Why is a patchwork acceptable?
Because the cookie was never universal either. It failed on mobile apps, in browsers that blocked it, and in logged-in environments. We remember it as complete because we built our measurement around what it could see and stopped counting what it could not.
Several overlapping identifiers with partial coverage each is a more honest version of the same situation. The work is accepting that measurement is estimation, and building for that rather than pretending to a precision that never existed.
The practical consequence
Stop optimising for perfect attribution. Build the first-party base, accept modelled measurement for the gaps, and spend the saved effort on the parts of marketing that were never measurable and always mattered.
What should a consumer brand actually do?
- Collect email consent properly. Not a dark pattern, a genuine value exchange. This is the substrate for UID2 and for everything else.
- Get your customer data into one place. Orders, sessions, engagement. Most brands have this scattered across three systems that do not join.
- Build audiences from your own base first, then extend with platform tools. Starting from a platform lookalike wastes the advantage you already have.
- Adopt UID2 where your partners support it, without expecting it to be complete.
- Keep contextual in the mix. It needs no identity at all, and it has quietly improved while everyone argued about identifiers.
Where does retail media fit?
It is the fastest-growing exception, because retail media networks have logged-in purchase data and can attribute to an actual transaction. A $196B market with real purchase attribution is a meaningful alternative to identity-based targeting on the open web.
For a consumer brand selling through retailers, this is worth more attention than most identity discussions.
Every brand asking which identifier to standardise on is asking a question that will be answered for them by their partners. The question they control is how much first-party data they have, and that one has the same answer whichever way the rest goes.
If you are an agency
The identity conversation is one where clients expect certainty and the honest answer is a range. Lead with the first-party recommendation, because it is the part that is definitely right, and treat identifier selection as a question of which partners the client works with. Agencies that oversold a single replacement have spent the last two years walking it back.
How we do this ourselves
Data Connectors brings your first-party base together: Shopify, Google Analytics 4, BigQuery, scheduled file drops, and LiveRamp for identity and audience connectivity. Audience Intelligence builds the segmentation on top, and Ads Manager pushes those audiences out rather than rebuilding them per platform. We are honest about what we do not connect to.
How Data Connectors work