Statsig Pricing and Statsig Cost Per Event From the Free Tier to Enterprise After the Amplitude Deal

Statsig costs $0 on its Developer tier for 2 million metered events a month, and $150 a month on Pro for 5 million events, then $0.05 per 1,000 events after that. Feature flag and config checks are free at any volume. Enterprise is quoted, and Vendr reports a median Statsig contract of $82,500 a year across 87 purchases. Statsig has been part of Amplitude since 5 May 2026, and its price list has not changed. Read from a US connection on 1 October 2026.

How much does Statsig cost?

Statsig publishes three plans and prices two of them. Every plan includes experimentation, feature flags, product analytics and session replay, and every plan has unlimited seats, so you never pay per person. The meter is metered events, which we cover in detail below.

Session replays are kept for 30 days on Developer and Pro, and that retention is custom on Enterprise. Support on Developer is the community Slack. Pro gets a 4 business hour response, and Enterprise adds a 99.95% service SLA and a private Slack channel with an account team.

What does Statsig Pro cost at your event volume?

The Pro overage works out to $50 per extra million events, with no volume tiers on the published plan. That makes the annual cost easy to model, and it also means the per-event price never falls on Pro however much you send. Volume discounts only exist on Enterprise.

Notice the last column. Because the $150 base includes 5 million events, Pro is cheapest per event at exactly 5 million, and the effective rate climbs toward $0.05 as you grow. Most metered tools get cheaper per unit with volume. Statsig Pro gets slightly more expensive, which is the strongest argument for asking about Enterprise once you pass 20 to 30 million events a month.

Two terms on the pricing page change the bill. Each Pro subscription applies to one project only, so a company running separate projects for web and mobile, or for two products, pays $150 per project and gets 5 million included events in each. And overages are billed in arrears: the $150 base is charged on the first day of each billing period, and any overage from that month is added to the next bill. Your first overage shows up a month after the traffic that caused it.

What are Statsig metered events?

Statsig bills four kinds of event and nothing else. The definitions are on its pricing page, and they decide whether your quote is accurate:

• Exposures, when your app asks the SDK whether a user should see an experiment, or a feature flag that has Metric Lifts switched on.
• Log events, every event you send through the SDK, plus events imported from a third-party source such as Segment.
• Ingested metrics, pre-computed metrics loaded from your data warehouse.
• Custom metrics, when existing events or metrics are combined into a new metric.

Three rules work in your favor. Flags at a 0% or 100% rollout, or with Metric Lifts off, generate no exposures at all, and finished or unstarted experiments do not either. Exposures are deduplicated per user and per flag or experiment, over a 60 minute window on client SDKs and one minute on server SDKs. And plain flag and config checks are unlimited on every plan.

One rule works against you. Each log event can carry one event dimension, the Value field, for free. Statsig charges an additional log event for every additional dimension you attach. A tracking plan that sends three extra properties on every event can bill as four events per action, so check this before you estimate from your current event count.

What happens if you exceed the Statsig free tier?

Nothing is cut off suddenly. Above 2 million events a month on Developer, Statsig prompts you to add a card and move to Pro, and sends several warnings first. If you stay on Developer, feature gates and dynamic configs keep working, but Metric Lift computations stop, which means your experiments keep running and you lose the results. For a team that relies on Statsig to decide what ships, that is the practical limit of the free plan.

What do companies actually pay for Statsig?

Vendr, which negotiates software purchases and publishes aggregate contract data, puts the median Statsig contract at $82,500 a year, based on 87 purchases, with a low of $14,650 and a high of $199,250. Its buyers report saving 27.69% on average against the first quote. These are third-party figures, not Statsig's.

The useful number is what $82,500 buys at list price. At Pro rates it covers roughly 140 million metered events a month. The median buyer is therefore a large-volume team on an Enterprise contract, and the $150 entry price describes a very different customer.

One caution. Vendr's own write-up describes Statsig as priced on monthly tracked users and feature seats. Statsig's pricing page says the opposite on both counts: it bills metered events and includes unlimited seats on every plan. Use Vendr for contract values and discount levels, and Statsig's page for how the bill is built.

Is Statsig owned by Amplitude now?

Yes. Statsig joined the Amplitude family on 5 May 2026, and its homepage now carries the line "Statsig is part of the Amplitude family". In a post dated 17 June 2026, Amplitude's VP of product wrote that a team of about 50 Amplitude staff maintains the platform and supports its customers, and that the original Statsig team is now at OpenAI. The stated priorities include more granular permissions, release pipelines for experiments, and evals and LLM traces.

The pricing page has not changed with the owner. Developer, Pro and Enterprise are sold exactly as before, with the same allowances. For a buyer that creates a real negotiating position, because Amplitude now sells experimentation on two price lists. Amplitude's own plan table lists feature A/B testing and experiment results as not included on its Free and Plus plans, so the closest Amplitude equivalent to Statsig Pro is a quoted Growth contract. Our Statsig vs PostHog vs Amplitude pricing comparison prices all three at 5 to 100 million events a month.

If you are signing a multi-year Statsig contract this quarter, ask in writing which roadmap the contract follows, whether the Enterprise rate is fixed for the full term, and what migration support applies if the two products are merged.

Statsig vs PostHog and Amplitude on price

PostHog is cheaper than Statsig for anonymous traffic and much more expensive once events are tied to known users, because identified events are a second meter. The full PostHog ladders are in our PostHog pricing breakdown, and the Amplitude annual ladder is in our Amplitude pricing guide. If you also need Mixpanel's numbers, the Amplitude vs Mixpanel vs PostHog pricing page sets all three side by side.

Is Statsig worth it?

For a product and engineering team that ships behind flags and runs several experiments a month, the free tier and the $150 Pro plan are among the best value in the category. You get a serious stats engine, unlimited seats and free flag checks, and the published rates let you forecast cost without a sales call. Feature flags also separate deploying code from releasing it, which pairs well with a zero-downtime deployment tool when you want each release to be both safe to ship and safe to roll back.

It is a weaker fit if your tracking plan sends many properties per event, because of the per-dimension billing, or if you need several separate projects on Pro. And above roughly 30 million events a month the flat Pro rate stops being a bargain, so plan on an Enterprise negotiation.

How to negotiate a Statsig contract

Pull one month of event counts from your current tool and recount them the way Statsig does: every extra property dimension as another event, and exposures only for partial rollouts with Metric Lifts on. That number, not your raw event count, is what your quote should be built on.

Then use the anchors that exist. Vendr's 27.69% average saving says the first Enterprise quote has room in it. PostHog's billing page prints every rate and gives you a checkable alternative. And Amplitude now owns Statsig, so an Amplitude quote for the same volume is a fair comparison to put on the table. Ask for the per-event rate at your volume, the rate if volume doubles mid-term, and whether warehouse-native deployment changes the price.

What Statsig will not tell you about why users churn

Statsig tells you which variant won and what users did. It does not read the support tickets, app reviews and survey comments where customers say what went wrong. UserInsight groups that feedback into themes, counts the customers behind each one, and ties each theme to product usage, so a losing experiment or a drop in retention comes with the reason attached. It runs next to Statsig on one plan price that does not change with your event volume.