I spent three months running PostHog and Amplitude side-by-side on the same product. Same events, same users, same questions. The experience clarified something that comparison blog posts usually gloss over: these two tools aren't really competing for the same buyer. PostHog is a developer tool that happens to do product analytics. Amplitude is a product-management tool that happens to have an SDK. The overlap is real, but the worldview behind each product shapes everything — pricing, data model, feature set, who actually opens the tool on Monday morning.
Here's what I found running both, and how I'd choose between them today.
What PostHog actually is
PostHog markets itself as a product analytics platform, but that undersells it. What you're really getting is a bundle: event analytics, session recordings, feature flags, A/B testing (experiments), and user surveys — all in one install. The analytics layer runs on ClickHouse, which matters because ClickHouse is a columnar database designed for fast aggregation over large event volumes. Queries that Postgres would choke on at 10M events, ClickHouse handles without breaking a sweat.
The all-in-one approach has a real cost advantage. A typical mid-stage startup might be paying for Amplitude (analytics) + LaunchDarkly (feature flags) + Hotjar (session recordings) + Statsig (experiments). PostHog replaces all four. Even if each individual component isn't best-in-class, the integration between them is tight: you can watch a session recording for a user who was in a specific experiment variant and triggered a specific funnel step. That kind of cross-feature drill-down usually requires gluing multiple tools together with user IDs and hoping the timestamps align.
The self-hosting story is real but comes with caveats. PostHog publishes a Helm chart for Kubernetes. You can run the full stack — ClickHouse, Kafka, PostgreSQL, Redis, the app servers — in your own infrastructure. This gives you full data control, which matters for companies in regulated industries (healthcare, fintech) or with European customers who care about where their data lives. But running a ClickHouse cluster in production is not trivial. You need someone who understands Kubernetes, storage provisioning, and ClickHouse-specific operational concerns (merges, replication lag, schema migrations). Most teams I've talked to start self-hosted and migrate to PostHog Cloud once they realize the ops burden isn't worth it. The notable exceptions are companies with genuine compliance requirements — they self-host because they have to, not because it's fun.
PostHog Cloud, for what it's worth, is solid. It runs on the same open-source codebase, hosted by PostHog's team, and the free tier gives you 1 million events per month, 5,000 session recordings, and unlimited feature flag evaluations. For a startup with under 50K MAU, you can run PostHog for months without paying anything.
What Amplitude actually is
Amplitude has been doing product analytics since 2012. That head start shows in the depth of its analytical capabilities. The behavioral graph — Amplitude's internal data structure that maps every user action to every other user action — powers features that PostHog doesn't have yet: behavioral cohorts (users who did X but not Y within Z days), predictive analytics (which users are likely to churn based on usage patterns), and causal analysis (did feature X actually cause metric Y to improve, or was it a confound?).
The Amplitude taxonomy is built around three concepts: events, user properties, and groups. Groups let you do account-level analytics in B2B products — "what features does Company X use" rather than "what features does User 47 use." PostHog supports group analytics too, but Amplitude's implementation is more mature and deeply integrated into every chart type.
Amplitude's CDP (Customer Data Platform) is worth calling out. It handles event ingestion, identity resolution, and audience syncing to downstream tools (ad platforms, email providers, CRMs). If you're running growth experiments where you need to build a cohort in your analytics tool and then push that cohort to Facebook Ads or Braze, Amplitude has that workflow built in. PostHog's approach to this is more manual — you'd export a CSV or build a webhook.
The experiment platform (Amplitude Experiment, formerly known as the Experiment product from the Experiment acquisition) handles feature flags and A/B testing. It's solid, but it's a separate product with separate pricing and a separate SDK. This is the fundamental philosophical difference: Amplitude is a suite of products, each priced independently. PostHog is one product, priced as one product.
Fastero
Connect your database. Ask questions. Get dashboards.
Postgres, BigQuery, Snowflake, and 10+ sources — live-connected, AI-powered, no dashboard builder learning curve.
Try free →Pricing: transparent vs opaque
This is where PostHog has an unambiguous advantage.
PostHog publishes its pricing on its website: $0 for the first million events per month, then $0.00031 per event beyond that. Session recordings: first 5,000 free, then $0.005 each. Feature flags: unlimited evaluations free, $0.0001 per API request beyond 1M. You can calculate your bill before you sign up. There's a pricing calculator on the site.
Amplitude's pricing is... not that. The free plan (Starter) gives you up to 50,000 monthly tracked users (MTUs) with core analytics. Beyond that, you're in "contact sales" territory. Enterprise plans are negotiated, and from what I've seen, they start around $30K-$50K/year and go up depending on volume. The Growth plan sits between free and enterprise but still requires talking to sales for exact pricing.
To be fair to Amplitude, this is how most enterprise SaaS tools price — they want to understand your use case before quoting a number. But if you're a developer evaluating tools on a Saturday morning, PostHog's transparency is refreshing. You know what you'll pay before you talk to anyone.
At scale, the comparison gets more nuanced. PostHog's per-event pricing can get expensive if you're tracking high-volume events (every click, every scroll, every API call). At 100M events per month, you're looking at roughly $30K/year on PostHog Cloud. Amplitude might actually be cheaper at that volume if you negotiate well, because enterprise contracts often include volume discounts that PostHog's self-serve pricing doesn't.
Feature-by-feature comparison
| Capability | PostHog | Amplitude |
|---|---|---|
| Core analytics (funnels, retention, trends) | Good | Excellent |
| Session recording | Built-in | Not available (use a third-party tool) |
| Feature flags | Built-in | Separate product (Amplitude Experiment) |
| A/B testing | Built-in | Separate product (Amplitude Experiment) |
| Surveys | Built-in | Not available |
| CDP / audience syncing | Limited (webhooks, exports) | Built-in |
| Behavioral cohorts | Basic | Advanced (behavioral graph) |
| Predictive analytics | Not available | Yes (churn prediction, lifecycle) |
| Group / account analytics | Supported | More mature |
| Self-hosting | Yes (Kubernetes) | No |
| Data warehouse | ClickHouse (included) | Snowflake/BigQuery export |
| Free tier | 1M events/month | 50K MTUs |
| Pricing transparency | Full public pricing | Contact sales for paid plans |
| SQL access | Yes (HogQL) | Limited |
| Open source | Yes (MIT-ish, functional source available) | No |
| Primary audience | Developers | Product managers |
Where PostHog wins clearly
Session recordings plus analytics in one tool. This is the feature combination that most teams tell me sold them on PostHog. You see a funnel drop-off, click into the users who dropped off, and watch their session recordings to understand why. No context-switching between tools, no trying to match user IDs across platforms. Amplitude simply doesn't offer session recordings. You'd need Hotjar, FullStory, or LogRocket alongside it.
Feature flags that share the analytics data model. When you run an experiment in PostHog, the experiment results, the feature flag assignments, and the analytics events all live in the same ClickHouse database. You can build a funnel that filters by experiment variant without any additional integration work. In Amplitude's world, this works too (between Amplitude Analytics and Amplitude Experiment), but it requires buying and configuring two products.
SQL access via HogQL. PostHog lets you write SQL queries directly against your event data using HogQL, their SQL dialect. If you're a developer who thinks in SQL, this is a major quality-of-life improvement. Amplitude has some SQL access through its Data Tables feature, but it's more constrained and not the primary interface.
Data ownership. Even if you use PostHog Cloud, you can export your data or migrate to self-hosted at any time. The codebase is open. If PostHog the company disappeared tomorrow, you could keep running the software. With Amplitude, your data lives in their infrastructure, and exporting it means CSV dumps or Snowflake integration — not the same as having the full queryable dataset on your own servers. If data ownership is a priority, see our guide on how to build a free analytics stack with open-source tools.
Where Amplitude wins clearly
Analytical depth. If your product team lives in the analytics tool — building cohorts, running lifecycle analyses, comparing retention curves across segments, doing statistical significance testing on experiments — Amplitude has more sophisticated tools for all of these. The behavioral graph is a genuine technical differentiator. It enables queries like "show me users who did action A then action B within 3 days but did NOT do action C" with a drag-and-drop interface that non-technical PMs can actually use.
Collaboration and governance for larger teams. Amplitude has better support for teams with 20+ people using the tool: notebooks, shared dashboards with access controls, chart annotations, audit logs. PostHog has dashboards and sharing, but the governance features are less mature. If your VP of Product needs to lock down certain reports so only her team can edit them, Amplitude handles that better.
CDP and downstream integrations. If you need to build audiences in your analytics tool and activate them in other platforms — send a push notification to users who haven't logged in for 7 days, exclude high-value users from a re-engagement ad campaign — Amplitude's CDP makes this straightforward. PostHog's approach requires more custom development: export a cohort, build a webhook, or use the API to pull user lists into your marketing tools.
Mature ecosystem. Amplitude has been around longer. The documentation is more complete. There are more blog posts, courses, and community resources. If you hire a product analyst, there's a good chance they already know Amplitude. PostHog is catching up fast, but the ecosystem gap is still real.
The licensing fine print
PostHog calls itself open-source, and the core product is indeed source-available under a permissive license. However, some premium features (like the SAML SSO, project permissions, and certain enterprise-grade capabilities) are under a proprietary license. The practical implication: you can self-host the full product, but some features require a paid license even on self-hosted deployments. This is common in open-core businesses (GitLab works the same way), but it's worth understanding before you assume "open-source" means "everything is free forever."
Amplitude is fully proprietary. There's no source code to inspect, no self-hosting option, and no way to run it independently of Amplitude's infrastructure.
When to pick PostHog
Your engineering team drives product decisions. You want analytics, session recordings, and feature flags in one tool rather than managing three vendor relationships. You care about data ownership, either for compliance reasons or on principle. Your event volume is in the low millions per month, where PostHog's free tier covers you or the per-event pricing stays reasonable. You prefer tools you can poke at — read the source code, write SQL, build integrations using a well-documented API.
When to pick Amplitude
Your product team includes non-technical PMs who need a visual interface for building complex behavioral queries. You're past the early stage and need enterprise features — SSO, role-based access, audit logs, SOC 2 compliance built into the vendor relationship. You want a CDP that syncs audiences to your marketing stack. Your analytics questions go beyond "what happened" into "why did it happen" and "what will happen next" — Amplitude's predictive and causal features are genuinely ahead here.
The third option nobody mentions
Both PostHog and Amplitude assume you want a dedicated analytics product with its own data store, its own event ingestion pipeline, and its own query engine. That's the right call for most teams, but it's not the only option. If your product data already lives in a database you control — Postgres, BigQuery, your data warehouse — you can run product analytics queries directly against that data without shipping events to a third-party platform. I wrote about this approach in how to do product analytics from your Postgres database.
If you're already pulling product data into a warehouse and want to layer analytics on top without adopting another event pipeline, Fastero connects to your existing database and lets you build dashboards and run analysis against your own data — no separate event ingestion required.
For more on how these tools stack up against the rest of the product analytics market, see our comparisons of Mixpanel vs Amplitude, Heap vs Mixpanel, and our three-way PostHog vs Amplitude vs Mixpanel comparison. If you're evaluating the broader visualization layer, check out best open-source alternatives to Tableau and our roundup of the best tools for data engineering teams in 2026.
Try Fastero free — connect your product database and ask questions in plain English — get the analytics you need without adopting another platform. No credit card required.