Introducing the Confidence Free plan and 30-day Growth trial
Unlimited feature flags, unlimited seats and up to 750,000 users in experiments a month, for $0. Or try Growth free for 30 days.

Confidence, Spotify’s product intelligence platform, now has a Free plan: run experiments on up to 750,000 users a month, with unlimited feature flags and unlimited seats, for $0 and no credit card. It also includes analytics, session replays (100K recordings a month), Confidence Agent and a managed warehouse. Teams that need more can try Growth free for 30 days.
It reads every signal in your product, tells you what’s worth building next, and proves it worked with a real, controlled experiment. Now any product builder can start with it for free, whatever the size of their team.
How generous is the Confidence by Spotify Free plan?
Most free plans limit the things you need most. Here’s what the Confidence Free plan includes, compared with each vendor’s public pricing as of November 2026:
750,000 users in experiments a month
Statsig’s free tier counts experiment exposures and logged events against 2M events a month.
Unlimited seats
GrowthBook’s free plan stops at 3 users.
Unlimited feature flags, never metered
PostHog’s free flags stop serving at 1M requests a month.
Experiments with results
Amplitude’s free plan allows 1 active experiment and doesn’t show results.
See the full side-by-side comparison below.
Why we’re making Confidence free to start
AI has made building cheap. A single product builder can go from idea to shipped feature in an afternoon. What hasn’t got cheaper is knowing whether any of it worked. Teams ship constantly, then make the next call from a Slack thread or a dashboard that shows a metric moved but not why.
At Spotify, we’ve spent more than a decade solving that problem for ourselves. Every meaningful change is tested before it reaches hundreds of millions of listeners, and the result decides what ships. We want every product team to work that way from day one, not once they’re big enough to justify a sales conversation. So the Free plan isn’t a cut-down demo. It runs on the same statistical engine, the same guardrails and the same feature flags we use internally.
How Confidence works: from signal to proof
Confidence closes that gap in four steps:
- 01Listen
- 02Decide
- 03Act
- 04Prove
Most tools stop at Listen and Decide: they show you a pattern and leave you to trust it. The Prove step is what sets Confidence apart. It runs on the same experimentation foundation Spotify uses for around 10,000 experiments a year. Sequential testing, guardrail metrics and CUPED variance reduction are on by default, so you can check results any time, catch regressions early, and reach an answer with fewer users. Every plan, including Free, gets the same methodology.
Warehouse-native, or fully managed
Every plan, including Free, comes with Confidence Cloud: a warehouse we set up and run for you, so you can send events and start experimenting without building data infrastructure.
When you’re ready, go warehouse-native from the Starter plan ($129 a month): connect your own Snowflake, BigQuery, Databricks or Redshift, and experiments run on the data you already trust. Define metrics in SQL, reuse your existing data models and keep your warehouse as the source of truth, with no duplicate event tracking. Most experimentation tools keep warehouse-native for enterprise contracts. On Confidence, it’s self-serve.
What’s included in the Free plan?
Every Confidence plan, including Free, comes with:
- Every signal in one place. Product analytics, exploratory analysis and session replays, with up to 100K recordings a month on Free.
- Confidence Agent, which watches your signals and suggests what’s worth building next.
- Unlimited feature flags. Flags and unmonitored rollouts never use credits.
- Unlimited seats. Bring engineers, PMs, designers and data scientists.
- 75,000 credits a month. Up to 750K users in experiments, 5M tracked events, or 250 hours of AI session analysis, or any mix.
- Confidence Cloud, a managed warehouse.
- Google SSO and EU or US hosting.
Who is the Free plan for?
The Free plan is built for the product builder: often an engineer who is also the de facto PM, analyst and data scientist on a small team. If you ship several times a week, already use AI coding tools like Cursor or Claude Code, and want something more rigorous than eyeballing a dashboard, it’s a good fit.
It also works for larger teams that want to evaluate Confidence properly before involving procurement. Because seats are unlimited, you can bring in engineering, product and data colleagues from the start.
How do credits work?
Confidence uses one monthly credit balance across its metered features. You only spend credits on things that cost compute:
| What you use | Credits |
|---|---|
| Users exposed to an experiment | 1 credit per 10 users |
| Tracked events | 15 credits per 1,000 events |
| AI recording analysis | 5 credits per active minute |
| Feature flags and unmonitored rollouts | Free |
We email you at 80% and 100% of your balance. We never silently rate-limit experiments or add surprise overage fees.
What can you do with 75,000 credits?
- Experiments20,000 credits
- Tracked events30,000 credits
- AI analysis24,000 credits
| Example month on Free | Credits used |
|---|---|
| One A/B test exposed to 200,000 users | 20,000 |
| 2 million tracked events | 30,000 |
| 80 hours of AI analysis on session replays | 24,000 |
| Unlimited feature flags and rollouts | 0 |
| Total | 74,000 of 75,000 |
If you’d rather put everything into experimentation, 75,000 credits covers up to 750,000 exposed users a month.
Get set up in minutes
The quickest way in is the Confidence setup wizard. Run one command in your project’s terminal and it installs the SDK and gets you creating flags and experiments in about three minutes.
npx @spotify-confidence/quickstart
If you already use PostHog, Statsig or Eppo, you don’t have to start from scratch. The Confidence plugin for your coding agent reads your existing flags, variants, targeting and rollout percentages, then rewrites your SDK calls one pull request per flag. Confidence is built on OpenFeature, the CNCF standard for feature flags, so once you’ve migrated, your code isn’t locked to any one vendor.
From there, invite your team. Seats are unlimited on every plan, so there’s no reason to wait.
How does the Confidence by Spotify Free plan compare?
Most experimentation tools have a free plan. The difference is what they limit, and Confidence’s Free plan is one of the most generous around. Here’s how the free plans compare, based on each vendor’s public pricing page as of November 2026.
| Confidence | Statsig | PostHog | Amplitude | GrowthBook | |
|---|---|---|---|---|---|
| Seats | Unlimited | Unlimited | Unlimited | Unlimited | Up to 3 |
| Flags | Unlimited | Unlimited | 1M req/mo | Unlimited | Unlimited |
| Experiments | Up to 750K users/mo¹ | In 2M events | Via flags | 1, no results | Unlimited |
| Tracked events | Up to 5M/mo¹ | 2M/mo | 1M/mo | 2M/mo | 1M/mo |
| Session replay | 100K/mo | 50K/mo | 5K/mo | 10K/mo | |
| AI replay analysis | Up to 250 h/mo¹ | 2,500 credits/mo | Growth | ||
| SSO | Enterprise | Add-on | Growth | ||
| EU hosting | Enterprise | Event data |
¹ Confidence figures are maximums from the 75,000 monthly credits, which you can spend on any mix of experiments, events and AI analysis. Units differ between vendors (credits, flag requests, metered events), so compare against your own usage.
Beyond the limits covered above, two things stand out:
- Seats stay free as you grow. GrowthBook’s Pro plan is $40 per seat per month. Confidence has no per-seat fees on any plan.
- Enterprise features at self-serve prices. Google SSO and EU hosting on Free, your own warehouse from Starter, SAML SSO on Growth.
On PostHog’s free plan today? PostHog counts every feature flag request against your free allowance. On Confidence, flags never use credits, and every experiment ends in a real, controlled result. Our migration plugin moves your flags across one pull request per flag. See Confidence by Spotify vs PostHog for the full breakdown.
Which plan is right for you?
| Plan | Price | Best for |
|---|---|---|
| Free | $0 | Developers and small teams getting started |
| Starter | $129/mo | Teams running their first experiments in production |
| Growth | From $449/mo | Product teams shipping fast and learning faster |
| Enterprise | Custom | High-volume programs with custom pricing at scale |
Early-stage startups can also apply for up to $75,000 in credits through our startup program.
When should you move to a paid plan?
You want your own warehouse
Starter ($129/mo) connects Snowflake, BigQuery, Databricks or Redshift.
You hit your credit limit
Starter doubles your credits, and Growth gives you 10 times as many.
Your team is growing
Growth adds SAML SSO, AI-powered reports and reviews, sticky assignments and priority support.
If you’re not sure, start the 30-day Growth trial and see which features your team actually uses.
How does the 30-day Growth trial work?
Growth is for product teams running experiments at scale. During the 30-day trial you get the full plan, including:
- AI-powered reports and reviews
- Bring your own warehouse: Snowflake, BigQuery, Databricks or Redshift
- Sticky assignments and materialized segments
- Insights with Learnings
- SAML SSO (Okta, Entra ID)
- Audit logs (6 months)
- Up to 15 surfaces and 750,000 credits a month
- Priority support
Growth starts at $449 a month, with no per-seat fees, so a team of 5 pays the same as a team of 500. No credit card is needed to start the trial, and when it ends you move back to the Free plan automatically, with nothing to cancel.
Still have questions about the Free plan? See the questions and answers on our pricing page.