PostHog

PostHog is the open-source product analytics platform that grew teeth. It started as a Mixpanel alternative you could self-host, then kept adding features until it became a legit all-in-one. Analytics, session replay, feature flags, A/B testing, surveys, and a CDP, all under one roof. The pricing is generous. The product is good.

The London-based team has been one of the loudest open-source success stories of the last few years. They publish their handbook publicly, ship features at a clip, and price aggressively to win against Amplitude and Mixpanel.

If you've been running Mixpanel and feeling the bill creep, or if you wanted Amplitude but couldn't justify it, PostHog is the obvious next look. The free tier alone covers most early-stage startups indefinitely.

What PostHog actually does

PostHog tracks user events, builds funnels and retention curves, records sessions, manages feature flags, runs A/B tests, ships surveys, and pipes data out to your warehouse. That's a lot. The thing that makes it work is one event graph powering everything.

You instrument once, then every product gets the same data. Funnel analysis sees the same events as feature flags. Session replay links to the same users as analytics. The unified data model is what bigger competitors charge a lot more for.

Product analytics

Funnels, retention, paths, trends, correlations. The standard product analytics primitives, executed competently. Insights save to dashboards. Dashboards share via URL. Cohorts sync to other tools.

The interface is busier than Amplitude's but more powerful. Once you know where things live, you can answer most product questions without leaving PostHog.

Session replay and feature flags bundled

Session replay records what users actually saw. It's bundled into the platform. Feature flags ship with targeting, rollouts, and experiments. Both would be separate paid tools elsewhere. PostHog gives them to you in one bill.

Who PostHog is for

Startups, indie hackers, and product teams who want analytics plus session replay plus flags without three separate vendors. Especially good for technical teams who like the open-source ethos and aren't afraid of a slightly busier UI.

It's also a strong pick for teams with privacy or sovereignty requirements. You can self-host PostHog on your own infrastructure. Few competitors offer that.

PostHog pricing

The free tier covers 1 million events, 5,000 session replays, 1 million feature flag requests, and 1,000 surveys per month. That's a lot. Past the free tier, pricing is usage-based per product.

Events cost $0.00005 each past the first million. Session replays cost $0.005 each past 5,000. Feature flag requests cost $0.0001 each past a million. The pricing scales gradually rather than jumping at tier boundaries.

For startups, this works out way cheaper than Amplitude or Mixpanel. For mid-size companies, it's competitive. For massive scale, you'd negotiate a custom contract.

Features worth knowing

Self-hosting

Open-source under the MIT-derived PostHog license. You can run PostHog on your own Kubernetes cluster, your own VMs, or in your VPC. The hosted version is easier, but the option to self-host changes risk profile for some buyers.

Data warehouse integration

PostHog can sync events to and from BigQuery, Snowflake, Redshift, and others. You can do analytics in PostHog or in your warehouse, depending on your team's comfort level.

HogQL

SQL-like query language for ad-hoc analysis. Power users build queries that the standard Insights can't express. Less polished than Amplitude's chart builder, more flexible than anything else.

Experiments and feature flags

Run A/B tests with proper stats, gate features by user attribute, do gradual rollouts. The flag SDK is solid. Tightly integrated with the analytics, so you can see how variants affect downstream events without extra plumbing.

The tradeoffs

PostHog's UI is denser than Amplitude or Mixpanel. There are more knobs. The learning curve is steeper. Don't expect a marketer to spin up a dashboard their first day without help.

The all-in-one bet is real. If you only need analytics, PostHog is overbuilt. If you only need flags, LaunchDarkly is more focused. PostHog wins when you need most or all of the bundle.

PostHog vs alternatives

The honest comparisons are PostHog vs Amplitude, PostHog vs Mixpanel, and PostHog vs LaunchDarkly. Amplitude is more polished. Mixpanel is more focused. LaunchDarkly is more enterprise on flags.

For best value across the bundle, PostHog wins. See PostHog alternatives or browse the best analytics tools.

Bottom line on PostHog

PostHog is the modern open-source pick for product analytics and adjacent tooling. It's not the prettiest. It's the most generous, the most open, and arguably the most ambitious.

If you're starting a product today, PostHog covers analytics, replay, flags, and experiments under one bill that's free until you have real traffic. That's a significantly better deal than what most teams default to.

Common PostHog questions

Is PostHog open source actually open? Yes, with an MIT-derived license. You can self-host the entire platform. Some enterprise features require a license, but the core analytics, replay, and flags are open.

Should I self-host PostHog? Only if you have ops capacity or hard data residency requirements. PostHog is non-trivial to operate (Postgres, ClickHouse, Kafka, Redis). Hosted PostHog is cheaper for most teams when you account for engineering time.

How does PostHog compare to Amplitude on usability? Amplitude is more polished. PostHog has more features but a steeper learning curve. For technical teams, PostHog wins. For marketers and PMs without engineering support, Amplitude is friendlier.

What about session replay quality?

Solid. Recording fidelity is good. Privacy controls (mask sensitive fields) are robust. Replays are linked to events so you can jump from a funnel drop to the actual session. Few competitors integrate this tightly.

Are feature flags fast enough for real apps?

Yes. Flag evaluation runs locally in the SDK after fetching configurations. Latency is sub-millisecond. The bootstrap pattern lets you ship initial flag values with your HTML to avoid flicker.

Workflow tips for PostHog

Standardize event names. Like Amplitude, PostHog rewards consistent taxonomy. Document your tracking plan. Use the Data Management tools to spot inconsistencies.

Bundle replay with feature flag rollouts. Roll a flag to 5%, watch the recordings of those users. You catch UX issues before scaling the rollout.

Use HogQL for complex questions. The visual builders cover 80% of needs. SQL covers the rest. Don't fight the tool; jump to HogQL when needed.

Self-hosted is easier on Kubernetes than docker-compose. The Helm chart is the supported path. Plan for ClickHouse storage growth and Postgres backups. Browse tools for product managers for related picks.

Real-world PostHog scenarios

An indie SaaS team runs everything on PostHog: analytics, session replay, feature flags, and A/B tests. Free tier covers their volume. They ship a feature, flag it, watch replays of beta users, decide on full rollout. One tool, one bill, one mental model.

A growth team uses PostHog session replay alongside Amplitude analytics. Amplitude handles the deep analysis. PostHog replays show them what users actually did. The combo is more powerful than either alone, though more expensive.

A regulated startup self-hosts PostHog on their VPC. Data residency requirement met. The product team gets analytics. The legal team gets sovereignty. The ops team gets a real ClickHouse cluster to babysit. Tradeoffs.

Operational tips

Use the bootstrap pattern for feature flags on first page load. Avoids flicker. The SDK supports it. Skip this and users see UI flash.

Tag releases. PostHog correlates errors and metric drops to releases when you do. Without release tags, debugging deploys is much harder.

Sample session replays. Recording every session is expensive and noisy. Sample 10% or 20%. Save more recordings on rare events (errors, key conversions).

For startups, PostHog is one of the rare all-in-one tools that delivers on the all-in-one promise. The pricing model rewards using more of the platform. Browse the PostHog page for community reviews.

Why PostHog keeps winning

Open source plus generous free tier plus all-in-one bundle is a hard combination to beat. PostHog launched with that thesis and executed on it for years. The result is a product that's grown faster than most product analytics competitors despite charging less.

The bundle matters. Buying analytics, session replay, feature flags, and experiments separately costs a fortune. PostHog packages them under one bill. The integration between them is real, not just bundled marketing. A funnel insight links to relevant replays. A feature flag rollout shows up in analytics.

The open source angle changes the buyer dynamic. Engineers can prototype on self-hosted, validate the product, and recommend it internally. By the time procurement gets involved, the team is already using PostHog. That bottom-up adoption beats sales-led competitors.

For teams in 2026 picking product analytics, PostHog is one of the few products where the price gets cheaper as you use more of the platform. That's an unusual property in SaaS, and it explains why PostHog keeps showing up on best-of lists.

PostHog's competitive moat

The PostHog moat is the bundle plus the open source story. Competitors can match individual features. Few can match the integrated experience across analytics, replay, flags, and experiments without dramatically higher prices. The bundle's value compounds the more features a team uses.

The open source angle protects against vendor lock-in concerns. Teams know they can self-host if hosted prices climb. That option, even if rarely exercised, changes the negotiation dynamic and limits how much hosted PostHog can charge. Pricing stays competitive because the alternative is real.

For 2026 and beyond, PostHog continues to ship features at a clip. The team is well-funded and motivated. The community contributions add up. For startups picking analytics today, PostHog is increasingly the obvious starting point. Outgrow it later if needed; benefit from the generous free tier now.