PostHog Alternatives: The Right Replacement for Each Job

The best PostHog alternative depends on which part of PostHog you want to replace. For product analytics, Mixpanel and Amplitude are the usual picks; for session replay, FullStory, LogRocket or Microsoft Clarity; for feature flags and experiments, Statsig, GrowthBook or LaunchDarkly; and for simple privacy-friendly traffic stats, Plausible, Fathom, Umami or Matomo. Because PostHog bundles many products, replacing it often means one focused tool, not one like-for-like clone.

Almost every page ranking for this query is written by a vendor that wants to be your replacement, PostHog included. This guide groups the options by job, covers the open-source and free routes, answers the PostHog vs Statsig and PostHog vs Plausible questions, and then deals with the parts the comparisons skip: estimating your own volume, why your numbers will not match after a switch, and what a marketing site should measure instead.

The best PostHog alternative for each job

What PostHog does, and what you would be replacing

PostHog started as open-source product analytics and grew into a suite. Today it covers product analytics (trends, funnels, retention, paths, cohorts and SQL access), web analytics, session replay with console and network logs, feature flags, experiments, surveys, error tracking, LLM observability, a data warehouse, a CDP, logs and workflows. Each product is billed separately once you pass its free allowance.

That free allowance is large. PostHog’s pricing page lists 1M product analytics events, 5K session recordings and 1M feature flag requests per month for free, plus separate allowances for error tracking, surveys and other products, and says 97% of companies use it for free. So before you leave, be clear about the reason. If the reason is cost, you may only be paying for one or two meters.

Why teams look for a PostHog alternative

The same reasons come up across vendor comparisons and PostHog’s own “when not to use us” list:

• It is built for engineers. Product managers, marketers and support teams often find the interface dense, and advanced questions can end in SQL. Adoption stalls outside the engineering team.
• All-in-one means broad, not always deep. Dedicated tools go further in their one area: experimentation statistics, frustration signals in replay, front-end performance monitoring, or flag governance with approvals and audit logs.
• Pricing runs on several meters. Events, recordings, flag requests, survey responses and exceptions are billed separately, which makes the bill harder to predict as traffic grows.
• Self-hosting is not the cloud product. PostHog’s self-hosting docs say it offers no customer support for self-hosted instances, that paid-plan features are cloud-only, and that it has “literally never seen” the self-hosting cost math work out.
• Missing pieces. PostHog has no no-code builder for product tours or tooltips, and no multi-touch marketing attribution.
• Client-side data gaps. Ad blockers block requests to known analytics domains, so PostHog recommends routing events through a reverse proxy on your own domain. Every browser-based tool on this list has the same problem.

Start with an inventory of what PostHog does for you

None of the comparison lists start here, but it decides everything else. Open PostHog’s billing and usage views and write down, for each product, whether anyone used it in the last 90 days and who. Then fill in a table like this one:

The example rows are illustrative. The point is the result: most teams find they rely on two or three products, not twelve. If only one row says “must keep,” a focused tool is the obvious answer. If four rows do, count the cost of running four vendors, four SDKs and four data models before you decide PostHog was the problem. Ask one more question for each row: is the issue the tool, or that nobody built a routine for reading its reports?

PostHog alternatives by category

Product analytics: Mixpanel, Amplitude, Heap

Mixpanel is the closest like-for-like replacement for PostHog’s analytics. It is event-based, with funnels, retention, flows and cohorts in an interface non-technical users pick up quickly, plus Metric Trees for mapping input metrics to a north star metric and data governance tools for cleaning up event names. It has added session replay, heatmaps, experiments and feature flags. It is not open source and cannot be self-hosted.

Amplitude suits larger product and growth teams. It adds predictive cohorts, multi-touch attribution, a built-in CDP, experimentation, and guides and surveys. The trade-offs are a steeper learning curve and paid tiers that require talking to sales.

Heap is known for autocapture with retroactive event definition: capture every interaction, then name the events that matter later through a visual labeler. Heap is now part of Contentsquare, and PostHog’s own comparison notes that Contentsquare is moving Heap customers to its platform, so evaluate where Heap is heading before committing. Autocaptured data also gets noisy unless someone curates it.

Also in this group: Countly, which pitches governed, self-hosted analytics with commercial support for teams that need agreements such as a DPA or BAA, and Userpilot, which pairs analytics with in-app engagement but has no free plan.

Session replay and UX: FullStory, LogRocket, Hotjar, Microsoft Clarity

FullStory is the choice for UX research teams: polished replay search, frustration signals (rage, dead and error clicks), mobile heatmaps and conversion maps. Pricing is through sales. LogRocket is for front-end engineers: replays come with console logs, network requests, state snapshots and performance data, which makes bugs reproducible. Its product analytics are basic. Hotjar is a lightweight replay and heatmap tool for marketing and UX teams, and Microsoft Clarity is free, though not open source.

Feature flags and experiments: Statsig, GrowthBook, LaunchDarkly and others

Statsig is the most common pick for teams leaving PostHog over experiments. It offers CUPED variance reduction, sequential testing, sample ratio checks, warehouse-native analysis and feature flag checks it does not bill for. GrowthBook is open source and warehouse-native, connecting directly to Snowflake, BigQuery and other SQL sources, with per-seat rather than per-request pricing. LaunchDarkly is the enterprise release-management option with guarded rollouts and automatic rollbacks, cloud-only.

For regulated industries, the flag-first tools emphasize governance PostHog does not center on: four-eyes approval on flag changes, audit logs, role-based access and private-cloud or on-premise deployment. Flagsmith and Unleash are open source and self-hostable; ConfigCat and Split (now part of Harness) are hosted options. For marketing-led website testing, VWO and AB Tasty offer visual editors that need no engineering.

In-app guides: Pendo and Userpilot

If the gap is “we shipped it but nobody found it,” you need tooltips, walkthroughs and onboarding checklists your product or success team can build without code. PostHog does not have this. Pendo is the established option, with feature tagging and NPS surveys; Userpilot combines guides with analytics and replay. Both have lighter analytics than Mixpanel or Amplitude.

Website analytics: Plausible, Fathom, Umami, Matomo, OpenPanel, GA4

If you mostly used PostHog for pageviews and sources on a marketing site, a full product suite is more than you need. Plausible, Fathom and Umami are cookieless dashboards with tiny scripts. Matomo is the most complete open-source replacement for Google Analytics, though heatmaps and A/B tests are paid plugins even when self-hosted. OpenPanel adds funnels and retention to web analytics. Pirsch tracks server-side, so client-side blockers do not affect it. GA4 is free and strongest on acquisition and Google Ads, and weak on in-product behavior.

Error monitoring: Sentry

PostHog’s error tracking is tied to replays and user context. Sentry goes deeper on the engineering side with stack traces, source maps, performance profiling and release tracking. Many teams run Sentry next to whatever analytics tool they pick.

Open-source and self-hosted PostHog alternatives

“Open source” covers very different licenses and very different self-hosted editions. Check both before you plan a server.

The license matters if you modify the code: MIT lets you build a closed product on top, while AGPL requires you to publish changes when you offer the modified tool as a network service. Hosting also has a cost that the license does not show: servers, upgrades, backups and someone on call when ingestion falls behind.

Free PostHog alternatives

No single free tool replaces all of PostHog, but each job has a free option. GA4 is the free default for website analytics. Microsoft Clarity gives free session replay and heatmaps. Mixpanel and Amplitude have free plans for product analytics. Statsig and GrowthBook have free tiers for flags and experiments. Umami, Matomo and Plausible Community Edition cost nothing in licenses if you host them, which is not the same as costing nothing. If you are asking “what is the best free web analytics tool,” the honest answer is GA4 for depth and ad integration, or a self-hosted Umami or Matomo if you want to own the data.

PostHog vs Statsig

Statsig is an experimentation platform that also does analytics; PostHog is an analytics suite that also does experiments. Statsig wins on statistics (CUPED, sequential testing, stratified sampling, mutual exclusion layers), warehouse-native deployment, and flag checks that do not count toward the bill. PostHog wins on breadth: heatmaps, surveys, error tracking, logs and LLM observability sit in the same place as your analytics. If flag request volume is what drives your PostHog bill, Statsig or GrowthBook will usually be cheaper for that job.

PostHog vs Plausible

These tools answer different questions. Plausible tells you how many people visited, from where, which pages they read and which goals they completed, without cookies and without recognizing a visitor from one day to the next. PostHog follows identified users across sessions and devices so you can study funnels, retention and individual journeys. Plausible’s script is a small fraction of the size of PostHog’s SDK, and it is easier to self-host. Choose Plausible for a content or marketing site; keep a product analytics tool for an app.

Has Google Analytics been replaced, and is it still free?

Google Analytics was not discontinued; its old version was. According to Google’s Analytics Help, Google Analytics 4 has replaced Universal Analytics, standard Universal Analytics properties stopped processing data on July 1, 2023, and access to Universal Analytics data ended in July 2024. GA4 itself is free; Google Analytics 360 is the paid enterprise version. Some teams that left Universal Analytics went to privacy-first tools like Plausible or Matomo instead of GA4, and some product teams use PostHog or Mixpanel next to GA4 rather than instead of it.

Estimate your volume before you compare prices

Every tool on this list prices on a unit: events, sessions, recordings, pageviews, flag requests, tracked users or seats. Comparison articles quote prices at round volumes, but your bill depends on how your traffic turns into those units. Do this conversion first.

In this illustrative case, the site sits just above PostHog’s 1M free events. Two settings change the answer more than switching vendors would: turning autocapture off or limiting it to an allow list, and capturing events anonymously until a user signs in (PostHog documents that anonymous events cost less to process than identified ones). Run the same arithmetic for replays: if you record every session, 60,000 sessions a month is twelve times a 5,000-recording allowance, while sampling 5% of sessions fits inside it.

Then price that volume on each candidate’s own pricing page or calculator. Replays, flag requests and seats often decide the comparison more than events do.

Where the comparison articles disagree

Reading the top results side by side turns up several contradictions. Here is how to settle them.

• Does Mixpanel have session replay and feature flags? One comparison says Mixpanel has no replay, flags or A/B testing; another says experiments and flags are Enterprise-only. Mixpanel’s current pricing page lists session replays and a limited number of feature flags on its free plan. Older articles describe an older product.
• How many companies use PostHog for free? PostHog’s alternatives post says 90%; its pricing page says 97%. Use the pricing page; the figure changes over time and either way the free tier covers most small teams.
• How much cheaper are anonymous events? One survey says 55 to 80% cheaper; PostHog’s cost guide says up to four times cheaper. These are close but not identical, and the exact rate depends on volume tier, so model it with PostHog’s own calculator.
• Free allowances for other tools. Free tiers for Amplitude, Statsig and others are quoted differently in different articles, sometimes in events and sometimes in tracked users. Treat any number in a comparison as a starting point and confirm it on the vendor’s pricing page the week you decide.

Switching without losing trust in your numbers

The most common failed migration is not technical. The new tool goes live, its numbers differ from PostHog’s, and the team decides one of them is broken. Expect differences and plan for them.

Why two tools never match

• Ad blockers and proxies. A tool loaded from a known analytics domain loses more events than one sent through your own domain or a server.
• Identity. Tools differ in how they merge anonymous visitors with signed-in users, so “users” and “unique visitors” are rarely comparable. Cookieless tools that reset identifiers daily count a returning visitor again on each new day.
• Sessions. Session timeouts and the rules that start a new session vary, which changes session counts and conversion rates.
• Bots and filters. Each vendor filters bots and internal traffic differently.
• Consent. If one tool waits for consent and the other does not, the gap can be large in regions with banners.

A parallel-run checklist

• Run both tools for at least two full weeks, so weekdays and weekends are both covered.
• Map every event you keep to its new name before launch, and write the definitions down.
• Compare trends, not totals: if both tools show the same week-over-week movement, both are working.
• Pick one source of truth per metric and say so in your dashboards.
• Export the PostHog history you need before you turn it off; historical data rarely migrates cleanly into a new data model.
• Remove the old SDK once you are done, so you are not paying in page weight for a tool nobody reads.

If you run a marketing site, measure revenue by channel, not just events

Many PostHog installs sit on a marketing site, where the real question is not “what do users click in the app” but “which channels and pages produce leads and sales.” Product analytics tools, PostHog included, are built for in-product behavior; they will count form submissions, but they will not tell you what organic search or AI assistants are worth to you unless you build that yourself.

For that job, you need three things: conversions defined (form submits, demo requests, calls, purchases), a value on each one, and a report that splits those conversions by channel and landing page. Search engines hide keywords, so the landing page is the level you can measure. AI assistants such as ChatGPT and Perplexity add a newer channel that sometimes passes a referrer and sometimes arrives as Direct.

The method is in our guides to SEO conversion tracking, calculating conversion value and tracking ChatGPT and AI assistant traffic. If you also want to drop cookie banners where the law allows, see cookieless conversion tracking. SEOConversion is built for this narrow job: one script, autocaptured form submits, calls and ecommerce events, a value on each conversion, and reports of conversions and revenue from Google, Bing and AI assistants by landing page. It does not replace product analytics, replay or feature flags.

Frequently asked questions