PostHog vs Plausible: Which Should You Self-Host?
Quick Verdict
These tools solve different problems. Plausible is a lightweight Google Analytics replacement for websites — simple dashboards, privacy-first, minimal resources. PostHog is a full product analytics platform with session replays, feature flags, and A/B testing for software teams. Most self-hosters want Plausible. SaaS builders want PostHog.
Overview
Plausible (launched 2019) is a privacy-focused web analytics tool built on Elixir and ClickHouse. It tracks pageviews, referrers, and top pages without cookies. The entire dashboard fits on one screen. The Community Edition is self-hostable under AGPL-3.0.
PostHog (launched 2020) is an open-source product analytics platform. It combines event tracking, session recordings, feature flags, A/B testing, user surveys, and a data warehouse. It runs on ClickHouse, Kafka/Redpanda, PostgreSQL, Redis, and ~20 Docker containers. Licensed under MIT for self-hosted.
Both use ClickHouse for data storage, but that’s where the similarity ends.
Feature Comparison
| Feature | PostHog | Plausible |
|---|---|---|
| License | MIT | AGPL-3.0 |
| Primary focus | Product analytics | Web analytics |
| Pageview tracking | Yes | Yes |
| Custom event tracking | Yes (unlimited properties) | Yes (goals + custom properties) |
| Session recording | Yes | No |
| Feature flags | Yes | No |
| A/B testing | Yes | No |
| Funnels | Yes (complex multi-step) | Yes (basic goal funnels) |
| User surveys | Yes | No |
| Cohort analysis | Yes | No |
| Retention analysis | Yes | No |
| Real-time dashboard | Yes | Yes |
| Cookieless tracking | Yes | Yes (default) |
| GDPR compliance | Yes | Yes (built around it) |
| Script size | ~75 KB | <1 KB |
| API | Full REST API | Read-only Stats API |
| Tracking snippet | JavaScript SDK | Single <script> tag |
| Docker containers | ~20 | 3 (Plausible + ClickHouse + PostgreSQL) |
| Min RAM | 8 GB | 2 GB |
| Min disk | 30 GB | 5 GB |
| GitHub stars | ~25K+ | ~22K+ |
Resource Usage
The resource gap is the deciding factor for many self-hosters:
| Resource | PostHog | Plausible |
|---|---|---|
| Docker containers | ~20 | 3 |
| Idle RAM | 4-6 GB | 500 MB-1 GB |
| Active RAM | 8-16 GB | 1-2 GB |
| Tracking script | ~75 KB | <1 KB |
| Disk footprint | 20+ GB | 2-5 GB |
| CPU (idle) | 2-4 cores | 0.5 core |
Plausible runs comfortably alongside a dozen other self-hosted apps on a $12/month VPS. PostHog realistically needs a dedicated server or a beefy VPS with 16+ GB RAM to run well.
The tracking script size also matters for page performance. Plausible’s <1 KB script has zero measurable impact on page load. PostHog’s ~75 KB SDK (with session recording enabled) adds noticeable weight.
Installation Complexity
Plausible: Clone the Community Edition repo, edit a .env file with your domain and secret key, run docker compose up -d. The setup takes under 5 minutes. Three containers total.
PostHog: Run the official deploy script, which generates a complex Docker Compose setup. The stack includes ClickHouse, Redpanda (Kafka alternative), PostgreSQL, Redis, Celery workers, and multiple application servers. Initial setup takes 15-30 minutes and requires health monitoring of all services. Upgrading involves following a detailed migration guide.
Analytics Capabilities
Plausible gives you:
- Visitor count, unique visitors, bounce rate, visit duration
- Top pages, entry pages, exit pages
- Traffic sources (referrers, UTM campaigns, search terms)
- Geographic breakdown, device and browser stats
- Custom goal tracking (pageview goals, custom event goals)
- Custom properties on events
- Revenue attribution
- All on a single, fast-loading dashboard
PostHog gives you everything above, plus:
- Arbitrary event tracking with unlimited custom properties
- Session recordings with DOM snapshots and console logs
- Feature flags with targeting rules (user properties, percentages, cohorts)
- Multivariate A/B testing with statistical significance calculations
- User surveys triggered by events or properties
- Retention tables and lifecycle analysis
- User path mapping
- Cohort analysis
- SQL access via HogQL for custom queries
- Data warehouse for combining external data sources
If your analytics needs are “how many people visited, where did they come from, and what did they read” — Plausible answers all of those cleanly. If you need “what sequence of actions do users take before converting, and does version A or B of the signup flow perform better” — that’s PostHog territory.
Privacy
Both tools are privacy-friendly, but Plausible was built around privacy as a core value:
- Plausible: No cookies by default. No personal data collected. Script is <1 KB. Designed to be used without consent banners. Data stays on your server.
- PostHog: Can run without cookies. Session recordings capture DOM state (need careful masking of sensitive fields). Feature flags and surveys involve user identification. More configuration needed for a fully privacy-compliant setup.
If avoiding consent banners entirely is a priority, Plausible is the safer choice.
Use Cases
Choose PostHog If…
- You’re building a SaaS product or web application
- You need session recordings to debug user experience issues
- Feature flags and A/B testing are part of your workflow
- You have dedicated server resources (8+ GB RAM)
- Your team includes product managers and engineers who need deep behavioral data
Choose Plausible If…
- You run a blog, content site, or marketing site
- You want a simple, fast Google Analytics replacement
- Server resources are limited
- Page load performance is critical (< 1 KB script)
- Privacy compliance without complexity is important
- You want analytics that “just work” with zero maintenance
Final Verdict
Plausible wins for websites. It does exactly what most site owners need — traffic metrics, referrer analysis, and goal tracking — with minimal resources and zero privacy concerns. The dashboard is beautiful and fast.
PostHog wins for software products. If you’re building an application with user accounts, onboarding flows, and feature releases, PostHog’s integrated platform (analytics + recordings + flags + experiments) replaces three or four separate tools.
For selfhosting.sh readers running personal sites, blogs, or small projects: Plausible. For those building the next B2B SaaS: PostHog.
FAQ
Can PostHog replace Plausible for basic web analytics?
Technically yes — PostHog tracks pageviews and referrers. But it is massive overkill for a blog or content site. You would use 8+ GB of RAM and 20 containers to get information that Plausible provides in 500 MB with 3 containers. PostHog’s value is in product analytics features (session replays, feature flags, experiments) that websites do not need.
Does PostHog’s 75 KB script affect page speed?
It can. On a fast connection, 75 KB loads quickly, but on mobile or slow connections it adds noticeable latency. With session recording enabled, the SDK also uses CPU to capture DOM changes. Plausible’s <1 KB script has no measurable impact on any device. If page speed is a ranking factor for your site (it is for Google), the script size matters.
Can I run PostHog and Plausible together?
Yes. Some teams use Plausible for public-facing website analytics and PostHog for product analytics inside their application. The two tools serve different purposes and do not conflict. Both tracking scripts can coexist on the same page, though be mindful of the combined script size.
Is PostHog’s self-hosted version fully featured?
PostHog self-hosted (open-source) includes event tracking, session recordings, feature flags, A/B testing, and surveys. Some features available on PostHog Cloud (like the data warehouse and certain integrations) may have limitations or require additional configuration in the self-hosted version. Check the PostHog docs for the current feature matrix.
Which is better for GDPR compliance?
Plausible is the safer choice for GDPR. It collects no personal data, uses no cookies, and does not store IP addresses. You do not need consent banners. PostHog can be configured for privacy (cookieless mode, IP masking), but session recordings inherently capture user behavior and potentially personal data visible on screen. You will likely need consent management with PostHog.
Can Plausible do A/B testing?
No. Plausible tracks what happens on your site but cannot change what users see. A/B testing requires server-side or client-side feature flagging — that is PostHog’s territory. If you need A/B testing alongside lightweight analytics, run PostHog for experiments and Plausible for traffic metrics.
Related
Get self-hosting tips in your inbox
Get the Docker Compose configs, hardware picks, and setup shortcuts we don't put in articles. Weekly. No spam.
Comments