TL;DR
- Seed Stage: PostHog wins on cost and speed. Free tier covers most early needs. Amplitude's MTU pricing punishes fast-growing trials.
- Series A: The decision splits. Engineering-led teams stay with PostHog; PM-led teams trend toward Amplitude for its deeper behavioural segmentation UI.
- Series B+: PostHog's data ownership advantage grows. Amplitude's governance and enterprise integrations become genuinely useful — but at a significant cost.
- Migration is expensive: Switching mid-growth is a 4-8 week engineering project. Choose with a 3-year view, not a 3-month one.
- The real differentiator: PostHog is an all-in-one platform. Amplitude is a specialist. Neither is universally better.
The Wrong Question
Most PostHog vs Amplitude comparisons read like spec sheets. They list features side by side, score each platform on a rubric, and produce a conclusion that means nothing to you specifically. You're not a generic SaaS company — you're at a particular stage, with a particular team composition, and you're making a decision that will be expensive to reverse.
The question isn't "which tool is better?" The question is "which tool is right for where we are now, and won't handcuff us twelve months from now?"
I've set up both platforms for B2B SaaS companies at Seed through Series B. The patterns in this comparison come from that direct experience, not from vendor marketing or G2 reviews. I'll give you the honest version — including where each tool falls short — so you can make the call with your eyes open.
Before we get into the stage-by-stage breakdown, it's worth understanding why these two platforms exist in the first place — because their philosophies genuinely shape what they're good at.
Two Different Philosophies
PostHog was built by engineers, for engineers. It started as an open-source project with a simple premise: you should own your data, and the analytics tooling should live inside your development workflow, not separate from it. That DNA is still visible in the product today — feature flags, session replays, A/B testing, and event analytics are all built into a single platform. You instrument once; everything works together.
Amplitude was built around a different conviction: that understanding user behaviour deeply — not just tracking events — is what drives product growth. Their Behavioural Graph technology, cohort analysis, and predictive features are genuinely more sophisticated than PostHog's equivalents for teams that have the user volume and PM bandwidth to use them. They've invested heavily in the data visualisation layer and in making product managers self-sufficient without requiring SQL.
| Dimension | PostHog | Amplitude |
|---|---|---|
| Primary user | Engineers + Data teams | Product Managers + Growth teams |
| Deployment | Cloud or self-hosted (open-source) | Cloud only |
| Pricing model | Event-based (pay per event beyond free tier) | MTU-based (pay per Monthly Tracked User) |
| Free tier | 1M events/month on Cloud; unlimited self-hosted | Up to 50K MTUs on Starter plan |
| Built-in tooling | Analytics + Session Replay + Feature Flags + A/B testing + Surveys | Analytics + Cohort analysis + Predictions |
| Data ownership | Full (especially when self-hosted) | Data resides with Amplitude |
| Behavioural depth | Good | Excellent (Behavioural Graph, Compass) |
| Open-source | Yes (PostHog OS) | No |
Neither philosophy is superior in the abstract. The question is which one maps to your team's composition and your company's current constraints. That's what the next section breaks down.
Stage-by-Stage Breakdown
Under $1M ARR. Typically 2-8 engineers. Finding out what works.
At Seed, your analytics problem isn't "we need better insights." It's "we don't have enough reliable data to know anything yet." You're trying to instrument cleanly, catch bugs early, understand which of the three features you shipped last month actually get used, and do it all without a dedicated data engineer.
PostHog is the obvious default here for most engineering-led teams. The setup takes an afternoon, the free cloud tier covers a meaningful amount of early usage, and you get session replay alongside your event data — which means you can actually watch what users are doing when the funnel drop-off confuses you. The autocapture feature lets you start collecting meaningful data before you've defined a full event taxonomy.
Amplitude at Seed is a reasonable choice if your founding team is PM-heavy and you want a more polished analysis interface from day one. Their Starter plan covers up to 50,000 MTUs, which is plenty for early testing. But there's a catch: Amplitude's MTU pricing model means every unique user session counts against your quota, regardless of how many events they generate. If you run a high-volume viral loop or onboarding sequence, you can burn through free-tier headroom fast without generating much usable signal.
PostHog at Seed
- Free tier covers most early needs without instrumentation pressure
- Session replay helps debug onboarding without user interviews
- Feature flags let you ship behind gates from day one
- Engineering setup is fast; no separate SDK per feature
- UI is functional but not elegant — expect a learning curve
Amplitude at Seed
- More intuitive analysis UI for non-technical founders
- Strong funnel and retention visualisations out of the box
- MTU quota can burn unexpectedly with onboarding flows
- No session replay, feature flags, or A/B testing built in
- You'll need additional tooling to cover what PostHog bundles
$1M–$5M ARR. Team growing. PMF confirmed, now optimising.
Series A is where the PostHog vs Amplitude decision gets genuinely interesting. You've confirmed something works. Now you need to understand why it works and replicate it faster. Your product team is growing — you likely have a dedicated PM, maybe a growth function, and engineering bandwidth for more sophisticated instrumentation.
This is also where the two tools start to diverge meaningfully in capability. PostHog's A/B testing framework becomes a real asset at Series A — you can run experiments properly, tied directly to your feature flag system, without duct-taping Optimizely to your event stream. The unified data model means your experiment results, funnel data, and session replays all live in one place. That's not a minor convenience; it's how you catch the anomaly that a funnel chart alone would never surface.
Amplitude at Series A starts to justify its pricing in specific scenarios. If you have a product with complex multi-step user journeys — think B2B SaaS where the path from signup to core value spans multiple sessions over multiple days — Amplitude's Journeys feature and Behavioural Cohorts are genuinely better than PostHog's equivalents. The segmentation depth, particularly the ability to create cohorts based on sequences of events with time constraints, is more powerful in Amplitude's current interface.
The team composition question matters here more than the feature list. A Series A company where the growth function is PM-led and the team uses a heavily visual analysis workflow will get more out of Amplitude faster. A Series A company where the growth function is engineering-led, where experiment setup lives in the same PR as the feature, and where the data team writes HogQL or SQL against the warehouse — PostHog is the faster, cheaper option.
PostHog at Series A
- Experiment setup integrated with feature flags — no separate stack
- HogQL gives SQL-literate teams full flexibility
- Data warehouse export (S3, BigQuery, Snowflake) available
- Cost predictability as event volume scales
- Behavioural cohort capabilities are improving but lag Amplitude
Amplitude at Series A
- Behavioural cohorts and Journeys are best-in-class for complex flows
- No-SQL analysis interface genuinely accessible for PMs
- Amplitude Compass predicts which behaviours correlate with retention
- MTU pricing requires active cost management as the user base grows
- No built-in feature flags or session replay — separate tool budget required
$5M–$20M ARR. Multiple product lines. Data team forming.
By Series B, you've almost certainly moved beyond "which tool" as the primary question and into "how do these tools fit our data architecture." You likely have a data warehouse, a growing data team, and stakeholders across product, marketing, and finance who all want access to product usage data in different forms.
PostHog at Series B is a serious platform if you're willing to invest in it. The data ownership story becomes genuinely valuable here — particularly if you're in healthcare, fintech, or any sector with meaningful privacy obligations. Self-hosted PostHog on your own ClickHouse infrastructure means your event data never touches a third-party server. PostHog Cloud still keeps your data off your own infra but has robust EU and US data residency options. The warehouse sync features let you pipe data to BigQuery or Snowflake, where your data team can join it with CRM data, billing data, and support ticket data without moving it through a vendor API.
Amplitude at Series B starts to show its enterprise muscle. The governance features — project-level permissions, data dictionary enforcement, audit logs — matter in a 30+ person product organisation where you need instrumentation discipline. Their CDP (Customer Data Platform) capabilities help unify user identity across touchpoints, which is a real problem at this scale. Dedicated customer success support from Amplitude is also meaningful when you have a complex integration that breaks and the whole BI stack depends on it.
The honest tension at Series B is cost. Amplitude enterprise pricing is not transparent (it's custom and negotiated), but multiple engineering leaders I've talked to describe it as "the moment you start wondering if the sophistication is worth it." PostHog at similar scale — particularly on the cloud tier — is meaningfully cheaper per event. But cheaper isn't always the right frame. If Amplitude's Compass predictions are saving your growth team a month of analysis every quarter, the licensing cost may be irrelevant compared to the velocity gain.
PostHog at Series B
- Data ownership advantage is most material in regulated industries
- Lower per-event cost at high volume on Cloud tier
- Warehouse sync supports modern data stack architecture
- Self-hosted requires a dedicated ClickHouse data engineer
- Governance and access control features are improving but not enterprise-grade yet
Amplitude at Series B
- Enterprise governance: data dictionary, audit logs, project permissions
- CDP capabilities help with cross-touchpoint identity resolution
- Amplitude Compass and predictive features are most valuable with high user volume
- Custom enterprise pricing — budget this before you're locked in
- Strong vendor support; dedicated CSM on enterprise plans
$20M+ ARR. Multiple teams. Enterprise customers.
At true Growth stage, the analytics platform question is less about capability and more about organisational fit. You have, or soon will have, hundreds of events, dozens of dashboards, multiple teams with competing instrumentation priorities, and enterprise customers who ask about your data practices in security reviews.
PostHog at Growth stage is a legitimate choice, particularly if you've been building on it since Seed and have invested in instrumentation discipline. Clean event taxonomy and a well-maintained data dictionary are your own responsibility on PostHog — there's less tooling to enforce this than Amplitude provides. That trade-off becomes more expensive as the organisation grows. The engineering investment to maintain a healthy PostHog installation at this scale is real, but the cost per event and the data sovereignty benefits also become more material.
For Growth-stage companies evaluating a switch to Amplitude at this stage: the main arguments are standardised governance tooling, enterprise-grade support SLAs, and the maturity of the Amplitude ecosystem. The main arguments against: high switching costs, vendor lock-in, and the fact that you'll likely be their customer for a decade if you sign an enterprise contract without a clear exit plan.
Pricing at Each Stage
Pricing comparisons between these two tools are tricky because their models are structurally different. PostHog charges per event; Amplitude charges per Monthly Tracked User (MTU). Depending on your product's event density — how many events an average user generates per session — one model will be materially cheaper than the other.
PostHog Pricing
PostHog Cloud's free tier includes 1 million events per month. Beyond that, pricing scales per event, with the rate decreasing at higher volumes. Products with high event density — think complex SaaS applications that fire many events per interaction — can accumulate events faster than expected. But for most Seed and Series A companies, the free tier covers genuine usage.
Self-hosted PostHog is open-source and free to run. The real cost is infrastructure (ClickHouse can be resource-intensive at scale) and engineering time for maintenance. At Seed, you can run PostHog on a modest VPS; at Series B scale with millions of events daily, you're looking at a meaningful infrastructure bill and dedicated operational overhead.
Amplitude Pricing
Amplitude's Starter plan is free up to 50,000 MTUs. The Growth tier requires custom pricing (contact sales), and the Enterprise tier is negotiated. The MTU model means you pay per unique user, not per event — which can be a significant advantage if your users generate high event volume per session, but a disadvantage if you have many low-activity users who still count against your quota.
The practical challenge with Amplitude's pricing is predictability. As your user base grows — particularly with trials, freemium, or viral loops — your MTU count can spike without a corresponding increase in revenue or product value generated. Teams frequently report cost surprises when trial cohorts or marketing campaigns drive unexpected MTU growth.
Total Cost of Ownership
The licensing cost is only part of the picture. When comparing the two tools honestly, the TCO calculation should include:
- Replacement tools: If you choose Amplitude, budget separately for session replay (FullStory, LogRocket, or similar) and feature flags (LaunchDarkly or equivalent). PostHog bundles these.
- Engineering overhead: Self-hosted PostHog requires infrastructure maintenance. Cloud PostHog at scale requires data pipeline management. Amplitude is fully managed.
- Migration cost: Switching between platforms later costs 4-8 weeks of engineering time minimum. That's a real TCO factor even if it's invisible in the month-one decision.
- Training time: Amplitude's UI is faster for PMs to become productive in. PostHog requires more technical fluency. Neither is free to onboard.
A common mistake: comparing PostHog's all-in-one price against Amplitude's analytics-only price. If you choose Amplitude, add the cost of a session replay tool, a feature flag system, and a survey tool to get an apples-to-apples comparison. That delta is often $500-2,000/month at Series A scale before you even reach Amplitude's paid tier.
Feature Maturity Comparison
Both platforms have matured significantly in the last two years. But they've matured in different directions, and the gaps that existed two years ago have partly closed and partly widened.
Core Analytics: Funnels, Retention, Dashboards
Both tools handle standard funnel analysis, retention curves, and dashboards competently. The difference is in the analysis layer beneath. Amplitude's query builder is more accessible for non-SQL users; PostHog's HogQL interface gives SQL-literate teams more flexibility. For complex analyses — multi-step retention broken by cohort and acquisition channel — Amplitude's interface is currently faster for most PMs. PostHog's interface is catching up but still requires more SQL comfort for equivalent queries.
Experimentation: A/B Testing and Feature Flags
This is PostHog's strongest differentiator. The feature flag and experiment system are natively integrated — you can ship a feature behind a flag, run an experiment on a subset of users, and view the analysis all in the same product. Amplitude's experimentation product (Amplitude Experiment) exists and is solid, but it's a separate product with separate pricing and separate instrumentation overhead. For teams doing frequent experiments, PostHog's native integration is a meaningful velocity advantage.
Session Replay and Heatmaps
PostHog includes session replay natively. You can jump from a funnel drop-off directly to a replay of a specific user's session. This is one of the most underrated features for early-stage teams doing qualitative debugging. Amplitude does not include session replay — you need a separate tool, which means a separate instrumentation layer and the overhead of context-switching between products when diagnosing a conversion problem.
Behavioural Intelligence and Predictions
Amplitude's Behavioural Graph, Compass (which identifies which early behaviours correlate with long-term retention), and predictive cohorts are genuinely more advanced than PostHog's equivalents. These features become most valuable when you have sufficient user volume — 50,000+ MAUs — to produce statistically meaningful predictions. At Seed, this capability is largely irrelevant. At Growth stage, it can be genuinely impactful for retention modelling and churn prediction.
Data Warehouse and Integrations
Both platforms export data to warehouses (BigQuery, Snowflake, Redshift). PostHog's data warehouse sync is a first-class feature that makes it easy to join product event data with external sources directly in your warehouse. Amplitude's integrations ecosystem is broader at the enterprise level, particularly for marketing and CRM tools. If your stack is heavily Salesforce or HubSpot integrated, Amplitude's connector ecosystem is more mature.
Getting the most out of PostHog requires clean instrumentation from the start.
If your analytics setup is a mess — inconsistent event names, missing properties, broken funnels — we can audit it and rebuild it properly.
When to Switch — and What Migration Actually Costs
The "when should we switch" question comes up in three scenarios: you're outgrowing PostHog's capabilities, you're outgrowing Amplitude's pricing, or a new data lead joins with strong platform preferences. Each scenario has a different answer.
Trigger Points for Switching from PostHog to Amplitude
- Behavioural analytics depth: Your growth team needs Compass-style behaviour-to-retention correlation and PostHog's current capabilities aren't sufficient.
- PM self-service: Your product organisation has scaled to a point where PMs need to run complex analyses without SQL, and the PostHog UI isn't fast enough for them.
- Enterprise governance: You're selling to enterprise and need audit-grade data governance, role-based access, and a formal data dictionary that enforces instrumentation standards.
- Vendor support SLAs: Your board or enterprise customers require formal SLAs that PostHog Cloud doesn't offer at your tier.
Trigger Points for Switching from Amplitude to PostHog
- Cost at scale: Your MTU count has grown faster than expected and the bill is no longer justified by the value extracted.
- Data sovereignty: You're entering a regulated market (healthcare, fintech, EU data localisation) where third-party data custody is a liability.
- Tool consolidation: You're paying for session replay, feature flags, and A/B testing separately and want to consolidate onto a single instrumentation layer.
- Engineering-led growth function: Your team composition has shifted and the engineering-centric workflow of PostHog maps better to how growth decisions actually get made.
What Migration Actually Costs
Both migrations — PostHog to Amplitude and Amplitude to PostHog — are significantly more expensive than they appear. The visible cost is SDK replacement and re-instrumentation. The invisible cost is everything else.
A realistic migration timeline for a Series A company with a mature event taxonomy looks like this: two weeks for event taxonomy redesign (the new platform's data model differs from your current one and you shouldn't just port events 1:1), two weeks for SDK implementation and testing across platforms, two weeks running both platforms in parallel to validate data parity, and an ongoing period of rebuilding dashboards, funnels, and cohorts in the new tool. The total is often 6-10 weeks of engineering and PM time, not 2-3.
Historical data is typically the biggest pain point. Most migrations don't include a historical data backfill — meaning you lose continuity in your retention cohorts, funnel benchmarks, and experiment baselines at the moment of migration. That's a real analytical cost, not just an operational one. See our migration guide for a more detailed view of what the process involves technically.
Never migrate your analytics platform during a critical growth experiment or a fundraising process. The data discontinuity will make both meaningfully harder to interpret. Plan migrations for relatively quiet quarters with no major analytical dependencies.
The Decision Framework
Rather than a score matrix, here's how to think through the decision based on your actual situation.
Your team and context favour PostHog
- Engineering-led growth function
- Seed or early Series A (cost matters)
- Need session replay + feature flags in one tool
- Regulated industry or data sovereignty requirements
- Want to avoid tool sprawl with a small team
- Running frequent product experiments tied to features
- Technical PMs comfortable with SQL or HogQL
Your team and context favour Amplitude
- PM-led growth function with low SQL fluency
- Complex multi-step behavioural analysis is core to your workflow
- Series B+ with budget for enterprise tooling
- Large product org needing governance and access control
- High-volume retention modelling (Compass / predictions)
- Salesforce or HubSpot-heavy integration requirements
- Need formal SLAs and dedicated vendor support
There are also scenarios where the answer is "neither by itself." Some Series B companies run PostHog for engineering and experimentation workflows while using Amplitude for the growth and marketing teams who need its visualisation layer. This dual-stack approach adds instrumentation complexity and data inconsistency risk, but it's a pragmatic solution when the two functions genuinely have incompatible tool preferences.
If you're evaluating now and still genuinely unsure, start with PostHog. The free tier means the cost of being wrong is low, the integrated tooling reduces decisions you'd otherwise have to make in parallel, and migrating from PostHog to Amplitude when you have the scale to justify Amplitude is a known problem with known solutions. Migrating from Amplitude to PostHog mid-Series B because the bill got out of hand is a harder conversation to have with your board.
For a broader comparison that includes Mixpanel in the picture, see our PostHog vs Amplitude vs Mixpanel breakdown, which walks through how a third option changes the calculus at each stage. And if you're already on PostHog and trying to decide whether the current implementation is working as well as it should be, the PostHog implementation guide covers when a fresh setup outperforms patching what's there.
FAQ
Is PostHog truly a free alternative to Amplitude?
For most Seed and early Series A companies: yes, in practice. PostHog Cloud's free tier covers 1 million events per month, which is substantial for a product with under 5,000 active users. Self-hosted PostHog is free to run but costs engineering time for maintenance. Amplitude's free tier is based on 50,000 MTUs — more intuitive to understand, but potentially more limiting for high-event-volume products. Neither "free" option is truly free once you factor in operational overhead, but PostHog's free tier is more generous in most early-stage scenarios.
Which tool is better for product-led growth (PLG) strategies?
PostHog's all-in-one architecture is a natural fit for PLG execution — feature flags let you gate features to user segments, session replay shows you where activation breaks, and experiments run directly on the instrumentation layer. Amplitude is better for PLG analysis — understanding which activation patterns predict long-term retention, modelling expansion behaviour in existing cohorts. If you're building a PLG motion, PostHog is the execution platform and Amplitude is the intelligence layer. Most early-stage PLG teams don't need both.
How difficult is migration between the two platforms?
More difficult than vendors suggest. The event schemas differ, the SDK implementations differ, and historical data rarely survives cleanly. Budget a minimum of 4-6 weeks for engineering time, plan for a period of running both tools in parallel, and accept that some historical analyses will have a discontinuity at the migration date. The migration guide for switching to PostHog has a realistic technical walkthrough of what the process looks like.
Can PostHog scale to Series B and beyond?
Yes, both on Cloud and self-hosted. PostHog Cloud has no hard scale limits and handles billions of events for large customers. Self-hosted PostHog at Series B scale (tens of millions of events per month) requires a proper ClickHouse configuration and dedicated data engineering oversight — it's not a one-person-part-time job. Cloud PostHog removes that operational overhead at the cost of slightly less control over infrastructure. Whether that trade-off makes sense depends on your data sovereignty requirements.
Which platform offers better data privacy and ownership?
PostHog, unambiguously, if you self-host. Your events never touch PostHog's servers. PostHog Cloud has data residency options and strong compliance certifications, but the data is on their infrastructure. Amplitude is a cloud-only SaaS — your data resides with them, governed by their privacy programme. For HIPAA, GDPR Article 25 data-by-design requirements, or companies that sell to government or regulated enterprise clients who audit your data practices, PostHog self-hosted is the only option that fully satisfies the requirement. See our guide to HIPAA-compliant product analytics for more detail on what this involves in practice.
The Bottom Line
PostHog and Amplitude are both serious products used by serious companies. This comparison isn't meant to declare a winner — it's meant to give you the frame for making the right call for your specific situation.
If you take one thing from this: the decision you make at Seed will likely persist longer than you expect. Switching platforms mid-growth is expensive in ways that don't show up on a feature comparison spreadsheet. Make the Seed decision with a three-year view, not a three-month one. If you're genuinely uncertain at Seed, PostHog's lower cost of entry means it's the lower-risk starting point. You can always migrate to Amplitude when the scale and organisational complexity genuinely justify it.
And if you're already mid-stack on either platform, questioning whether it's still the right call: the answer isn't usually "switch." It's usually "get more from what you have." A well-instrumented PostHog installation beats a poorly used Amplitude account every time, and vice versa. The platform matters less than the discipline of instrumentation and the habit of actually using the data to make decisions.
If you'd like a second opinion on your current analytics setup — whether you're on PostHog, Amplitude, or something else — get in touch. We work with B2B SaaS companies at Series A-C to audit and rebuild analytics stacks that have drifted from their original intent.
Sources & Further Reading
- PostHog Pricing — official pricing page with event tier details
- Amplitude Pricing — official pricing page with MTU model explanation
- PostHog's own comparison with Amplitude — useful for feature specifics
- PostHog Experiments documentation — for A/B testing implementation detail
- PostHog vs Amplitude vs Mixpanel — how a third tool changes the decision
- Migrating to PostHog — technical process and data continuity guide
- PostHog Setup Guide — instrumentation best practices
- HIPAA-compliant product analytics — data sovereignty requirements in regulated industries
Evaluate your tools
Series B teams need different analytics than Series A. We assess and recommend.
See Analytics Audit Sprint →
About the Author
Jake McMahon is a product analytics consultant who has set up and audited PostHog and Amplitude installations across B2B SaaS companies at Seed through Series B. He specialises in instrumentation strategy, PLG analytics, and helping product teams get from data collection to actual product decisions.