

Pricing
PostHog Pricing Explained: What a SaaS Pays
PostHog's pricing page claims 97% of companies use it for free, and for most MVPs that holds. The bill that does arrive rarely comes from the analytics you installed it for, but from session replay, identified events, a hasty add-on or a staging project eating your free allowance. Here's how each product is billed, what sits behind paid packages, and the real cost at four stages.
How much does PostHog cost for a SaaS?
$0 at MVP stage, and roughly $15 to $50 a month at 1,000 to 10,000 active users with sensible settings: PostHog's monthly free allowances — 1 million events, 5,000 session recordings and 1 million feature flag requests — cover most of a young B2B app's usage, and the first dollars usually come from session replay. Leave every default switched on, though, and the same 10,000 users can cost ~$686 a month.
There's no base fee and no per-seat charge. PostHog bills each product separately by usage — events, recordings, flag requests, survey responses — with a free allowance that resets monthly and per-unit prices that fall as volume grows. Allowances belong to the organisation, not the project, so development, staging and production all draw from the same pool.
What does each PostHog product cost?
Prices are in USD, excluding tax, as of October 2026, from PostHog's pricing page. Vendors change pricing, so check it before you commit.
| Product | Billed per | Free each month | First paid tier |
|---|---|---|---|
| Product analytics (includes web analytics) | Event | 1 million | $0.00005 (1–2 million) |
| Identified events add-on (person profiles) | Identified event | 1 million | +$0.000198 (1–2 million) |
| Group analytics add-on | Identified event | 1 million | +$0.000071 (1–2 million) |
| Session replay (web) | Recording | 5,000 | $0.005 (5,001–15,000) |
| Session replay (mobile) | Recording | 2,500 | $0.01 (2,501–15,000) |
| Feature flags and experiments | Flag request | 1 million | $0.0001 (1–2 million) |
| Surveys | Response | 1,500 | $0.10 (1,501–2,000) |
| Error tracking | Exception | 100,000 | $0.00037 (to 325,000) |
| Data warehouse sources | Synced row | 1 million | $0.000015 (to 10 million) |
| AI observability (LLM calls) | Event | 100,000 | $0.00035 (to 150,000) |
| Logs | GB ingested | 10 GB | $0.25 per GB |
Every rate falls with volume: events drop to $0.000009 above 250 million a month, recordings to $0.0015 above 500,000. Put plainly, the first paid tier of events is $50 per extra million. Web analytics is billed as ordinary events, and experiments share the feature flag allowance.
Is the PostHog free plan enough for an MVP?
For building and validating, yes; once customers depend on the app, add a card — it costs nothing extra until you pass the allowances. Both plans get the same free tier every month. The free plan's catches:
- One project. No separate staging or development project until you add a card; pay-as-you-go gives you six.
- Usage stops at the limits. Extra events are permanently dropped — safe for your wallet, bad for your data the week a launch goes well.
- Shorter retention. Events are kept for a year and recordings for 30 days, against seven years and 90 days on pay-as-you-go.
- Community support only. Pay-as-you-go adds email support.
Adding a card moves you to pay-as-you-go: a $0 base, the same allowances, then usage pricing. The trap with extra projects: three each sending 400,000 events count as 1.2 million, not three free tiers.
How do identified and anonymous events change the price?
Every event pays the base rate; events from identified users also pay a person-profile rate, so above the free million an identified event costs $0.000248 against $0.00005 for an anonymous one. That's $248 versus $50 per extra million in the first paid tier, though person profiles get their own free million too.
The web SDK's default, person_profiles: 'identified_only', keeps visitors
anonymous until you call identify(); after that, their events are
identified. Inside a SaaS app that's right — cohorts and flag targeting by plan
need person profiles — so the saving is on anonymous marketing traffic. Two
traps catch B2B teams:
- Server-side events are identified by default. Backend SDKs create person
profiles unless you send
$process_person_profile: false— right for a Stripe webhook tied to a user, wasteful for system events. - Group analytics bills on every identified event. It's the add-on B2B products want, analysing by customer account rather than user. Once enabled, it charges on all identified events, not just those with a group attached.
Which PostHog features are paid?
No core product is locked behind a package: a card buys volume, projects and retention, while team security and administration — enforced SSO, SAML, audit logs, access control, priority support — sit in platform packages from $250 a month, added on top of usage — see PostHog's platform packages page.
| Boost | Scale | Enterprise | |
|---|---|---|---|
| Price | $250/month | $750/month | Custom |
| Projects | Unlimited | Unlimited | Unlimited |
| Enforce SSO login and 2FA for your team | Yes | Yes | Yes |
| Access control, white labelling, HIPAA BAA | Yes | Yes | Yes |
| Activity (audit) logs | 7 days | 2 months | 60 months |
| Session replay retention | 12 months | 12 months | 60 months |
| SAML and OIDC sign-in | No | Yes | Yes |
| Approval workflows for flag changes | No | Yes | Yes |
| Priority support target | Not included | 24 hours | 8 hours |
| Role-based access control, SCIM, property access control | No | No | Yes |
| Dedicated account manager, custom MSA | No | No | Yes |
Without a package there's no access control at all: every member can see every project. Signing in with Google, GitHub or GitLab needs no package; enforcing SSO needs Boost, and SAML needs Scale. All of this governs your team's access to PostHog, not anything customers see — so an MVP rarely needs it until an enterprise buyer asks who can watch recordings of their staff.
What will PostHog actually cost at each stage?
Illustrative estimates from October 2026 list prices for a B2B web app where every in-app user is signed in, with the marketing site on a separate tool. We assume ~90 events per monthly active user — PostHog publishes 87 for its own B2B app with autocapture on — and eight sessions per user a month.
| Stage | Assumptions | Arithmetic | Monthly total |
|---|---|---|---|
| (a) Pre-launch | Free plan; one project; a few hundred test users; under 1 million events and 5,000 recordings | Inside free allowances | $0 |
| (b) Launched, ~1,000 MAU | Pay-as-you-go; 90,000 identified events plus 30,000 from staging; every session recorded (8,000); ~16,000 flag requests | Events and flags free; 3,000 recordings × $0.005 | $15 |
| (c) ~10,000 MAU, tuned | 900,000 identified events plus 50,000 from staging; replay sampled to 15,000 of 80,000 sessions; ~240,000 flag requests | Events and flags inside allowances; 10,000 recordings × $0.005 | $50 |
| (d) ~10,000 MAU, defaults left on | Autocapture untouched at 300 events per user (3 million, all identified); every session recorded (80,000); group analytics on | Base $50 + $34.30; person profiles $198 + $69.70; groups $71 + $30; replay $50 + $122.50 + $60 | ~$686 |
The gap between (c) and (d) is configuration, not customers. Keeping identified events under the free million zeroes three lines at once — base rate, person profiles and group analytics — so trimming autocapture matters most, with replay sampling second.
PostHog is one line of your running costs; the rest is broken down in what it costs to maintain a SaaS.
How do you avoid a surprise PostHog bill?
Set a billing limit on every product the day you add a card, then cut volume at the source: trim autocapture, sample replays and keep anonymous traffic anonymous.
- Billing limits. Set a dollar limit per product in your organisation's billing settings. At the limit PostHog stops ingesting — that data is lost, not billed — and owners get emails at 80% and 100% of both the limit and the free allowance. Logs custom retention is the one charge no limit caps.
- Autocapture. Every recorded click is a billable event. Use an allow or ignore list, or switch it off and capture the handful of events that matter.
- Replay triggers. Trigger groups combine a sample rate, a minimum duration and URL, event or feature-flag triggers: record all of onboarding and checkout, a sample elsewhere, nothing under a few seconds.
- Pageviews, once. With
defaultsset to'2025-05-24'or later, the JavaScript SDK captures pageviews on route changes itself. Delete any manual$pageviewcomponent an older Next.js tutorial had you add, or every page counts twice. - Anonymous stays anonymous. Keep
identified_only, and mark server-side system events with$process_person_profile: false. - Flag requests. Server-side checks without local evaluation cost a request each; local evaluation polls cost roughly 86,400 requests a month per server instance. Requests that evaluate no active flag aren't billed, so switch flags off once a rollout is done.
Does PostHog's EU cloud matter for UK founders?
PostHog runs a US cloud in Virginia and an EU cloud in Frankfurt; you pick one at sign-up. For UK founders the EU cloud is the low-paperwork choice, because UK adequacy regulations cover the EEA. Consent is the bigger question: under PECR, identified tracking and session replay need opt-in, while cookieless mode stores nothing on the device but gives up replay — our PostHog vs Plausible vs GA4 comparison covers the set-up, and our UK GDPR guide for SaaS startups the rest. This is general guidance, not legal advice.
Our default: PostHog in the app, Plausible on the site
Our default stack is Next.js, Postgres (often on Supabase), Clerk, Stripe and Resend on Vercel — reasoning in our SaaS MVP tech stack for 2026 — with PostHog for product analytics inside the app and Plausible or Vercel Analytics for the marketing site.
That split saves money: marketing traffic is anonymous, high-volume and needs no replay, so it stays off the PostHog bill entirely.
In practice: EU cloud; pay-as-you-go from launch so staging gets its own project; a billing limit on every product; autocapture trimmed to an allowlist; replay sampled, with onboarding recorded in full; billing events sent server-side from the Stripe webhook. Group analytics goes on once paying teams need account-level reporting. We don't buy platform packages for an MVP; Boost earns its $250 when a customer contract demands access control or an audit trail. Choosing the events that matter, starting with activation, is part of our product strategy work, and the weekly rhythm for reading them is in the first 90 days after an MVP launch.
We deviate when a client wants one tool and no consent banner: PostHog alone in cookieless mode, giving up session replay.
Frequently asked questions
Is PostHog free to use?
Yes, with no time limit, as long as you stay inside the monthly allowances. The free plan needs no card and stops collecting at the limits, so it can't bill you. Adding a card keeps the same allowances and charges only for usage above them.
Is PostHog cheaper than Plausible?
For a SaaS app, usually yes — PostHog is free to 1 million events a month, while Plausible has no free plan and starts at $9 a month after a 30-day trial. But they price different jobs. Plausible charges by pageviews for cookieless marketing-site analytics; PostHog charges by events, recordings and flag requests for what users do inside the product. Most SaaS teams run both, and at MVP scale the pair can cost as little as $9 a month.
Does PostHog offer startup credits?
Yes — eligible startups get $50,000 in credits for 12 months. PostHog's programme is for companies under two years old that have raised less than $5 million; Y Combinator companies get a separate, renewable deal. Credits matter most once replay or event volume is pushing you past the free tier.
What happens when you go over the PostHog free tier?
On the free plan the extra data is dropped; on pay-as-you-go you're charged per unit up to any billing limit you've set. At a limit, events are discarded and flags return a default response until you raise it or the next billing period starts — so set limits high enough that a good week doesn't blind your dashboards.
Want an analytics setup costed before you build?
Your PostHog bill is mostly decided in week one: what autocapture records, which sessions get replayed, who gets identified and where the marketing site's traffic goes. To make those choices deliberately, book a free scoping call. We'll map the analytics around your product, running costs included, and quote the build fixed-price.





Leave a comment