GA4 or PostHog: which one should an early-stage startup use?

If you are setting up analytics for a young company, you have probably reached the GA4 vs PostHog question. Both are free to start, both have fans, and most comparisons are written by one of the two sides. I have used GA4 with nearly every client I have worked with, and I chose PostHog for my own two websites this autumn, so I can tell you how I decide.

GA4 vs PostHog: what is the actual difference?

GA4 is Google’s free website analytics tool. It is built to tell you how many people visit, where they come from and which campaigns bring them in. PostHog is a product analytics tool. It is built to tell you what people do once they arrive: which steps they complete, where they drop off, and whether they come back.

There is overlap, because both can count visitors and both can show a funnel. The difference is in what each one does well. In my experience, GA4 answers marketing questions comfortably and product questions badly. PostHog is the other way round, although its web analytics view has become good enough for a small site.

PostHog also puts several tools in one place. Alongside the analytics there are session recordings, heatmaps, feature flags, A/B tests and surveys. With GA4 you would need a separate tool for each of those.

When I choose GA4

I choose GA4 when the main questions are about acquisition, and above all when there is real money going into Google Ads.

Linking GA4 to Google Ads lets Google import your key events and use them for bidding, and it lets you build remarketing audiences from what people did on your site. At Augmentive, Google Ads was where the work started, so GA4 and Tag Manager were part of the setup from day one. No other tool would have done that job as well.

I also keep GA4 when a team already knows it and only needs it to be right. In that case the job is to fix it, and these are the six checks I run on GA4 first.

One thing to know before you rely on it. On a standard GA4 property, explorations and funnel reports only keep the detailed data for 2 or 14 months. If you use GA4, set it to 14 months on day one, because you cannot get the lost months back.

When I choose PostHog

I choose PostHog when the most important journey happens inside the product or a multi-step form, and the developers are the people who will add the events.

That is exactly the situation I was in with RoastMyFlow, the product I built with Nazia Hassan. The original plan was GA4 plus Mixpanel. In October I changed my mind and set up PostHog on its own, for three reasons:

• One tool instead of two. I wanted traffic sources, every step of the sign-up form and the clicks on the blog in one dashboard. PostHog covers all of it.
• Recordings next to the numbers. When a step loses people, I can watch a few recordings of that step straight away. We mask everything people type, so the recordings never show personal details.
• The free plan is generous. At the time of writing, PostHog’s free plan includes 1 million events and 5,000 recordings a month. A young product takes a long time to outgrow that.

I made the same choice for this website, using PostHog’s EU servers. On both sites, nothing loads until a visitor accepts analytics cookies.

When I run both together

Many of the companies I work with need both kinds of answer. They spend on Google Ads, so they need GA4. They also have a booking flow or a sign-up flow that matters more than any single page, so they need a product analytics tool as well.

In that case I run both, and I follow one rule. Both tools get their events from the same written tracking plan, with the same names and the same meaning. At Givematch I set up GA4 and then Mixpanel from a single plan, so the two tools never disagreed about what a donation was. At Qured we did the same with GA4 and Mixpanel before we started testing the booking funnel.

If you are not sure what goes in that plan, I explain it in what I set up first, and in what order.

PostHog vs Mixpanel: which product analytics tool?

This is the other question I get asked a lot. Mixpanel is the product analytics tool I have used most with clients. At Givergy I set Mixpanel up from scratch, and their in-house team went on to run it without me. That is the best test of a tool I know.

Both tools do funnels, retention and user paths well. The main differences, as I see them, are these:

• PostHog does more in one place. Recordings, heatmaps, feature flags, A/B tests and surveys sit next to the analytics, so a small team has fewer tools to pay for and connect.
• Mixpanel is more focused. It does product analytics and does it very well, and non-technical team members tend to find their way around it quickly.
• Your team matters more than the feature list. If your developers already like one of them, that is usually the right choice, because they are the ones adding the events.

For a brand new setup I currently lean towards PostHog. If a client already has a Mixpanel setup that works, I would not move them.

What I would do in your position

Here is the short version of my advice for an early-stage startup:

• You spend real money on Google Ads and your site is mostly marketing pages. Use GA4, set up properly, and add Tag Manager.
• Your main journey is inside the product or a long form, and you do little paid search. Use PostHog on its own.
• You run Google Ads and you have a product funnel that matters. Run GA4 and PostHog (or Mixpanel) together, from one tracking plan.
• You are not sure yet. Start with PostHog’s free plan and a written tracking plan. Adding GA4 later is easy when the events are already defined.

Whatever you pick, remember that a consent banner means you only see the visitors who accept. On my own products I treat the analytics numbers as a floor, and I count real sales from Stripe, which sees every one of them.

How I set this up for clients

Choosing the tool is one small part of my Growth Foundations project. Before anything is installed, we agree in writing what a conversion is and what each event is called. Then your developers add the events from a brief I write, and I check every one of them against real traffic before anyone looks at a dashboard.

If you are about to choose your analytics tools, or you have two tools that disagree, have a look at how Growth Foundations works or book a thirty-minute call below. I will tell you honestly which tool I would use in your position, even if the answer is the one you already have.