Cross-channel measurement in a privacy-first world: how to build a tracking foundation you can trust

(updated ) 6 min read ppctrackingprivacy
Server-side tracking, first-party data, Consent Mode v2 and the Conversions API. The foundation under your advertising on Google, Meta and TikTok.

You can run the sharpest campaigns in the world, but if your tracking is shaky, every bidding algorithm is flying blind. And that foundation is exactly what has shifted under our feet in recent years. Browsers block trackers, visitors dismiss your cookie banner and the pixel in the browser misses a growing share of what really happens. In this post, I explain how to build a tracking foundation that holds up across Google, Meta and TikTok, within the privacy rules.

Why your old tracking setup leaks

First, let’s clear up a misunderstanding. Third-party cookies have not disappeared. In April 2025, Google decided to keep them in Chrome. But that doesn’t solve your problem. Safari and Firefox have blocked them by default for years, and a sizable share of your visitors use an ad blocker that stops your browser pixel.

The result is tracking that consistently comes in too low. You see fewer conversions than there really are, and the ones you do see are harder to tie to the right click. For a bidding algorithm, that’s poison: if your data is incomplete, your budget goes to the wrong place.

The solution comes down to two moves: tracking from your own server instead of only the browser, and feeding that tracking with first-party data you collect with consent.

Server-side tracking: measuring from your own server

With classic tracking, a script in the browser fires directly to Google, Meta or TikTok. With server-side tracking, you first send the data to your own server, usually a server-side Google Tag Manager container, and from there on to the platforms.

That has three concrete benefits:

  • You miss less. Ad blockers and browser restrictions have much less grip on a request that comes from your own domain.
  • You stay in control. You decide which data you pass on and which you don’t. That’s cleaner, and it’s necessary to comply with the GDPR.
  • It’s platform-independent. One setup feeds Google, Meta and TikTok at the same time. One foundation, three channels.

Note: this is not a trick to get around consent. Consent still comes first. It improves data quality within the rules, not around them.

First-party data is your most valuable asset

First-party data is everything you collect yourself in your relationship with the customer: the email address at checkout, a phone number from a form, an order ID. With consent and in hashed form, you pass that data to the platforms, so they can still tie a conversion to the right click.

This is what drives the modern measurement integrations. The more high-quality signals you send along, the better the matching. At Meta, you see this in Event Match Quality, a score for how well your server events can be matched. Advertisers with a high score demonstrably capture more attributed conversions. It works the same way on the other channels. And the better your own data is in order, the less you depend on cookies you have no control over.

If you want to advertise through Google in the EEA, Consent Mode v2 is not optional. You pass on the visitor’s consent choice through four parameters:

  • ad_storage: storage for advertising purposes (cookies).
  • analytics_storage: storage for analytics.
  • ad_user_data: sending user data to Google for advertising.
  • ad_personalization: personalized ads.

Without a correct setup through a Google-certified consent platform, you fall back to almost nothing: no remarketing lists that fill up, no complete conversion tracking in Europe. If you do get consent, Consent Mode partly fills the gaps left by visitors who decline, using modeling.

Watch out for a change that took effect on June 15, 2026: since then, Google Analytics uses the consent settings within Google Ads as the only control for advertising data. The role of Google Signals is narrowed to analytics reporting in GA4. In practice, this means you really need ad_storage, ad_user_data and ad_personalization in order, because they now determine the behavior.

The Conversions API: the same idea at Meta and TikTok

What Consent Mode and server-side tagging do at Google, you handle at Meta and TikTok through their own server-side integrations.

At Meta, that’s called the Conversions API (CAPI). The advice is clear: run the pixel and CAPI side by side. The pixel captures behavior in the browser, and the Conversions API confirms the purchase from your server. Meta now gives more weight to what is confirmed through CAPI. Deduplication is crucial: give the pixel event and the server event the same event_id, or Meta counts the same conversion twice.

At TikTok, the Events API does the same job, and the same rule applies: if you send the same event through the pixel and the Events API, give both an identical event_id. TikTok deduplicates overlapping events with the same event_id within 48 hours on its own. Also note that standard integrations, such as the TikTok integration in Shopify, often only place a browser pixel and don’t automatically send data through the Events API.

What this means for you as an advertiser

Concretely, this is what I would get in order before you put another euro into your campaigns:

  • Set up server-side tracking. One container that feeds Google, Meta and TikTok at the same time. That’s your foundation.
  • Activate your first-party data. Send hashed email addresses, phone numbers and order IDs wherever consent allows it. This is your biggest lever on match quality.
  • Check your Consent Mode v2. Do all four parameters fire, and does your consent platform work well with your tags? Test it.
  • Get deduplication right. One consistent event_id across pixel and server, or you inflate your numbers.
  • Check your standard integrations. An integration that promises “tracking” often only places a browser pixel. Is the server-side part actually running?

Measurement isn’t the finishing touch. It’s the starting point. A bidding algorithm is only as smart as the data you give it. Fix the foundation first, and only then does the rest really work.

Not sure whether your tracking setup is right, or whether you’re quietly under-measuring? Request a Free Performance Quick Scan, no strings attached. I’ll look at where your data leaks and what that costs you.

Sources

Want to win back return?

Start with a Quick Scan and find your 3 biggest opportunities.