Do I still need GA4 when I'm using PostHog?

Why this question matters more in 2026 than ever

Anyone setting up a new website today — whether a classic marketing site, SaaS project, or lead-gen funnel — eventually arrives at the question of web analytics. And almost every developer, founder, or marketing lead asks themselves the same thing:

"We have a modern analytics tool like PostHog, Plausible, or Matomo in place. Do we still need GA4 on top — for SEO, for Google Ads, for audience building?"

The answer you'll read on most blogs amounts to: "Better to run both, doesn't hurt to have it." From our perspective as a performance marketing agency, that's a reflex that creates more overhead than value. We've been seeing for years that GA4, in a surprisingly high share of our mandates, contributes nothing that a modern stack with server-side conversions doesn't already deliver — while adding legal and operational complexity.

Let's go through this using the three most common misconceptions we keep hearing in mandate conversations.

Misconception 1 — "GA4 helps SEO rankings"

It doesn't. This assumption is stubborn but factually wrong.

Google itself has clarified this multiple times over the past years — representatives like John Mueller and Gary Illyes have repeated it in official statements, podcasts, and on the Search Off the Record format — that GA4 data is not a ranking signal. The Google Search algorithm does not have direct access to your GA4 property data. That wouldn't be defensible under antitrust or privacy law.

What Google actually uses for rankings:

• Google Search Console — click and impression data from the search results themselves
• Core Web Vitals as field data from the Chrome User Experience Report (CrUX)
• Backlink profile, domain authority, on-page factors
• Content quality, semantic relevance, crawl behavior

GA4 plays no role in any of these mechanisms.

What you actually need for SEO: Google Search Console. It's mandatory, free, privacy-clean, and delivers the data Google itself has about your visibility. For Core Web Vitals, either Search Console or a modern analytics tool like PostHog with web vitals capture covers it — both surface the same fields (LCP, CLS, INP).

Misconception 2 — "Google Ads audiences don't work without GA4"

This answer is more nuanced — and the nuance is rarely made cleanly.

Remarketing lists

The most important function: targeting people who've visited your site with ads later. This does not require GA4.

The Google Ads tag (gtag.js) sets the _gcl_aw cookie on its own — assuming it's correctly embedded and consent is given — and builds your remarketing list directly inside your Google Ads account. Lists like "all website visitors of the last 30 days" or "visited services page without conversion" work without GA4.

Lookalike and similar audiences

The situation fundamentally changed in 2023. Google discontinued Similar Audiences and replaced them with automated mechanisms:

• Optimized Targeting in Display and Search campaigns
• Audience Expansion in Performance Max campaigns
• Smart Bidding on conversion value

These mechanisms don't need GA4 audiences as input. They need good conversion signals. That's exactly what we deliver with the PostHog → Google Ads Conversions destination — at least as cleanly as a GA4 setup, often more cleanly, because in PostHog we can freely filter conversion conditions (e.g. "only count Calendly booking when status = confirmed and UTM source = google").

Cross-device tracking via Google Signals

This is the one area where GA4 historically had a real technical edge: for users signed in to their Google account, GA4 can recognize them across devices. For e-commerce sites with high cross-device share (browse on mobile, buy on desktop), this can matter.

For a lead-gen site with server-side conversion reporting, it's nearly irrelevant. The Calendly booking event is sent with a hashed email address via Enhanced Conversions to Google Ads — Google handles the cross-device matching itself through its login profile, regardless of whether a GA4 tag was firing.

Misconception 3 — "More data points make Smart Bidding better"

We hear this argument often, and it sounds intuitive. But it's backwards.

Smart Bidding doesn't optimize on pageviews. Smart Bidding optimizes on conversion signals.

What makes the algorithm strong:

1. Clearly defined conversion actions — not twenty, but three to five important ones, with clear hierarchy and realistic conversion values
2. High Event Match Quality — meaning gclid plus Enhanced Conversions with hashed email or phone, so Google can reliably attribute the conversion to the ad click
3. Sufficient volume — rule of thumb: at least around 30 conversions per month per campaign, or the algorithm doesn't learn stably
4. Conversion values that reflect real business value — a Calendly booking is worth more than a newsletter signup, and that has to be reflected in the value property

All of this is delivered cleanly by a properly configured PostHog → Google Ads Conversions destination. More GA4 pageviews in the background don't make the model better. They can make it worse, if they're defined sloppily and Smart Bidding learns false correlations.

The one exception — Performance Max with audience signals

We're arguing clearly against GA4 as a standard setup in this article. But there is one use case where GA4 brings real value — and we don't want to hide it, or it would surprise you later.

Performance Max (PMax) is Google's automated cross-channel campaign type (Search, Display, YouTube, Discover, Gmail, Maps in one). PMax uses so-called audience signals as input — hints about what your ideal customers look like. PMax then finds similar users in the Google ecosystem.

Audience signals can be:

• Customer Match lists (hashed email addresses of your existing customers)
• Existing remarketing lists
• GA4 audiences with behavioral granularity — e.g. "users who have read more than three blog articles"

Point 3 is the gap. PostHog cannot provide these audiences in a PMax-compatible format. If you want to run an extensive PMax strategy with your own behavioral audiences, GA4 is the faster path there.

When is this relevant for you? From our mandate practice: at typical ad spend from around €5,000/month and a site with enough content for meaningful behavioral segmentation. Below that, PMax with audience signals is technically possible but statistically too thin — the algorithm has too little training data.

For a newly launched marketing site with campaigns just ramping up: not relevant yet. GA4 can be added later without losing historical data — audiences rebuild themselves over the following 30 days anyway.

What you should set up instead

From our mandate practice, this is the stack combination that works most cleanly for lead-gen sites and SaaS projects in 2026:

1. Google Search Console. Mandatory for SEO. Free, no privacy discussion, delivers the data Google itself has about you.

2. PostHog (EU cluster). Web analytics, funnels, heatmaps, session replays, conversion definition. GDPR-compliant with DPA, EU hosting in Frankfurt, fully consent-gated. Everything GA4 does — plus product analytics features GA4 doesn't have.

3. Google Ads tag (gtag.js) in conversion-linker mode. Not as a tracking tag, but only to set the _gcl_aw cookie and persist the gclid. That makes remarketing lists and Enhanced Conversions work cleanly.

4. PostHog → Google Ads Conversions destination. Server-side transmission of real conversions (Calendly bookings, contact form, lead magnets) to Google Ads. With hashed email for Enhanced Conversions and maximum Event Match Quality.

5. Meta Conversions API (optional). If you also advertise on Meta: PostHog → Meta Ads Conversions destination as a server-side complement to the pixel.

This gives you complete performance marketing tracking that is privacy-compliant and doesn't require a GA4 migration in two years when the next regulator decision lands.

What you avoid

Deliberately skipping GA4 doesn't just mean "fewer tools". You avoid a list of concrete problems:

• Schrems II discussion with regulators — even after the EU-US Data Privacy Framework 2023, GA4 remains a recurring audit topic in several EU countries
• Cookie consent complexity — GA4 only works partially without Consent Mode v2, and the modeling layer gets technically and conceptually demanding fast
• Privacy policy risk — GA4 has been the most frequently flagged web analytics tool by German data protection authorities since 2022
• Data model complexity — the GA4 event model with custom dimensions, custom definitions, and explorations has a noticeable learning curve that many teams haven't built
• Sampling and thresholding — GA4 reduces your data at higher volumes without always being transparent about it

Conclusion

If you're setting up a new website in 2026 and have a modern analytics tool like PostHog in place, you don't need GA4 in most cases. The commonly cited arguments — SEO benefit, Google Ads audience building, more data points for Smart Bidding — don't hold up to closer inspection. What you really need is clean server-side conversion transmission to Google Ads, the Google Ads tag in conversion-linker mode, and Google Search Console.

The one real exception — Performance Max with your own audience signals — is only relevant for most mandates at significant ad spend and can be added later.