What is PostHog?

PostHog is an open-source product analytics platform designed for engineering and product teams that want behavioral data without relying on third-party SaaS. Founded in 2020 and backed by Y Combinator, PostHog consolidates product analytics, session replay, feature flags, A/B experimentation, surveys, error tracking, and a data warehouse into a single platform with one SDK. Unlike Mixpanel or Amplitude, which target non-technical product managers, PostHog assumes the user is comfortable with SQL, event schemas, and developer tooling. As of 2026, PostHog has surpassed 10,000 company installs, and the posthog-js npm package logs over 1.3 million weekly downloads.

Key Takeaways

• Open-source, all-in-one analytics platform replacing five or six separate tools with a single SDK
• Usage-based pricing with no per-seat charges and a generous free tier (1M events, 5K recordings, 1M feature flag requests per month)
• Self-hosting via Docker Compose keeps behavioral data on your own infrastructure for GDPR and HIPAA compliance
• New LLM Analytics module (2026) tracks prompts, completions, token counts, and latency for AI-powered products
• Developer-first positioning with growing demand in fractional engineering roles at startups and growth-stage companies

Core Features and What Sets PostHog Apart

PostHog's value proposition is consolidation. Product Analytics tracks events, funnels, retention curves, cohorts, and path analysis with both autocapture and explicit event tracking, plus the ability to query raw data via SQL directly in the UI. Session Replay records anonymized user sessions so teams can trace drop-off points visually without bolting on FullStory or Hotjar. Feature Flags and Experimentation let teams roll out changes to targeted segments and measure statistical impact, eliminating LaunchDarkly for most use cases. Surveys collect in-product feedback with up to 1,500 free responses per month. The Data Warehouse and CDP connects to external sources and routes event data to 60+ destinations including BigQuery, Snowflake, CRMs, and marketing tools.

The all-in-one architecture means events, flags, and replays share the same data model, which reduces data drift and makes debugging significantly faster than stitching together point solutions.

PostHog Pricing (2026)

PostHog uses a purely usage-based pricing model with no per-seat charges. Each product module is billed independently, and the free tier is generous: 1 million analytics events, 5,000 session recordings, and 1 million feature flag requests per month at no cost. Beyond free limits, analytics events cost roughly $0.00005 each for the first tier (1-2 million events) and drop to around $0.000009 per event at very high volumes (250 million+). There is no opaque enterprise tier with negotiated pricing — all rates are published and calculator-friendly. Self-hosting eliminates cloud costs but shifts the burden to infrastructure and engineering time. At scale, PostHog is significantly cheaper per event than Amplitude or Mixpanel, which makes it especially attractive for high-traffic products where per-user or per-event pricing compounds fast.

PostHog vs Mixpanel vs Amplitude

Mixpanel is PostHog's most direct competitor for event-based analytics but is built for non-technical users with a polished point-and-click interface. Mixpanel has narrowed its focus over time, deprecating A/B testing to concentrate on analytics alone, which makes PostHog the broader platform for teams that want experimentation alongside analytics.

Amplitude excels at visualizing complex user journeys and is built for enterprise scale, but charges by monthly tracked users, which gets expensive quickly. Amplitude lacks autocapture and session replay natively, requiring additional tooling to match PostHog's feature surface.

Heap pioneered autocapture analytics and is the closest functional analog to PostHog's event model. PostHog is often described as the open-source alternative to Heap, replicating autocapture while adding feature flags, experimentation, and self-hosting options that Heap does not offer.

The Organic Growth Engine Behind PostHog's Adoption

PostHog attributes 97% of its growth to organic channels, which is unusually high for a B2B SaaS product and reflects how deeply the tool is embedded in developer communities. Growth is peer-to-peer rather than sales-led: engineers adopt PostHog on a project, then bring it to their next company or recommend it to their network. The decision to open-source the core product was originally a distribution play, but it created a secondary effect that proprietary competitors struggle to replicate. Companies in regulated industries use PostHog's self-hosted option as their compliance path for GDPR and HIPAA, a segment that cloud-only analytics platforms simply cannot serve at the same price point. This organic flywheel matters for hiring managers because it means PostHog experience is spreading bottom-up through the engineering talent pool, especially at startups between Series A and Series C.

LLM Analytics and What's New in 2026

PostHog's most significant 2026 development is the LLM Analytics product, which lets teams building AI-powered applications track prompt/completion pairs, model usage, token consumption, and latency. With 100,000 free events per month, it fills a gap that no other major analytics platform had addressed for AI-native products. This is an early-mover position in a category without a dominant player, and it could become a significant wedge as AI product instrumentation becomes standard practice.

Beyond LLM Analytics, PostHog has increased free-tier limits across multiple products — surveys went from 250 to 1,500 free responses per month — signaling a strategy of broadening the free tier to drive bottom-up adoption before upselling on volume. The company also formally ended support for new Kubernetes/Helm self-hosted installations, consolidating self-hosting around Docker Compose and pushing larger deployments toward the managed cloud.

Limitations Worth Knowing

Self-hosting PostHog is more operationally demanding than the documentation suggests. Real users report that the recommended minimum specs (8GB RAM, 2 CPU cores) are insufficient for even moderate traffic before the instance becomes unresponsive. The community edition also lags behind the cloud version in feature parity, so teams that choose self-hosting to save money sometimes find themselves missing features only available on the managed tier.

Query performance for large event volumes can be slow without tuning. PostHog uses ClickHouse under the hood, and complex funnel or path queries over hundreds of millions of events require careful index management. The all-in-one pitch also means individual modules are less polished than dedicated point solutions — its A/B testing is functional but less statistically rigorous than Optimizely, and session replay is solid but lacks the UX research depth of FullStory. For non-technical stakeholders accustomed to Mixpanel's or Amplitude's simplified query builders, PostHog's interface feels more raw, and building dashboards for business audiences often requires engineering involvement.

PostHog in the Fractional Talent Market

PostHog proficiency is showing up with increasing frequency in job postings for senior product engineers, growth engineers, and data engineers at developer-tool and AI-native startups. The skill is rarely listed as a hard requirement but commonly appears in "our current stack" sections, signaling that companies prefer to hire people already familiar with its data model and feature flag API. Demand is concentrated at companies between 10 and 200 employees — enterprise postings still overwhelmingly reference Amplitude or Mixpanel.

On Pangea, we see PostHog experience surfacing alongside Next.js, Vercel, and Supabase as part of the modern startup stack. Familiarity with PostHog increasingly signals a broader developer-first product sensibility, making it a useful credential for engineers moving into product or growth roles. For hiring managers, a fractional engineer who already knows PostHog can set up event tracking, configure feature flags, and build analytics dashboards without a lengthy onboarding period.

The Bottom Line

PostHog has established itself as the go-to analytics platform for engineering-led teams that want product analytics, session replay, feature flags, and experimentation without managing a fragmented stack of point solutions. Its open-source foundation, usage-based pricing, and self-hosting option make it especially attractive for startups and compliance-conscious companies. The new LLM Analytics module positions PostHog well for the growing wave of AI product instrumentation. For companies hiring through Pangea, PostHog experience signals an engineer who thinks in product terms and can own the analytics layer end to end.