# PostHog

> PostHog is an open-source product and web analytics platform that bundles event analytics, session replay, feature flags, experiments, surveys, error tracking, and a data warehouse into one usage-priced suite, aimed at engineering-led teams that would rather run one tool with generous free tiers than stitch together five point solutions.

- Category: GTM Analytics (https://saastracker.org/categories/gtm-analytics)
- Website: https://posthog.com
- Starting price: $0 (generous monthly free tiers; pay only past the allowance)
- Free plan: 1M analytics events, 5K session recordings, 1M feature-flag requests, 100K exceptions, 1,500 survey responses, and 1M data warehouse rows per month, on 1 project with 1-year retention.
- Free trial: Not applicable; the free tier is permanent, not a trial
- Founded: 2020, HQ: Remote-first; US-incorporated (San Francisco), team distributed globally, Ownership: Venture-backed
- Profile last reviewed: 2026-08-22
- Canonical profile: https://saastracker.org/products/posthog

## Overview

PostHog started in 2020 as an open-source alternative to product analytics incumbents and has since turned into something closer to a developer-tools conglomerate: the company now ships 10+ paid products off one event pipeline, including product analytics, a simplified web analytics dashboard, session replay, feature flags, A/B experiments, surveys, error tracking, data pipelines, and a queryable data warehouse. Everything shares the same underlying events, so a spike on the web analytics dashboard can be chased into funnels, then into individual session recordings, without exporting anything.

The commercial model is the other defining trait. Every product has a monthly free allowance (1 million analytics events, 5,000 session recordings, 1 million feature-flag requests, 1,500 survey responses, and so on), and usage beyond that is billed per unit at rates that fall steeply with volume. There are no seat fees on the core products; over 190,000 companies have signed up, a large share of them paying nothing, which PostHog treats as deliberate strategy rather than leakage.

For a GTM audience, the relevant framing is that PostHog covers the measurement layer from marketing site to product usage, but it is not a revenue-attribution tool: it will tell you which channels drive signups and what those users then do, and it can join CRM-ish data in its warehouse, but pipeline and closed-won attribution against Salesforce or HubSpot objects is a build-it-yourself exercise, not a packaged report the way Dreamdata or HockeyStack package it. Buyers coming from Google Analytics get a far simpler dashboard; buyers coming from Amplitude or Mixpanel get comparable product analytics with more surface area and a cheaper entry.

## How it works

1. You install a JavaScript snippet or one of the server/mobile SDKs, and PostHog begins capturing events, either through autocapture (clicks, pageviews, form submits recorded without manual instrumentation) or through explicit custom events. The same stream feeds every product: the web analytics dashboard renders visitors, pageviews, sessions, bounce rate, referrers, UTMs, and Core Web Vitals out of the box, while product analytics exposes funnels, trends, retention, paths, and SQL-style querying for deeper questions.

2. Session replay records real user sessions (with masking controls for sensitive fields), and because recordings share identity with events, you can jump from a failed funnel step to watching the sessions that failed it. Feature flags and experiments run off the same identity layer, so a team can gate a rollout to a cohort it built in analytics and measure the experiment without a separate testing tool.

3. The data warehouse side ingests external sources (Stripe, HubSpot, and similar) so revenue and CRM records can be queried next to behavioral events; data pipelines push events onward to destinations like BigQuery, Snowflake, or S3. Teams choose US or EU (Frankfurt) cloud at signup, and an open-source Hobby build can be self-hosted with Docker for those who insist, though PostHog itself steers production workloads to its cloud.

4. Billing is metered per product each month: you set spending limits per product, burn through the free allowance first, then pay per event, recording, flag request, or row. A paid account (any spend) also upgrades the platform itself: six projects instead of one, seven-year data retention instead of one, and email support.

## Best for

Engineering-led SaaS and product teams that want web analytics, product analytics, replay, flags, and experimentation consolidated in one usage-priced platform, and that are comfortable doing their own revenue-attribution modeling if they need it.

## Not the right fit for

- Marketing teams that want packaged B2B revenue attribution against CRM pipeline; PostHog stores the data but does not ship Dreamdata-style attribution models or account-journey reporting out of the box.
- Non-technical teams wanting a set-and-forget dashboard; PostHog's depth assumes someone on the team is willing to define events, insights, and queries.
- Privacy-minimalist site owners who only need pageview counts; a one-product tool like Plausible or Fathom is simpler, lighter, and easier to defend in a GDPR review than a full event platform.
- Organizations that need contractual on-premise deployment with vendor support; the self-hosted Hobby build is explicitly not positioned for production scale, and PostHog pushes cloud for anything serious.
- Teams with unpredictable traffic and no billing discipline; usage pricing is cheap per unit but needs spending limits configured, or a traffic spike becomes an invoice surprise.

## Features

### Web and product analytics

One event stream feeding both a simple marketing-site dashboard and full product analytics.

- **Web analytics dashboard**: A pre-built, GA-style view of visitors, pageviews, sessions, bounce rate, referrers, UTMs, and Core Web Vitals, designed as an escape hatch from GA4's complexity.
- **Autocapture**: Clicks, pageviews, and form interactions are captured without manual instrumentation, so historical questions can be answered about events nobody thought to track.
- **Funnels, trends, retention, paths**: The standard product-analytics toolkit for conversion and engagement analysis, with breakdowns by person, cohort, or group properties.
- **SQL access to events**: A SQL dialect over the event store lets analysts write arbitrary queries when the visual insight builder runs out of road.
- **Group and B2B analytics**: Events can be aggregated at company or account level rather than individual users, the building block for B2B activation and usage reporting.
- **Cookie-less mode**: An optional configuration tracks without cookies for teams that want to reduce consent-banner scope, at some cost to cross-session identity.

### Session replay and debugging

Qualitative evidence attached to the same identities as the quantitative data.

- **Session recordings**: Watch real sessions with console logs and network context; 5,000 recordings per month are free, then $0.005 per recording with volume discounts.
- **Replay linked to insights**: Jump from a funnel drop-off or an error straight to the recordings of affected sessions, because replay and analytics share one identity graph.
- **Mobile replay**: Session recording for mobile apps, billed separately per recording.
- **Error tracking**: Exception capture and grouping with a 100,000-exceptions-per-month free tier, overlapping a chunk of what teams pay Sentry for.

### Feature flags, experiments, surveys

The shipping-and-learning layer most analytics tools leave to third parties.

- **Feature flags**: Boolean and multivariate flags with percentage rollouts and cohort targeting; 1 million flag requests per month free, then $0.0001 per request.
- **A/B experiments**: Experimentation built on the flag infrastructure with significance reporting, billed through flag requests rather than as a separate SKU.
- **Surveys**: In-app surveys (NPS, PMF, custom) targeted by cohort or behavior; 1,500 responses per month free, then per-response pricing.
- **Toolbar**: A browser overlay for inspecting live pages, viewing click data in place, and toggling flags without opening the app.

### Data platform

Warehouse, pipelines, and integrations that make PostHog the system of record rather than a silo.

- **Data warehouse**: Query external data (Stripe, HubSpot, and other sources) alongside product events; 1 million rows per month free, then fractions of a cent per row.
- **Data pipelines (CDP)**: Transform events and forward them to destinations such as BigQuery, Snowflake, S3, and webhook targets, with batch exports for warehouse syncs.
- **LLM and AI observability**: Tracing and cost analytics for LLM-powered features, with a 100,000-events-per-month free allowance, a newer product line aimed at AI app builders.
- **API and SDKs**: Client and server SDKs across major languages plus a full REST API for capture, query, and administration.
- **Open-source core**: The codebase is public on GitHub and a Hobby build can be self-hosted with Docker; the company recommends its cloud for production scale.

### Platform, pricing, and controls

The account-level mechanics that make the usage model livable.

- **Per-product free tiers**: Every product has a monthly free allowance (1M events, 5K recordings, 1M flag requests, 100K exceptions, 1,500 survey responses, 1M warehouse rows) that resets monthly, forever.
- **Billing limits**: Hard spending caps can be set per product so a traffic spike degrades collection rather than producing a runaway invoice.
- **US or EU cloud**: Data residency choice at signup between US hosting and an EU cloud in Frankfurt, relevant for GDPR-sensitive deployments.
- **Platform packages**: Paid platform add-on packages (Boost, Scale, Enterprise) layer on advanced support and compliance options such as HIPAA BAAs and SSO on top of usage pricing.
- **Projects and retention upgrades**: Any paid usage upgrades the account from 1 project and 1-year retention to 6 projects and 7-year retention, plus email support.

## Use cases

- **Growth engineer at a product-led SaaS startup**: Marketing reports traffic in GA4, product reports usage in a separate analytics tool, and nobody can connect a channel to activation, so channel budget arguments run on opinion. Outcome: One PostHog snippet feeds the web analytics dashboard and product funnels from the same events, so the team can follow a UTM-tagged visitor from landing page through signup to activation and settle the channel argument with a single funnel.
- **Founding team pre-revenue, allergic to SaaS sprawl**: Needs analytics, replay, flags, and an experimentation tool but cannot justify four subscriptions before product-market fit. Outcome: PostHog's free allowances (1M events, 5K recordings, 1M flag requests monthly) cover all four jobs at typical early-stage volume for $0, and usage pricing kicks in only as traffic grows.
- **Product engineer debugging a conversion regression**: Checkout conversion dropped 15% after a release and aggregate charts don't show why. Outcome: The funnel isolates the failing step, linked session replays show a payment widget erroring on mobile Safari, and error tracking confirms the exception spike, one tool, one identity graph, no data export.
- **Data-privacy-conscious EU SaaS team**: Wants full product analytics but legal requires EU data residency and wants the option to audit the code that processes user data. Outcome: EU cloud hosting in Frankfurt plus an open-source codebase gives legal both residency and inspectability, without downgrading to a pageview-only privacy tool.

## Pricing

Usage-based pricing per product: every product has a monthly free allowance, then per-unit billing (per event, recording, flag request, survey response, or row) with steep volume discounts; optional platform packages add support and compliance features.

- **Free (no card)**: $0 per month. All products with monthly free allowances; 1 project, 1-year data retention; Community support. The allowances reset monthly and are large enough that many small products never pay.
- **Pay-as-you-go**: Usage-based per month, metered per product. Analytics events from $0.00005/event, falling to ~$0.000009 at 250M+; Session replay from $0.005/recording; feature flags from $0.0001/request; Surveys from $0.10/response; warehouse from $0.000015/row; 6 projects, 7-year retention, email support once paying. Per-product spending limits are configurable and worth setting on day one.
- **Platform packages (Boost, Scale, Enterprise)**: Quoted / package pricing added on top of usage. Advanced support and onboarding; Compliance options including HIPAA BAA; SSO and enterprise controls on the top package. Usage rates stay the same; the packages buy platform capabilities, not cheaper units.

Billing notes:

- Pricing is metered monthly per product; there is no seat pricing on core products, so large teams pay the same as small ones at equal volume.
- Volume discounts are automatic and aggressive (analytics events drop roughly 80%+ per unit at the highest published tiers), which rewards consolidation onto one tool.
- Free-tier allowances apply every month even on paid accounts: the first 1M events are free before metering starts.
- A paid account upgrades platform limits (6 projects, 7-year retention); heavy compliance needs (HIPAA BAA) require the Boost, Scale, or Enterprise package.
- Costs are workload-sensitive: session replay and data pipelines can dominate the bill at high traffic, so spending caps per product are the practical guardrail.

Value assessment: At small and mid scale PostHog is close to unbeatable on price: the permanent free tiers cover a real startup's entire measurement stack, and the per-unit rates undercut buying analytics, replay, flags, and experimentation separately. The honest caveat is that usage pricing shifts the budgeting burden onto you: a replay-heavy or event-noisy implementation can quietly cost more than a flat-rate point tool, and the platform packages needed for enterprise compliance are quoted, not listed. Treat it as extremely cheap by default and only as cheap as your instrumentation discipline at scale.

## Strengths

- Breadth with real integration: analytics, replay, flags, experiments, surveys, error tracking, and a warehouse genuinely share one event stream and identity graph rather than being bolted-on acquisitions.
- The most generous free tier in the category, 1M events and 5K recordings monthly, permanently, which makes evaluation and early-stage use genuinely free.
- Open-source codebase plus a choice of US or EU (Frankfurt) cloud gives privacy and procurement teams inspectability and residency options most rivals lack.
- Usage pricing with automatic volume discounts and per-product spending caps scales from hobby project to hundreds of millions of events without a sales call.
- Web analytics dashboard is a credible, simpler GA4 replacement that comes free with the rest of the platform rather than as a separate purchase.
- Strong engineering brand and documentation; the product assumes and rewards technical users, and 190,000+ signed-up companies suggest the wedge works.

## Limitations

- No packaged B2B revenue attribution: connecting spend and touchpoints to CRM pipeline and closed-won revenue is a data-warehouse project in PostHog, not a built-in report as in Dreamdata or HockeyStack.
- Depth demands technical investment; non-technical marketers can read the web dashboard but will struggle to self-serve funnels, SQL insights, or warehouse joins.
- Usage-based billing is unpredictable without configured limits, and replay-heavy or autocapture-noisy sites can generate surprising invoices.
- The full event platform carries more GDPR surface than minimalist tools: consent, masking, and retention need deliberate configuration, where Plausible or Fathom are compliant nearly by default.
- Self-hosting is nominally possible but the supported Hobby build is explicitly not for production scale, so the open-source angle is more inspectability than a true on-prem offering.
- Product surface expands fast (10+ products, AI features, workflows), and the pace means some newer modules are visibly less mature than the analytics core.

## Comparisons

- **PostHog vs Plausible Analytics**: Plausible is a single lightweight dashboard for marketing-site traffic; PostHog is a full event platform where web analytics is one module of ten. Choose Plausible when you want cookie-free simplicity, a sub-1KB script, and nothing to maintain; choose PostHog when the website is the top of a product funnel you intend to analyze end to end.
- **PostHog vs Fathom Analytics**: Fathom and PostHog barely overlap beyond the traffic dashboard: Fathom sells privacy-first simplicity at a flat $15/month, PostHog sells depth on usage pricing that starts free. A content site or agency should buy Fathom; a product team that will ever want funnels, replay, or experiments should start on PostHog's free tier and skip running both.
- **PostHog vs HockeyStack**: HockeyStack starts where PostHog stops: it ingests CRM, ad, and marketing-automation data to attribute pipeline and revenue for B2B GTM teams, at quote-only pricing reportedly starting around four figures monthly. PostHog is the better system of record for product and web behavior at a fraction of the cost; teams with real attribution budgets sometimes run both, PostHog for product data, HockeyStack for the board's pipeline questions.

## Implementation

- Setup time: Minutes to first data with the JS snippet and autocapture; days to weeks to define clean custom events, dashboards, and cohorts; warehouse sources and pipelines are a separate project.
- Learning curve: Low for the web analytics dashboard; moderate for funnels, cohorts, and replay; high for SQL insights, warehouse modeling, and experimentation statistics.
- Onboarding: Self-serve by default with docs and community; paid platform packages add support tiers, and there is no mandatory sales process below enterprise needs.
- Migration: Coming from GA4 or Mixpanel-style tools means re-instrumenting events (autocapture softens this) and accepting that historical data largely stays behind; exporting out of PostHog later is easier than most rivals thanks to batch exports and the API.

## Platform, API & security

- Platforms: Web app, JavaScript snippet and web SDK, Server SDKs (Python, Node, Go, PHP, Ruby, and others), Mobile SDKs (iOS, Android, React Native, Flutter), Self-hosted Hobby build (Docker, open source)
- API: Full REST API for event capture, querying, and account administration, plus SDKs across major languages; batch exports push data to warehouses on schedule.
- Compliance: SOC 2 Type II, GDPR, HIPAA (BAA on Boost, Scale, or Enterprise package)
- Data residency: Customer's choice of US cloud or EU cloud hosted in Frankfurt, selected at signup.
- SSO: SSO/SAML available on upper platform packages; standard login for self-serve accounts.
- Security notes: Open-source codebase allows independent inspection; session replay ships masking controls for sensitive fields, and a public trust center documents audits and controls.

## Support

- Channels: Community forum and GitHub, Email support (paying customers), Dedicated support on platform packages
- Documentation: Extensive public docs, tutorials, and an unusually candid company handbook covering product and pricing decisions.
- Community: Active GitHub (open-source repo) and community Q&A; PostHog claims roughly 65% of each Y Combinator batch uses the product.

## Company

- Founded: 2020
- Founders: James Hawkins (co-CEO), Tim Glaser (co-CEO)
- Headquarters: Remote-first; US-incorporated (San Francisco), team distributed globally
- Ownership: Venture-backed
- Employees: ~200-211 (mid-2026, per PitchBook and Owler)
- Funding: ~$194M raised, including a $70M Series D (June 2025, led by Stripe, $920M valuation) and a $75M Series E (October 2025, led by Peak XV Partners, $1.4B valuation).

Funding history:

- Seed (2020): $3M. Y Combinator W20 batch.
- Series A (2020): $9M. Led by GV with YC Continuity.
- Series B (2021): $15M. YC Continuity and GV.
- Series C (2025): ~$10M primary. Announced alongside the Series D; structured partly for employee liquidity.
- Series D (2025): $70M. Led by Stripe at a $920M valuation.
- Series E (2025): $75M. Led by Peak XV Partners at a $1.4B valuation.

Timeline:

- 2020: Founded by James Hawkins and Tim Glaser in Y Combinator's W20 batch; MVP shipped to Hacker News four weeks after starting, followed by a $3M seed and $9M Series A.
- 2021: $15M Series B led by YC Continuity and GV as the open-source product analytics wedge takes hold.
- 2023: Platform strategy solidifies around multiple products (replay, flags, experiments, surveys) sharing one event pipeline and usage-based pricing.
- 2025: Raises a $70M Series D led by Stripe at a $920M valuation (June), then a $75M Series E led by Peak XV at $1.4B (October); 176,000+ companies signed up.
- 2026: Roughly 200 employees, 190,000+ signed-up companies, and 10+ paid products including error tracking, LLM observability, and workflows.

## Integrations

Stripe (data warehouse source), HubSpot (data warehouse source), BigQuery, Snowflake, S3 (batch exports and pipelines), Slack and webhook destinations, Segment (event source), Zapier, SDKs for all major web, server, and mobile stacks

## FAQ

### What is PostHog?

PostHog is an open-source analytics platform that combines product analytics, a web analytics dashboard, session replay, feature flags, experiments, surveys, error tracking, and a data warehouse in one tool. Everything runs off a single event stream, and each product is priced on usage with a permanent monthly free allowance.

### How much does PostHog cost?

There is no base subscription: each product has a free monthly allowance (1M analytics events, 5K session recordings, 1M feature-flag requests, 1,500 survey responses, 1M warehouse rows) and usage beyond that is billed per unit, for example $0.00005 per analytics event falling to around $0.000009 at very high volume. Optional platform packages (Boost, Scale, Enterprise) add support and compliance features on top.

### Is PostHog really free for small teams?

For most early-stage products, yes. The free allowances reset every month and are high enough that a site with modest traffic can run analytics, replay, and feature flags indefinitely at $0, on one project with 1-year retention. Paying anything upgrades you to 6 projects, 7-year retention, and email support.

### Is PostHog a good Google Analytics replacement?

Its web analytics product is a deliberate GA4 alternative: a pre-built dashboard of visitors, pageviews, sessions, bounce rate, referrers, UTMs, and Core Web Vitals without GA4's report maze. The difference is that the same events also power product analytics and session replay, so it suits teams that want to go deeper than traffic reporting; if you only want traffic reporting, a lighter tool like Plausible or Fathom is simpler.

### Does PostHog do B2B revenue attribution?

Not as a packaged product. PostHog can capture the touchpoints and, via its data warehouse, sit CRM and payment data next to behavioral events, but building pipeline attribution models is your analysts' job. Teams that want turnkey account-level attribution against Salesforce or HubSpot pipeline should look at Dreamdata or HockeyStack, possibly alongside PostHog.

### Can I self-host PostHog?

There is an open-source Hobby build deployable with Docker, and the full codebase is public on GitHub. PostHog itself recommends its managed cloud for production workloads and does not position self-hosting for scale, so treat self-hosting as an option for small or exploratory deployments and the open code as an auditability benefit.

### Where is PostHog data stored, and is it GDPR compliant?

You choose at signup between US cloud and EU cloud hosted in Frankfurt. PostHog is SOC 2 Type II audited and documents its GDPR posture, including cookie-less tracking options and replay masking; HIPAA use requires a BAA, which is available on the Boost, Scale, or Enterprise platform packages.

### How does PostHog compare to Amplitude or Mixpanel?

Feature-for-feature on product analytics they are comparable for most teams' needs, but PostHog bundles session replay, flags, experiments, surveys, and a warehouse that the incumbents either lack or charge separately for, and its usage pricing starts at zero without a sales conversation. The incumbents retain an edge in some enterprise analytics depth and services; PostHog's pitch is consolidation and price.

### What are PostHog's free tier limits?

Per month: 1M product analytics events, 5K session recordings, 1M feature-flag requests, 100K error-tracking exceptions, 1,500 survey responses, 1M data warehouse rows, and 10K data pipeline events, among others. The allowances apply every month even after you start paying.

### Who owns PostHog and how is it funded?

PostHog is a venture-backed company founded in 2020 by James Hawkins and Tim Glaser through Y Combinator. It has raised roughly $194M, most recently a $70M Series D led by Stripe (June 2025, $920M valuation) and a $75M Series E led by Peak XV Partners (October 2025, $1.4B valuation), and employs around 200 people.

## Editorial verdict

PostHog is the default recommendation in this category for any team with an engineer on it: no rival matches the combination of a free-forever tier that covers real workloads, usage pricing that scales without a sales call, and genuinely integrated replay, flags, and experimentation on top of web and product analytics. The two honest reservations are that marketing teams without technical support will use a fraction of it, and that B2B pipeline attribution, the question GTM leaders most want answered, remains a do-it-yourself exercise on PostHog's warehouse rather than a shipped feature. Buy it as the behavioral system of record; budget separately if you need turnkey revenue attribution.

## SaaSTracker awards

- Category Leader (GTM Analytics, Summer 2026): "An open-source, usage-priced suite that keeps absorbing adjacent tools until the answer to most analytics questions is already installed."

---

Source: SaaSTracker (https://saastracker.org), an independent editorial project. This profile is compiled from public information, carries no peer reviews or paid placement, and was last reviewed 2026-08-22. Awards are judged on published criteria: https://saastracker.org/methodology