PostHog

This is best used for product teams needing comprehensive analytics without vendor lock-in or data privacy concerns. PostHog consolidates 10+ tools into one platform—product analytics, session replay, feature flags, A/B testing, surveys, error tracking, and data warehouse—eliminating the need to integrate Mixpanel, Hotjar, LaunchDarkly, and others separately.

The platform's MIT-licensed open source core means you own your data completely. Self-host for full control or use PostHog Cloud (US/EU) for managed infrastructure. This matters for teams uncomfortable sending user behavior data to third-party SaaS tools or facing GDPR/compliance requirements. 90%+ of companies use PostHog completely free thanks to generous monthly allowances: 1M events, 5K session replays, 1M feature flag requests, 100K error exceptions, 1,500 survey responses—renewing every month.

Pricing: Usage-based with volume discounts. Product analytics starts at $0.00031/event after free tier. Session replay at $0.005/recording. Feature flags at $0.0001/request. Billing limits prevent surprise charges. Typical costs: $0-100/month (startups), $500-2,000/month (scale-ups), $5,000+/month (enterprises). Platform add-ons: HIPAA compliance ($250/mo), advanced permissions ($450/mo), enterprise features ($2,000/mo).

PostHog vs alternatives: Cheaper and more transparent than Amplitude/Mixpanel. More features than Plausible/Fathom. Developer-friendly unlike Heap. Open source unlike everything except Matomo. Y Combinator uses PostHog internally and recommends it to batch companies—65% adoption rate proves product-market fit.

Integration: SDKs for JavaScript, React, Node, Python, PHP, Ruby, Go, iOS, Android, Flutter. Auto-capture tracks clicks/pageviews without manual instrumentation. SQL access for custom analysis. Data warehouse connects Stripe, HubSpot, Salesforce, Zendesk for unified analytics.

Team fit: Product engineers, growth teams, and startups prioritizing speed and data ownership. Ideal for technical teams comfortable with self-service tools. Less suitable for non-technical marketers needing hand-holding (use Google Analytics 4 instead).

Watch-outs: Usage-based pricing can spike unexpectedly if auto-capture tracks too many events—set billing limits. Self-hosting requires 4 vCPU, 16GB RAM minimum. Free tier generous but costs scale with volume. Complex implementations need thoughtful event taxonomy planning.

Customers: Y Combinator (internal tool), SpaceX, Staples, Grafana, 65% of Y Combinator batch companies. Launched on Hacker News 2020 as "most successful B2B software launch since 2012."