PostHog Pricing 2026: What Usage-Based Pay-As-You-Go Actually Costs
PostHog does not charge per seat. Model the real monthly bill from official per-event, per-recording, and per-request pricing at 1, 5, and 10-seat startup scale, including where usage-based billing actually hurts.
Decision Brief
What to do with this research
PostHog's pricing page lists per-unit rates, not a monthly number, so most teams cannot tell what they will actually pay until the first invoice arrives. Team size barely matters — PostHog does not charge per seat. What drives the bill is event volume, session replay volume, feature flag request volume, and survey response volume, each billed on its own sliding scale after a free monthly allowance. A 2-person app doing 2M events a month can land near $50. A 10-person product doing 80M events, 100K replays, and 40M flag requests can land above $4,000, with feature flags and session replay — not the core analytics events — often the line items that surprise teams first.
SaaS Pricing Checks changes
Get a practical ToolPick alert when pricing, free-plan limits, policy risk, or alternatives change.
Weekly at most · one-click unsubscribe
PostHog's pricing page lists per-unit rates, not a monthly number, so most teams cannot tell what they will actually pay until the first invoice arrives. Team size barely matters — PostHog does not charge per seat. What drives the bill is event volume, session replay volume, feature flag request volume, and survey response volume, each billed on its own sliding scale after a free monthly allowance. A 2-person app doing 2M events a month can land near $50. A 10-person product doing 80M events, 100K replays, and 40M flag requests can land above $4,000, with feature flags and session replay — not the core analytics events — often the line items that surprise teams first.
- Pricing is per-product and per-usage-unit, not per-seat — team headcount is not the cost driver, usage is
- Session replay and feature flag requests scale faster than most teams expect and often exceed the analytics event bill
- Billing limits exist and are the correct guardrail before enabling replay or flags broadly, because overage past a limit means data loss, not a bigger bill
Keep reading for the full analysis.
Where this decision goes next
Skip the scroll: the pages most readers open after this one.
AI-Powered Cybersecurity for SaaS ApplicationsRead the next related article.PostHog's pricing page is honest about one thing and quiet about another. It is honest that pricing is usage-based, not seat-based. It is quiet about what that actually means in dollars once a product has real traffic. The pricing page shows per-event, per-recording, and per-request rates in scientific-notation-adjacent numbers like $0.0000343 per event, and while PostHog does offer an on-page calculator for single-product estimates, what the calculator cannot do is model how five separately-metered products compound as a real product grows.
This guide does that multi-product growth modeling. It pulls the live per-unit tiers straight from PostHog's pricing page and billing documentation as of today, then runs three realistic usage scenarios — a 1-person indie project, a 5-person early-stage startup, and a 10-person growth-stage team — through the actual tiered math PostHog bills on. The existing PostHog review on ToolPick covers feature fit and workflow trade-offs; this piece is narrower and more concrete: what does the invoice say.
Quick Decision
Team size is close to irrelevant to the PostHog bill. What matters is product usage: monthly events, session replay recordings, feature flag/experiment requests, and survey responses, each billed on an independent sliding scale after its own free allowance. A low-traffic solo project can run PostHog for the price of a coffee subscription. A growth-stage product with heavy replay sampling and flag-gated rollouts can land in four figures a month before anyone touches the Enterprise tier. The decision that matters is not "can we afford PostHog" — it is "which usage dial do we control before it controls our bill."
How PostHog Actually Bills: Per Product, Not Per Plan
PostHog does not sell "Starter," "Growth," and "Enterprise" tiers the way most SaaS pricing pages do. Every product inside the platform — product analytics, session replay, feature flags and experiments, surveys, error tracking, data warehouse, logs — has its own monthly free allowance and its own pay-as-you-go rate ladder. There is no bundled monthly minimum on the standard plan, and every seat is included at every usage level; PostHog states plainly that team members are unlimited on all tiers.
That structure has one immediate consequence worth internalizing before doing any math: adding a tenth engineer to the workspace costs nothing. What costs money is turning on more products, sampling more session replay, or shipping a product with ten times the users. The spend conversation for PostHog is a usage conversation, not a headcount conversation — which is a genuinely different buying calculus than Amplitude or Mixpanel, where the cheapest plans cap on Monthly Tracked Users and a growing team competes with a growing user base for the same allowance. (For that seat-and-user-count comparison, see Amplitude vs. Mixpanel.)
Free monthly allowance, per product
| Product | Free tier every month |
|---|---|
| Product analytics | 1,000,000 events |
| Session replay | 5,000 recordings (2,500 for mobile) |
| Feature flags & experiments | 1,000,000 requests |
| Surveys | 1,500 responses |
| Data warehouse | 1,000,000 rows synced |
| Error tracking | 100,000 exceptions |
Everything above that line, on every product, is billed on a declining per-unit rate as volume climbs — the more a team uses a given product, the cheaper each additional unit gets, but the absolute bill still climbs because volume is climbing faster than the discount.
The Tiered Rate Cards That Actually Set the Bill
These are the live per-unit rates, pulled straight from PostHog's pricing page. This is the table most teams never actually read past the first line, and it is the table that determines whether a given usage pattern costs $40 or $4,000.
| Product | Volume band | Price per unit |
|---|---|---|
| Product analytics | 1M–2M events | $0.0000500 |
| Product analytics | 2M–15M events | $0.0000343 |
| Product analytics | 15M–50M events | $0.0000295 |
| Product analytics | 50M–100M events | $0.0000218 |
| Product analytics | 100M–250M events | $0.0000150 |
| Product analytics | 250M+ events | $0.0000090 |
| Session replay (web) | 5K–15K recordings | $0.0050 |
| Session replay (web) | 15K–50K recordings | $0.0035 |
| Session replay (web) | 50K–150K recordings | $0.0020 |
| Session replay (web) | 150K–500K recordings | $0.0017 |
| Session replay (web) | 500K+ recordings | $0.0015 |
| Feature flags/experiments | 1M–2M requests | $0.000100 |
| Feature flags/experiments | 2M–10M requests | $0.000045 |
| Feature flags/experiments | 10M–50M requests | $0.000025 |
| Feature flags/experiments | 50M+ requests | $0.000010 |
| Surveys | 1.5K–2K responses | $0.100 |
| Surveys | 2K–10K responses | $0.035 |
| Surveys | 10K–20K responses | $0.015 |
| Surveys | 20K+ responses | $0.010 |
A pattern worth noticing before the scenarios: session replay is priced roughly two orders of magnitude higher per unit than product analytics events, and feature flag requests sit in between. That single fact is why the "biggest lever" answer in the FAQ is replay sampling, not event volume — teams instinctively worry about tracking too many events, when the line item that actually spikes bills is usually recording too many sessions.
Modeling the Real Bill at 1, 5, and 10 Seats
These are cost models built from the official rate cards above and stated usage assumptions, not a vendor-published bill or a guarantee — actual invoices depend on a team's real event volume, replay sampling rate, and flag usage, which PostHog's own usage dashboard reports exactly once a product is live. Use these as planning anchors, then verify against real numbers within the first billing cycle.
1 seat — solo indie project, early traffic
Assumptions: roughly 500 monthly active users generating about 2M product-analytics events a month, light session replay sampling that stays under 2,000 recordings, and feature-flag usage under the 1M free allowance.
- Product analytics: 1M free + 1M billed in the 1M–2M band at $0.00005 = $50
- Session replay: under 5,000 free allowance = $0
- Feature flags: under 1,000,000 free allowance = $0
- Estimated monthly total: ~$50
At this scale, PostHog is close to free in practice. The free tier alone (1M events, 5K replays, 1M flag requests) comfortably covers most pre-revenue and early-beta products, and the first paid dollar typically shows up only once event volume crosses 1M/month.
5 seats — early-stage startup, ~20,000 MAU
Assumptions: 15M product-analytics events a month, session replay sampled at roughly 5% of sessions (about 25,000 recordings), feature flags evaluated on most page loads plus a couple of active experiments (about 5M requests), and an NPS-style survey collecting around 3,000 responses a month.
- Product analytics: 1M free, then 1M @ $0.00005 + 13M @ $0.0000343 = $50.00 + $445.90 = $495.90
- Session replay: 5K free, then 10K @ $0.0050 + 10K @ $0.0035 = $50.00 + $35.00 = $85.00
- Feature flags: 1M free, then 1M @ $0.0001 + 3M @ $0.000045 = $100.00 + $135.00 = $235.00
- Surveys: 1.5K free, then 0.5K @ $0.10 + 1K @ $0.035 = $50.00 + $35.00 = $85.00
- Estimated monthly total: ~$901
Notice the shape here: analytics events are still the largest line item, but feature flags and session replay combined ($320) already outweigh what most teams budget mentally for "the flags and replay stuff." This is the scale where a team's first billing-limit conversation should happen — before flags and replay usage double alongside user growth.
10 seats — growth-stage product, ~150,000 MAU
Assumptions: 80M product-analytics events a month, session replay sampled at roughly 10% of sessions (about 100,000 recordings), heavier flag usage from ongoing experiments plus gradual rollouts (about 40M requests), and a broader in-product survey program collecting 15,000 responses a month.
- Product analytics: 1M free, then 1M@$0.00005 + 13M@$0.0000343 + 35M@$0.0000295 + 30M@$0.0000218 = $50.00 + $445.90 + $1,032.50 + $654.00 = $2,182.40
- Session replay: 5K free, then 10K@$0.0050 + 35K@$0.0035 + 50K@$0.0020 = $50.00 + $122.50 + $100.00 = $272.50
- Feature flags: 1M free, then 1M@$0.0001 + 8M@$0.000045 + 30M@$0.000025 = $100.00 + $360.00 + $750.00 = $1,210.00
- Surveys: 1.5K free, then 0.5K@$0.10 + 8K@$0.035 + 5K@$0.015 = $50.00 + $280.00 + $75.00 = $405.00
- Estimated monthly total: ~$4,070
At this scale, feature flags ($1,210) have grown to nearly rival product analytics as a share of the bill, driven entirely by how many flags and experiments are live and how often they get evaluated per pageview — not by seat count. A team running fewer, coarser flags at this same traffic level could plausibly cut this line by half without losing rollout capability.
Bill scenario summary
| Scale | Events/mo | Replays/mo | Flag requests/mo | Survey responses/mo | Estimated monthly bill |
|---|---|---|---|---|---|
| 1 seat, early beta | ~2M | ~2K (free) | ~0.5M (free) | 0 | ~$50 |
| 5 seats, ~20K MAU | ~15M | ~25K | ~5M | ~3K | ~$901 |
| 10 seats, ~150K MAU | ~80M | ~100K | ~40M | ~15K | ~$4,070 |
The jump from 5 to 10 seats in this model is roughly 4.5x on the bill despite headcount only doubling — because the underlying user base and product usage grew by roughly the same 4.5-8x, not because PostHog charged more per person. That is the core mental model this pricing structure demands: budget for user growth and feature-usage growth, not for team growth.
Who Should Pick PostHog's Pay-As-You-Go Plan
Pick PostHog PAYG when a team wants product analytics, session replay, feature flags, and experiments in one tool and is comfortable owning usage discipline in exchange for not paying per seat. It is a strong fit for:
- Early-stage teams under roughly 1M monthly events across all products, where the free tier likely covers the whole workload and the first invoice may be $0.
- Engineering-led teams that will actually use the SQL/API access to build billing-limit alerting and usage dashboards rather than treating the bill as a black box.
- Teams replacing three or four separate tools (an analytics tool, a session-recording tool, a feature-flag tool, a survey tool) where the consolidated PostHog bill is compared against the sum of those four separate subscriptions, not against any single one in isolation.
When Not to Choose It
Skip PostHog's usage-based plan, or at minimum negotiate the Enterprise tier instead, when:
- Finance needs a fixed, predictable monthly number and cannot tolerate a bill that moves with product growth — a flat per-seat competitor may be the better budgeting fit even if it costs more at low usage.
- The team plans to record session replay broadly (near 100% of sessions) rather than sampling — replay's per-unit price is roughly 30-100x the analytics event price, and unsampled replay is the single fastest way to a five-figure surprise.
- Nobody on the team owns billing-limit configuration. Without a limit set, usage keeps billing past free tiers indefinitely; with a limit set carelessly low, PostHog stops ingesting data entirely once the cap hits, which can silently create analytics gaps during a traffic spike.
Keeping the Bill Predictable: Billing Limits Are Not Optional
PostHog's own billing limits documentation is direct about the mechanic that matters most here: billing limits can be set per product, and when a limit is hit, PostHog "will stop ingesting and processing your data so you are not charged over the set limit." That is a hard stop, not a throttle — data past the limit is lost, not queued or degraded. The organization owner also receives automatic email alerts at 80% and 100% of any configured limit, which gives a real window to react before data loss.
The practical implication: a team that never sets a billing limit is implicitly opting into unlimited spend as usage grows, while a team that sets limits without watching the 80% alert is implicitly opting into data loss during its highest-traffic days — often the worst possible time for a launch or campaign spike. The correct sequence for any team past the 1-seat/early-beta scale in this guide is to (1) estimate a monthly ceiling using the scenarios above, (2) set billing limits at roughly 1.5x that ceiling per product, and (3) treat the 80% alert as the trigger to review usage, not the 100% hard stop.
Bottom Line
PostHog's pricing page tells the truth about mechanics — usage-based, unbundled, unlimited seats — but it does not tell a team what its own product will actually cost, because that number does not exist until real usage data exists. The rate cards above make that number computable in advance instead of discovered on the first invoice. For a team below roughly 1M events, 5K replays, and 1M flag requests a month, PostHog is close to free. Past that point, the bill scales with product growth roughly the way infrastructure costs do — which is the correct comparison to make, and the reason billing limits belong in the initial setup checklist, not the incident retrospective after the first surprise invoice.
Related ToolPick Decisions
- PostHog Review 2026: Product Analytics, Feature Flags, and Session Replay Fit — The feature and workflow comparison against Amplitude, Mixpanel, and Google Analytics that this cost model complements.
- Best Product Analytics Tools in 2026 — See how PostHog's usage-based model compares to seat-and-MTU-capped competitors across the category.
- LaunchDarkly Alternatives in 2026: Statsig, PostHog, GrowthBook, and ConfigCat — If the feature-flags line item above is the concern, compare PostHog's flag pricing against dedicated flag platforms.
USD 1 Decision Check
Need one outside check before you choose?
Use Polar checkout, then email one narrow buying question. ToolPick replies with one practical buyer-risk answer and the next action: buy, trial, delay, compare, or ask a vendor.
No subscription. Manual email answer. If the question is outside one software decision, ToolPick will say so before work starts.
Vendor? Run free check- PayUse Polar checkout. No subscription.
- SendEmail help@neogenesis.app with the order ID and one narrow question.
- ReceiveGet one manual buyer-risk answer by email and the next action.
- One concrete tool, pricing, stack, or alternative question.
- One concise buyer-risk answer by email after checkout handoff.
- One next step: buy, trial, delay, compare, or ask a vendor.
Frequently Asked Questions
Does PostHog charge more as a team adds people?
No. PostHog does not meter by seat on any current plan — unlimited team members are included from the free tier up. The bill moves with product usage (events, replays, flag requests, survey responses, rows), which usually grows with the user base a team is serving, not with the number of people on the team itself.
What is the single biggest lever for controlling the PostHog bill?
Session replay sampling. Recording every session is the fastest way to blow past the free 5,000-recording allowance, and recording price per unit is far higher than the event price per unit. Most teams that keep bills low sample replay at 5-20% of sessions instead of 100%.
Can a team accidentally get a surprise five-figure invoice?
Only if billing limits are left unset. PostHog's default behavior without a billing limit is to keep ingesting and billing usage past free tiers. With a billing limit set, PostHog stops ingesting data once the cap is hit rather than over-charging — the trade-off is lost data past the cap, not a runaway bill. Set limits before enabling replay or flags at scale, not after the first invoice surprises you.
Send it to a teammate or save it for the next renewal check.
📬 Get the Weekly SaaS Digest
New tool reviews, pricing-change alerts, and stack cost tips — one email a week, one-click unsubscribe. No spam, no fake urgency.
Turn this article into a decision path
Every ToolPick article should lead to a second useful page: another article, a hub, or a calculator action.
AI-Powered Cybersecurity for SaaS ApplicationsRead the next related article.Related Articles
AI-Powered Cybersecurity for SaaS Applications
AI-Powered Cybersecurity for SaaS Applications compared for SaaS pricing and budget fit: pricing fit, workflow trade-offs, setup risk, and when to choose each option.
AI-Powered Security for Remote Development Teams
AI-Powered Security for Remote Development Teams compared for SaaS pricing and budget fit: pricing fit, workflow trade-offs, setup risk, and when to choose each option.
Notion Review 2026: Best All-in-One Workspace for Solo Devs?
Solo dev deep dive: free tier limits, real costs for 1-10 users, and honest warnings about when Notion is NOT the right pick.