PostHog vs Plausible: Features, Real Costs and Which to Pick

PostHog vs Plausible comes down to the question you need answered. Plausible is simple, cookieless web analytics: visitors, sources, pages and goal conversions on one dashboard, with a script of about 1.3 KB gzipped. PostHog is a product analytics suite: funnels, retention, session replay, feature flags and experiments tied to individual users, free up to 1M events a month. Use Plausible for a content or marketing site, PostHog for a software product, and both if you have both.

This comparison covers features, pricing with a worked cost example, cookie banners (PostHog can run without one, which most comparisons miss), self-hosting, data retention, and the places where the popular comparison pages contradict each other and the vendors’ own docs. All vendor facts were checked against official pages on October 6, 2026.

PostHog vs Plausible at a glance

The table hides one important point: these are not two versions of the same product. Plausible tells you how many people came, from where, to which pages, and how many completed a goal. PostHog tells you what a specific user or cohort did inside your product, and lets you change what they see.

What Plausible does well

Plausible puts everything on one page: unique visitors, visits, pageviews, bounce rate, visit duration, sources and UTM campaigns, top and entry pages, countries, devices and goals. Nobody needs training to read it, which is why it suits teams where marketers, writers and founders all check the numbers.

• No cookies and no stored identifiers. It counts unique visitors with a daily hash of the site domain, IP address and user agent, salted with a value that is rotated and deleted every 24 hours, according to its data policy. A returning visitor is a new visitor the next day.
• A tiny script. We fetched plausible.io/js/script.js on October 6, 2026: 2.8 KB raw, about 1.3 KB gzipped.
• Goals and custom events. Form submissions, outbound links, file downloads and your own events can be goals on every plan. Funnels, custom properties and ecommerce revenue are on the Business plan.
• Sharing. Public dashboards and shared links make it popular with agencies and open-startup founders.

What it does not do: session replay, heatmaps, user profiles, cohorts and retention by user, feature flags or experiments. If you need any of those, Plausible is the wrong tool, not a cheaper version of the right one.

What PostHog does that Plausible does not

PostHog started as open-source product analytics and now bundles web analytics, session replay, feature flags, experiments, surveys, error tracking, a data warehouse, data pipelines, LLM observability and logs. Each product has its own free allowance and its own meter.

• Users, not just visits. Call identify() after login and PostHog links a person’s events across sessions and devices. That is what makes retention curves, cohorts and “why did this account churn” questions possible.
• Autocapture. Clicks, form submits and pageviews are recorded without writing event code, which helps early on and inflates volume later.
• Replay, flags and experiments in one place. You can roll a feature out to 10% of users, watch recordings of those who struggled, and measure the result against a control group.
• SQL. You can query raw events directly, which matters when the built-in reports stop short.

The price is complexity. The interface is dense, useful analysis means planning events and properties, and the full browser bundle is far heavier than Plausible’s. Teams that only wanted a traffic dashboard often find PostHog is more tool than they need. If you are already on PostHog and that is your situation, our guide to PostHog alternatives maps each PostHog feature to a replacement.

Cookie banners: both can run without one

Most comparisons say PostHog “requires” a cookie banner and Plausible does not. Half of that is out of date. By default, PostHog stores a distinct ID, session ID and device ID in a first-party cookie and localStorage, and that kind of storage is what consent rules care about. But PostHog has an official cookieless mode, documented in its cookieless tracking tutorial:

• cookieless_mode: ‘always’ stores nothing in cookies, localStorage or sessionStorage. PostHog counts users with a server-side hash of the project, a daily salt, IP address, user agent and hostname. This is the same idea as Plausible’s method.
• cookieless_mode: ‘on_reject’ shows your banner, uses full tracking for visitors who accept, and still counts visitors who decline with the hash.

The trade-offs, per PostHog’s docs, are real:

• Session replay and surveys are disabled for visitors without consent.
• Weekly and monthly unique users run high, because the same person gets a new hash every day.
• GeoIP enrichment and bot detection stop, so the web analytics world map stays empty.
• You should avoid identify(), because a persistent ID is personal data. Set person_profiles: ‘never’ to make it a no-op.

So the honest version is: Plausible is cookieless with nothing to configure. PostHog can be cookieless, but you give up the features that made you choose PostHog. Whether you need a banner at all depends on where your visitors are and what else your site loads, not only on the analytics tool. Neither vendor setting is legal advice. Our guide to cookieless tracking explains how hash-based counting works and what it costs you in accuracy.

Pricing: PostHog vs Plausible cost

Plausible

Plausible’s pricing table lists three self-serve plans at the 10k monthly pageview tier: Starter at $9 (1 site), Growth at $14 (up to 3 sites and 3 team members) and Business at $19 (up to 10 sites, plus funnels, custom properties, revenue attribution and the Stats API). Prices rise with volume, yearly billing gives two months free, and there is a 30-day trial with no card. There is no free cloud plan.

The detail most comparisons skip: Plausible’s billing docs say your tier is based on pageviews and custom events across all sites in the team. Outbound link clicks, file downloads, form submissions and 404 tracking all count. One month over the limit is tolerated; two consecutive months over means an upgrade request.

PostHog

The PostHog pricing page lists a monthly free allowance that resets every month, with or without a card: 1M analytics events, 5K session recordings, 1M feature flag requests and 1,500 survey responses, among others. Beyond that, each product is billed per unit, with volume discounts, and you can set billing limits. PostHog’s cost-cutting docs also say anonymous events are cheaper than identified ones, so how often you call identify() changes the bill.

Worked example: two sites, two bills

The numbers in this section are illustrative. Plug in your own traffic. The point is to count what each vendor counts.

Site A needs a Plausible tier above 10k, so it pays every month. On PostHog it sits at about a fifth of the free allowance and pays nothing for analytics. Cost alone favors PostHog here, but Site A would be paying with complexity and a heavier script for features it does not use.

Site B pays PostHog for the 800,000 events above the free allowance. It can cut that by trimming autocapture with allow lists, dropping pageleave events, and calling identify() once per session. Plausible would be cheap here, but cheap and unable to answer the question is the expensive option.

Data retention

None of the comparison pages mention how long each vendor keeps your data, and the difference is large. The figures below come from the same Plausible pricing table and PostHog pricing page linked above.

If you want year-over-year comparisons on PostHog’s free plan, export the data or add a card before the first year runs out. On Plausible, three years covers most content sites.

Self-hosting Plausible vs PostHog

Both are open source, and the gap in effort is wide.

• Plausible Community Edition is AGPL-licensed and runs with Docker Compose. It needs Postgres and ClickHouse, and ClickHouse wants memory headroom. The Community Edition is community supported and lags the cloud product on some features. A mid-size VPS is a realistic floor.
• PostHog offers a free “hobby” Docker Compose deployment under the MIT license. Its own self-host docs say self-hosted instances are officially unsupported, there are no tagged releases, and you accept the risk of data loss. Its decision flowchart treats more than 300k events or 1k recordings as the point where you need real infrastructure skills, and it recommends PostHog Cloud for most users.

For both, the real cost is not the server. It is backups you have tested, upgrades on the maintainer’s schedule, and the afternoon you lose when ingestion falls over. Self-host when you have a data residency requirement or genuinely enjoy running ClickHouse. Otherwise the cloud plans are cheaper than your time.

Where the comparison pages contradict each other

Reading the pages that rank for this query side by side, several facts disagree. Here is what the vendors’ own pages say:

Pricing and limits change often. Treat any third-party table, including this one, as a starting point and confirm on the vendor page before you budget.

Ad blockers and proxying

Both scripts are on common blocklists, so both undercount, and the gap is bigger for technical audiences. Both vendors let you serve the script and the event endpoint from your own domain through a reverse proxy. Two things to get right: proxy the event endpoint, not just the script, and forward the real client IP. If every request arrives from your proxy’s IP, country data collapses to one row and the daily hash merges different visitors into one. A geography chart that flattens on the day you shipped a proxy is the tell.

Installation in one paragraph each

Plausible: add one deferred script tag with your domain to the site head; on Next.js, use next/script in the root layout. Single-page app route changes are tracked by the script. PostHog: install posthog-js, initialize it once with your project token and host, and in a Next.js App Router project wrap the app in a client-side provider so route changes send pageviews. Then decide which events to name, which autocapture to keep, and where to call identify(). That planning step is where most of PostHog’s setup time goes.

Can you run both?

Yes, and many teams do: Plausible on the marketing site and blog, PostHog inside the app after sign-up. Plausible answers “which pages and sources bring sign-ups”; PostHog answers “what do those users do next.” Keep the boundary clean. If both scripts run on the same marketing pages, you pay twice for the same pageviews and get two visitor counts that will never match, because one hashes daily and the other uses a persistent ID.

Turning either tool into revenue by channel

For a marketing site, the question behind “PostHog or Plausible” is usually “which pages and channels bring customers.” Neither tool answers it out of the box:

• Both see referrers, so Google, Bing and AI assistants such as ChatGPT or Perplexity appear as sources when they pass a referrer. Neither can show the search keyword, because search engines do not send it.
• Plausible gates funnels and revenue attribution to the Business plan. On lower plans you get goal counts by source and page, without values.
• PostHog can do it, but you build it: a conversion event with a value property, a breakdown by referring domain or UTM source, and landing page captured on the first pageview.

Whichever tool you pick, the arithmetic is the same:

Illustrative: a pricing-comparison post brings 90 organic visits a month that start a trial at 4%, and a trial is worth $60 to you. That is 3.6 trials and $216 a month from one page. A how-to post with 2,000 visits converting at 0.2% gives 4 trials, $240. Traffic ranks them about 22 to 1; value ranks them almost even. Our guides on calculating conversion value and SEO conversion tracking walk through the setup.

Where Umami, Matomo and GA4 fit

• Umami is the closest thing to Plausible: simple, cookieless, MIT-licensed, and lighter to self-host because it runs on Postgres or MySQL without ClickHouse. Pick it if you want Plausible-style stats on a server you already run.
• Matomo is a full Google Analytics-style suite: acquisition reports, goals, ecommerce, tag manager and optional heatmaps, self-hostable on PHP and MySQL. Against PostHog, Matomo is for marketing analytics on websites; PostHog is for product behavior in apps. Against Plausible, Matomo is deeper and harder to read.
• Google Analytics 4 is free and goes furthest on attribution, audiences, Google Ads and BigQuery export, at the cost of cookies, consent setup and a steep interface. If you are weighing GA4 against both, our GA4 alternatives guide groups the options by job.

Who should pick which

The forum threads that rank for this query ask the same thing: “which is better?” The useful answer is to write down the three questions you will ask your analytics most often. If they are about pages and sources, pick Plausible. If they are about users and features, pick PostHog.

Frequently asked questions