- Classic remarketing lists built on site visits are limited now, especially when users reject tracking. First-party audiences and Customer Match are what still reach past visitors. [6]
- Server-side signaling (Meta's Conversions API, Google's Enhanced Conversions) is the durable layer. One agency estimates it recovers 15-30% of otherwise-lost conversion signal. [5]
- Customer Match runs on consented, hashed identifiers, not cookies. One vendor's write-up citing Google's Help docs reports a 5.3% conversion uplift. [3]
- Advantage+ and Performance Max are automation layers now. Feed them first-party data and clean conversions instead of rebuilding the old cookie stack by hand. [3] [6]
- Contextual targeting is a real complement for anonymous traffic, not a fallback. [1] [2]
Your retargeting line item is doing two different jobs, and only one of them still works. The first job is reaching people you already know, past buyers, cart abandoners, account holders. The second is chasing anonymous browsers around the open web with a cookie. That second job is the one that quietly stopped paying, and a lot of budgets are still funding it out of habit.
The reason is not a single deprecation date. It's that the cross-site identifier layer retargeting relied on broke down: browsers block third-party cookies, and consent rejection removes most of what's left. The advice in every credible source now points the same way, toward consented first-party data, server-side events, and platform-native audiences rather than the browser pixel alone. [1] [2]
The identifier did the work. You did not need to know who the person was, only that a browser carried the tag.
You have to capture the identity yourself. What you own, you can still retarget. What you rented from a cookie, you lost.
What still works, and what's now theater
The honest split is not cookie versus cookieless. It's between the tactics that run on data you own and the ones that run on an identifier that no longer resolves.
| What you're doing | Status | Why |
|---|---|---|
| Broad third-party retargeting pools | Theater | the cross-site identifier layer it relied on is gone [1] [2] |
| Browser pixel as the only signal | Theater | blocked or incomplete without a server-side backup [4] [5] |
| "Cookie-free" retargeting with no identity capture | Theater | branding, not capability, until you collect consented IDs [1] |
| First-party audiences (email, CRM, loyalty, app IDs) | Works | the central asset in a cookieless setup [1] |
| Customer Match on hashed lists | Works | consented identifiers, not cookies [3] |
| Server-side events (Conversions API, Enhanced Conversions) | Works | recovers 15-30% of lost signal in one estimate [5] |
| Contextual for anonymous traffic | Works | within 5-8% of behavioral on click-through in one reading [5] |
The through-line is identity capture. First-party data (email signups, account holders, loyalty members, purchase history, app IDs, on-site behavior collected directly) is the asset every source names as central. [1] Everything downstream, the audiences, the automation, the measurement, is only as good as the identifiers you feed it.
Feed the machine, don't fight it
Server-side is the layer that makes the rest hold up. Sending events from your backend, through Meta's Conversions API, Google's Enhanced Conversions, or offline imports, is what survives a browser that drops the cookie. These tools exist specifically to solve the deprecation problem, and one agency puts the recovered signal at 15-30% of what pixels miss. [4] [5] If you want the mechanics of moving collection off the browser, we covered the plumbing in server-side tracking, and the consent side that gates all of it in Consent Mode.
Advantage+ and Performance Max are the execution layer, not the strategy. Both are automation systems that reward better inputs over manual tinkering. The practical guidance for Advantage+ and Performance Max is the same: improve the signal with first-party data and clean conversions, and stop trying to rebuild the old cookie retargeting stack by hand. Poor signal quality, not a shortage of targeting knobs, is the limiting factor. [3] [6]
The line in 2026 is not cookie versus cookieless.
It's between teams doing consented, incrementality-aware first-party activation and teams paying to re-run cookie-era theater. One has an asset that compounds. The other has an invoice.
So the working stack is short, and it's in order for a reason. Get the identity first, then the signal, then let the platforms do what they're good at.
- Capture consented identifiers before anything else: email, phone, account, loyalty, app install, checkout.
- Wire server-side events (Conversions API, Enhanced Conversions, offline imports) so signal survives a blocked pixel.
- Upload and refresh hashed Customer Match lists, and treat them as the retargeting base, not a bonus.
- Feed Advantage+ and Performance Max clean conversions and creative rather than manual list edits.
- Use contextual to prospect into anonymous traffic instead of resurrecting cross-site pools.
- Test for incrementality. A retargeting audience that would have converted anyway is a cost, not a win.
Retargeting didn't die. It stopped being free. The reach that a cookie handed you for nothing now has to be earned by capturing the identity, sending the signal, and proving the lift. Cut the theater, move that money into first-party activation, and you still have a retargeting program worth the name next year.
Sources
- Adrenalead · Advertising without cookiesnames first-party data as the central pillar of a post-cookie setup
- Usercentrics · Google and third-party cookieslists first- and zero-party data, server-side tagging, identity solutions, and contextual as the resilient approaches
- Xictron · Retargeting without third-party cookies in 2026reports the Customer Match 29-62% match rate and 5.3% uplift, citing Google Ads Help; not independently primary-verified, treat as directional
- Cometly · Third-party cookie deprecation impactframes the Conversions API and Enhanced Conversions as the tools built to solve deprecation
- Digital Applied · Cookieless marketing strategy 2026the 15-30% server-side recovery and 5-8% contextual figures are secondary commentary citing DoubleVerify and IAS, not primary studies
- Ads Made Awesome, with Andy Splichal · Cookieless Google Ads remarketingclassic site-visit remarketing lists are limited when users reject tracking; first-party audiences and Customer Match remain usable
Frequently asked questions
Is retargeting dead after third-party cookies?
No, but the cookie-based version of it is fading. Classic remarketing lists built on site visits are limited now, especially when users reject tracking, per the practitioner guidance in an Ads Made Awesome session. What still reaches past visitors is consented first-party data: Customer Match lists, engagement audiences, and platform-native retargeting fed by server-side conversion signals.
What replaces cookie-based site retargeting?
A bundle, not a single tool. Capture consented identifiers (email, phone, account, loyalty, app installs), send events server-side through Meta's Conversions API or Google's Enhanced Conversions, activate hashed Customer Match lists, and let Advantage+ and Performance Max use those signals. Contextual targeting fills the gap for anonymous traffic you cannot identify.
Does server-side tracking actually recover lost conversions?
It recovers some signal that browser pixels miss, because events come from your backend rather than a cookie the browser may block. One agency estimate puts the recovery at 15-30% of otherwise-lost conversion signal, though that figure is secondary commentary rather than a primary study, so treat it as directional and measure your own lift.
How do Advantage+ and Performance Max retargeting work without cookies?
Both are signal-driven automation layers now, not manual audience tools. You improve results by feeding them better inputs, first-party data, clean conversion tracking, Customer Match, and consent mode, rather than rebuilding the old cookie retargeting stack by hand. Poor signal quality, not a lack of targeting knobs, is usually the limiting factor.
Is contextual targeting a real retargeting substitute?
Not a direct substitute, but a durable complement. Contextual reaches anonymous traffic you cannot identify and, in one blog's reading of DoubleVerify and IAS figures, comes within 5-8% of behavioral targeting on click-through. Use it to prospect into unknown visitors rather than to resurrect the old cross-site retargeting model.
What is Customer Match and does it lift conversions?
Customer Match activates hashed, consented identifiers like email and phone rather than cookies, so it keeps working after third-party cookies fade. One vendor write-up citing Google's Help docs reports a 5.3% conversion uplift and match rates of 29-62%, though those come from a single source rather than an independent study, so treat them as directional. Upload and refresh the lists and treat them as the retargeting base, not a bonus.
What counts as first-party data for retargeting?
It is the identity you capture directly: email signups, account holders, loyalty members, purchase history, app installs, and on-site behavior you collect yourself. Every credible source names it as the central asset of a post-cookie setup, because the audiences, automation, and measurement downstream are only as good as the identifiers you feed them. What you own you can still retarget, and what you rented from a cookie you lost.
Should I test my retargeting for incrementality?
Yes, because a retargeting audience that would have converted anyway is a cost, not a win. Past buyers and cart abandoners often intercept sales already in motion, so a healthy-looking return can hide spend that created nothing. Run incrementality tests before you credit the channel, the same discipline that separates real lift from recycled demand.
Mikołaj Salecki
Writes about media, tech, and AI business for people who actually run digital. Former agency lead. Skeptic of frameworks that read better than they perform.
Get the next issue by email.
One letter, once a week. Sharp coverage of media, tech, and AI business. No filler.
