PostHog vs Google Analytics: pricing, sampling limits, and where Google Analytics vs PostHog and GA4 actually differ

Yes, GA4 is genuinely free for standard properties, with no monthly bill and no published cap on how many events you can collect. Google makes its money from the advertising products GA4 feeds, not from the analytics tool. The cost arrives as limits rather than invoices.

Those limits are published and worth reading before you commit an instrumentation strategy to them. A standard property retains user and event level data for at most 14 months, samples explorations above 10 million events per query, allows 50 event-scoped custom dimensions and 50 custom metrics, 25 parameters per event, 30 key events, 100 audiences and 5 calculated metrics, and exports at most 1 million events per day to BigQuery. Every one of those numbers roughly doubles or better on Analytics 360.

GA4 360 is the paid tier, and it has no public price list at all. Contracts are quoted, sold on an annual commitment through Google or a reseller, and reported entry points sit around $50,000 a year rather than the $150,000 figure still repeated across the web, which dates from the Universal Analytics era. Treat any specific 360 number you read, including that one, as reported rather than official. The honest summary is that GA4 is free until your questions outgrow the limits, and the next step up is a negotiation, not a price you can look up.

For understanding what users do inside a product, yes, clearly. For understanding how people found your website, no, and it is not close. They are built for different halves of the funnel and the better tool is decided entirely by which half you are asking about.

Google Analytics has a decade and a half of accumulated integration with the rest of the advertising and search ecosystem. Campaign attribution, channel grouping, Search Console and Google Ads integration, and the reporting vocabulary every marketer already speaks. Nothing in PostHog competes with that, and replacing GA4 for marketing measurement is a downgrade nobody should attempt on principle.

PostHog wins the moment the question involves a logged-in user. Unsampled queries, seven year retention on paid plans, unrestricted event and person properties, session replay to watch the drop-off you just measured, and feature flags to change it. GA4 offers none of those. It also has no replay, no experimentation since Google Optimize was retired in 2023, and no surveys, so a team using GA4 for product work usually ends up buying two or three other tools anyway. The comparison people should really be making is not PostHog against GA4, but PostHog against the combined cost of the tools GA4 forces you to buy around it.

There is a free plan, and it is unusually generous in breadth, but it is not free in the way GA4 is free. PostHog gives a separate monthly allowance on each of roughly eleven metered products, and once any one of them runs out you add a payment method and start paying published per-unit rates.

The monthly free allowances are 1 million analytics events, 5,000 web session recordings, 2,500 mobile recordings, 1 million feature flag requests, 1,500 survey responses, 100,000 error tracking exceptions, 1 million data warehouse rows and 10 GB of logs. A small product can run genuinely useful analytics, flagging, error tracking and surveys on that without paying anything, which is far more than any free tier in the category offers.

What usually ends a PostHog free account is not the event volume. It is the two structural limits: a free account gets one project and one year of data retention, against six projects and seven years once you add a card. Teams with a staging environment hit the project wall, and teams doing any year over year analysis hit the retention wall. Both arrive before the million events do for most B2B products. Add the identified event multiplier, where logged-in traffic costs up to four times the anonymous rate, and the realistic answer is that PostHog is free to evaluate and priced to run.

Use both, and be deliberate about which question goes to which tool. That is not a hedge, it is what most teams running PostHog actually do, because dropping GA4 costs them marketing measurement they still need and dropping PostHog costs them product measurement GA4 cannot provide.

The split that works: GA4 owns everything before signup. Acquisition channels, campaign performance, landing page conversion, Search Console and Ads integration. PostHog owns everything after signup. Activation, feature adoption, funnels through the product, retention by cohort, session replay, flags and experiments. The handover point is the account creation event, and instrumenting that event in both tools is what lets you connect a channel to a retained customer later.

If you genuinely must run one, decide by where your unanswered questions live. A company selling a website, a media property or an ecommerce catalogue should stay on GA4 and spend the saved money elsewhere. A B2B SaaS company whose growth problem is activation or churn should run PostHog and accept that marketing attribution gets thinner. The one option to avoid is stretching GA4 across both jobs and quietly making product decisions on sampled data, which is the situation most people are already in when they search for this comparison.

GA4 counts sessions and traffic on a website. PostHog counts what identified people do inside a product. Every other difference, including the pricing model, follows from that one distinction.

The data models diverge immediately. GA4 organises around sessions, channels and campaigns, with users as a derived concept and a session definition that restarts on a timeout. PostHog organises around persons and events, where a person persists across sessions and devices and carries properties you define. When you want to ask which accounts on the enterprise plan used the new export feature in the last quarter and then renewed, you are asking a person-shaped question, and GA4's session-shaped model fights you the whole way.

The second structural difference is scope. GA4 is one product. PostHog is roughly eleven, billed separately: analytics, session replay, feature flags, experiments, surveys, error tracking, data warehouse, pipelines, logs and AI observability. That makes PostHog a consolidation play, which is how most teams justify the cost, since it can absorb what would otherwise be a separate flagging vendor and a separate error tracking vendor. Both tools share one real limitation worth naming: they both record behaviour and neither can tell you why. That answer lives in tickets, reviews and survey replies, not in the event stream.

PostHog supports GDPR compliant deployments and offers EU cloud hosting, but compliance is a property of how you configure and use a tool, not a badge the vendor hands you. The same is true of GA4, and this is one area where the two have a genuinely different history.

GA4's complication is that data is processed by Google and has historically involved transfers to US infrastructure, which produced a run of adverse rulings from European data protection authorities against the previous Universal Analytics. Google has since made changes, and the legal position has moved, but any European team should verify the current position with its own counsel rather than trusting a comparison page, including this one. PostHog offers an EU cloud region, gives you controls over what is captured, and supports anonymising events, which is also the cheaper billing mode.

The practical advice for either tool is the same and it is unglamorous. Decide what you actually need to capture before you instrument anything, avoid sending personal data into an analytics event unless you have a reason and a lawful basis, use the anonymous event mode where identification is not needed, and document your retention settings. Doing that on PostHog has a pleasant side effect, since anonymous events are billed at a large discount to identified ones.

Yes, and it is the most common arrangement among PostHog customers rather than an unusual one. The two tools overlap far less than the comparison framing suggests, so running both duplicates less work than running two product analytics tools would.

The setup is straightforward. Both are JavaScript snippets and both can be deployed through a tag manager. GA4 stays on the marketing site and the signup flow, PostHog goes into the authenticated product. The important detail is the handover: capture the signup or account creation event in both tools with a shared identifier, so a marketing channel in GA4 can later be connected to a retention cohort in PostHog. Without that shared key you get two accurate systems that can never answer a question spanning both.

Budget for the ongoing cost of the arrangement, which is not money but maintenance. Two instrumentation paths drift apart unless one person owns keeping the event definitions aligned, and the two tools will report different numbers for anything that looks superficially similar, because a GA4 session and a PostHog person are not the same object. Agree in advance which tool is authoritative for which metric and write it down, or you will spend the next year relitigating whose number is right in meetings.