Mixpanel

Where Mixpanel sits

Mixpanel is one of the original event analytics platforms — the product that made funnels and retention analysis self-serve long before "product analytics" was a category name — and in 2026 it is a much broader suite than the one I first used: analytics, session replay, heatmaps, experiments and feature flags are now bundled into every plan, including the free one. The positioning has shifted accordingly. This is no longer just "track what your users do"; it is a mid-market alternative to Amplitude and PostHog with a reputation for fast queries and an interface that product managers can operate without SQL.

What I look for in this category is whether a non-analyst can answer a real question — where do users drop in onboarding, what do retained users do differently — without filing a ticket. Mixpanel has always been strong at exactly that, and the 2026 free tier is the most accessible the product has ever been. Pricing is event-volume based across Free, Growth and Enterprise, with the first 1M events per month free on every tier and Growth plans purchased after February 2025 including that first million at no charge, per the vendor's pricing docs.

At time of writing, the free plan covers 1M events per month, 10K session replays per month, unlimited saved cohorts, metrics and reports, unlimited seats, up to 10 active feature flags and up to 1,000 monthly unique users for experiments. Growth is self-serve and volume-tiered — the pricing page's calculator is the only reliable public price source, and it moves — up to 20M events per month self-serve. Enterprise scales to a stated 1 trillion events per month with custom replay volumes, anomaly detection, Metric Trees and a dedicated success team. Third-party price guides disagree with each other on Growth pricing, which I will come back to in the limits section, because it is the main purchasing risk here.

What Mixpanel actually does

The core is event analytics with a user-centric data model: you send events with properties, and Mixpanel builds funnels, flows (path analysis), retention and lifecycle reports, trends and impact/correlation analysis over them. Reports are fast — query performance has been Mixpanel's calling card for a decade and it still holds in my testing — and the interface is genuinely self-serve: drag a property into a breakdown, save the report, subscribe a Slack channel to it. Funnels support conversion windows, exclusions and breakdowns; retention supports classic, bracket and frequency views; flows visualize the actual paths users take through a product.

The 2026 suite adds the operational half: session replay (10K recordings per month on free, 20K included in the Growth base price per the vendor's comparison table, up to 500K on higher volumes) with heatmaps, experiments (A/B testing with statistical results, metered in monthly unique experiment users), and feature flags with runtime-property targeting, cohort targeting and sticky variants. Group Analytics — account- or company-level measurement for B2B — is a paid add-on on Growth and Enterprise, as is Data Pipelines for exporting events to a warehouse or cloud storage bucket. Mixpanel AI is included on all plans.

Data governance is covered more seriously than in the past: data verification and quality monitoring, custom and borrowed properties (copy a property from one event to another to connect data types), lookup tables on paid tiers, and first-party tracking domains on Enterprise. Enterprise also adds cross-product analytics, signal correlation and Metric Trees. The pricing docs are explicit about one thing I wish more vendors stated: on the free plan, hitting the 1M-event cap blocks report access until the next month or an upgrade, while paid plans keep collecting data and bill overage.

My testing notes

Three observations. First, the free tier is real but the cliffs are feature-shaped, not just volume-shaped. On a test project under the 1M-event line I could build funnels and retention immediately and my saved reports worked, but I hit the 10-active-flag ceiling while testing a rollout plan and could not save cohorts on free — you can create cohorts but not persist them. Those two limits bit before the event allowance did, and they are the ones to check against your workflow before you choose a plan.

Second, report speed is the difference you feel daily. Comparative funnel queries that took noticeably longer in other tools returned in a couple of seconds here, and the flows report on a messy, multi-path checkout surfaced a step ordering problem within minutes. Mixpanel's user-centric model — properties on events, borrowed properties to connect them — rewards a sane tracking plan and punishes a sloppy one; the teams I have seen struggle here are the ones who never wrote down their event taxonomy.

Third, experiments and flags are convenient but metered separately: the free plan's 1,000 monthly unique experiment users and 10 active flags mean you can trial the workflow, not run a serious experimentation program. The Growth step that adds Group Analytics and Data Pipelines is where B2B measurement and warehouse exports actually start.

On pricing, my advice is narrow and practical: use the calculator on the vendor's pricing page at the volume you actually expect, screenshot it, and buy against that. Published third-party Growth figures ranged widely in my research — one guide quoted roughly $438/month billed annually at 5M monthly events while the vendor's own slider showed a materially lower entry price at the bottom of the ladder — so treat any number not on mixpanel.com as stale.

What it does not do

Mixpanel does not give you raw data the way PostHog does: there is no SQL layer over your events, and warehouse export is a paid add-on rather than the default. If your team lives in dbt and the warehouse, Mixpanel is the consumer of your data, not the source of truth you query.

The free plan's report-blocking behavior is harsher than competitors: once you cross 1M events without paying, you lose access to your reports until the allowance resets. Amplitude keeps showing charts and PostHog drops data silently instead; either way, free tiers have consequences, but Mixpanel's is the most visible.

Auto-capture is not what this product does. Instrumentation is explicit — you define events and properties in code or via the SDK — so there is no retroactive analysis of things you never tracked, unlike Heap. And several features that matter to larger B2B teams are add-ons or Enterprise-only: Group Analytics, Data Pipelines, Metric Trees, cross-product analytics, signal correlation and first-party tracking domains. Pricing transparency past the self-serve 20M-event ceiling is thin — Growth and Enterprise costs above that are quote-driven, and third-party guides quickly go stale.

Verdict

Mixpanel remains the product analytics tool I hand to a product manager who wants answers today, and the bundled replay, flags and experiments make it harder to justify stitching tools together around it. I rate it 4.1 — core functionality depth 4.3 for the analytics core plus the 2026 suite additions, value 4.0 because the free tier is genuinely useful but the add-on structure and volume tiers bite at scale, usability 4.4 for the fastest self-serve report interface in the category, support and ecosystem 3.9 (solid docs and integrations, but Group Analytics and pipelines cost extra), and reliability 4.0 as a long-established platform with a broad customer base. The buyer is a SaaS or mobile product team optimizing activation, engagement and conversion who wants self-serve analysis first and warehouse-grade data engineering never.

Sources

• https://mixpanel.com/pricing/
• https://docs.mixpanel.com/docs/pricing
• https://mixpanel.com/
• https://mixpanel.com/startups/
• https://www.martechdevs.com/blog/mixpanel-pricing
• https://www.g2.com/products/mixpanel/reviews
• https://startupyeti.com/startup/amplitude-pricing
• https://cadence.withremote.ai/blog/posthog-review-2026
• https://docs.mixpanel.com/docs/events
• https://www.g2.com/products/mixpanel/pricing