Head to head
PostHog vs Amplitude
The whole toolbox, or the sharpest tool. Scored on ten weighted measures, with every price read on the same day.
PostHog
7.1/10
No. 6 of 30
Amplitude
6.4/10
No. 15 of 30
Both give away a lot, both are aimed at product teams, and both will answer the core question of who did what and whether they came back. The difference is scope.
PostHog wants to be the only thing you install: analytics, session replay, feature flags, experiments, error tracking, surveys and a data warehouse behind one signup. Amplitude has stayed an analysis product and made the analysis better than anyone else's.
The verdict
It depends on whether you have an analyst.
Take Amplitude if somebody's job includes answering behavioural questions. Behavioural cohorts, pathfinder analysis and correlation reports go after which early behaviour predicts retention, which is a harder and more valuable question than what happened last week. Its free tier is the most generous in the category at 2 million events a month, with the whole platform included and unlimited seats on every plan.
Take PostHog if the alternative is buying four tools. Shipping a feature behind a flag, measuring its effect on a funnel, watching recordings of people failing to use it and reading the errors they hit, without leaving one product, is a materially faster loop. It is open source and self-hostable, and the company states that 97 per cent of its customers never pay.
The deciding practical point is the failure mode. A half-configured Amplitude that nobody understands is worse than a well-used PostHog, and half-configured Amplitude implementations are common. If you cannot name the person who will own it, take the broader, blunter tool.
Side by side
| PostHog | Amplitude | |
|---|---|---|
| Overall score | 7.1 / 10 | 6.4 / 10 |
| Rank on this list | No. 6 of 30 | No. 15 of 30 |
| From | Free then usage-based | Free then event-based |
| What that buys | 1M events, 5,000 session replays and 1M feature flag requests free every month, forever | 2M events a month free forever, with unlimited seats on every plan |
| Free tier | Yes, and it is the most generous on this list | Yes, 2M events a month |
| Cookies | Yes, by default | Yes, by default |
| Script size | Tens of kilobytes | Tens of kilobytes |
| Hosting | US or EU cloud, or self-hosted | US or EU |
| Trading since | 2020 | 2012 |
| Free events per month | 1 million | 2 million |
| Free session replays | 5,000 | 10,000 |
| Feature flags and experiments | Core to the product, 1M requests free | Included, limited on Free |
| Error tracking | 100K exceptions free | Not offered |
| Behavioural cohorts and paths | Basic | The deepest here |
| Self-hosting | Yes, open source | No |
| Learning curve | Broad and shallow | Steep |
The ten measures, scored
| Measure | PostHog | Amplitude |
|---|---|---|
| Insight quality | 8 | 8 |
| Privacy architecture | 5 | 4 |
| Value for money | 9 | 8 |
| Setup and time to data | 5 | 4 |
| Script weight | 3 | 3 |
| Analytical depth | 10 | 10 |
| Behaviour tooling | 9 | 7 |
| Data ownership | 9 | 8 |
| Accuracy | 6 | 6 |
| Vendor conduct | 8 | 7 |
| Weighted overall | 7.1 | 6.4 |
Higher is better. The overall figure is the weighted mean, relative to the 30 tools on this list. How this works.
Which one should you pick
Choose Amplitude if
You have an analyst who will genuinely use behavioural cohorts and path analysis rather than admire them.
Choose PostHog if
You would otherwise buy separate tools for flags, experiments, replay and error tracking, or you need to self-host.
Common questions
Which has the better free tier?
Amplitude on volume, at 2 million events a month with the full platform and unlimited seats. PostHog on breadth, at 1 million events plus 5,000 replays, 1 million flag requests and 100,000 exceptions. Pick by which dimension you will exhaust first.
Is Amplitude harder to learn than PostHog?
Yes. Amplitude's analytical model rewards someone who understands behavioural cohorts and rewards nobody else. PostHog is broader and blunter, which makes it easier to start with and easier to leave half-configured.
Can either replace Google Analytics on a marketing site?
Neither should. Both are tens of kilobytes, both set cookies by default and both need a consent banner in the UK and EU. Use a cookieless web analytics tool on the public site and keep these for the product.
Does PostHog do experiments as well as Amplitude?
PostHog's flags and experiments are more central to the product and free at high volume. Amplitude's analysis of the results is deeper. Teams that care most about experiment rigour often prefer a dedicated platform such as Statsig to either.
Both entries above are desk reviews written from published evidence, and every judgement is my opinion rather than a statement of fact. Read the full reviews for PostHog and Amplitude, or tell me where I have got something wrong.