Mixpanel

Name: Mixpanel
Brand: Mixpanel
Price: 010 USD

Event analytics that a non-analyst can actually drive

Mixpanel is an event-based product analytics platform that tracks what users do inside a web or mobile app and answers questions about funnels, retention, feature adoption, and cohorts through a self-serve query interface, with session replay, warehouse connectors, and a free tier of 1 million events per month; it bills on event volume rather than seats, so the whole team can have logins.

Visit website

Overview

Mixpanel is the oldest surviving name in product analytics and, for a small team, still the easiest one to get value out of in an afternoon. It was founded in 2009 out of Y Combinator by Suhail Doshi and Tim Trefren on a single argument that has aged well: page views tell you nothing about a product, so track events instead. Everything in the interface follows from that. You define events, you group them into funnels, you look at retention curves, you break results down by a property, and you do it without writing SQL or filing a ticket with a data team you probably do not have.

The reason Mixpanel keeps showing up on small-company shortlists is the pricing meter. Billing is on events ingested, not on seats and not on monthly tracked users, which means a five-person startup can give logins to everyone including the founder who checks numbers on Sunday night and pay nothing extra. The free plan covers 1 million events and 10,000 session replays per month, which is a genuinely usable product rather than a demo, and the startup program makes the first year free for companies under five years old with less than $8M raised. That combination is why Mixpanel is often the first analytics tool a company installs and the one it stays on until either the event bill or the data team gets big.

The last few years have widened the product beyond charts. Session replay arrived across web, iOS, Android, React Native, and Electron and is now tied to the same event data, so you can jump from a funnel drop-off straight to watching the person drop off. Autocapture landed for web via the JavaScript SDK, giving teams a no-instrumentation on-ramp. Warehouse connectors let Mixpanel read modeled tables out of Snowflake, BigQuery, Redshift, and Databricks instead of demanding that everything arrive through the SDK. In October 2025 Mixpanel acquired DoubleLoop, a product strategy and impact mapping tool, pushing the platform toward tying shipped work to metric movement.

The catch, and it is a real one, is that Mixpanel does not do the things adjacent to analytics. There are no feature flags, no experimentation engine, no in-app guides, no error monitoring. If you want an integrated suite where the flag that gated the feature and the chart that measured it are the same system, PostHog and Statsig are built that way and Mixpanel is not. Mixpanel's answer is that it does one job unusually well and connects to everything else, which is a defensible position but does mean a second and third vendor on the invoice.

Best for

Product and growth teams at startups and small software companies who want funnels, retention, and cohort analysis that a product manager can drive without SQL, who value event-based rather than seat-based billing so everyone gets a login, and who are happy to buy flags, experiments, and error monitoring from someone else.

Not the right fit for

Teams who want one vendor for analytics, feature flags, experiments, and error tracking; Mixpanel ships none of those, so an integrated buyer should look at PostHog or Statsig instead of assembling three invoices.
Content sites and marketing pages measuring page views and traffic sources; Mixpanel is expensive and awkward for that job compared with Plausible, Fathom Analytics, or Matomo, and its event meter punishes high-volume anonymous traffic.
Companies with no engineering capacity to instrument events; Autocapture softens this for web only, and a mobile app with no developer time will produce an empty Mixpanel project.
Buyers who need session replay as the primary product; Mixpanel's replay is a companion to the charts, with a 30 day default retention window, and it does not approach LogRocket or FullStory in debugging depth.
Organizations that need strict EU data residency retrofitted to an existing project; residency is chosen when a project is created and is not something you can flip on later without a migration.

How it works

1
You install an SDK (JavaScript, iOS, Android, React Native, Python, Node, and others) or route events through a CDP like Segment or RudderStack. Each event is a named action with properties attached, and each user gets an identity so events can be stitched into a lifetime. For web you can switch on Autocapture instead, which collects page views, scrolls, form interactions, clicks, dead clicks, and rage clicks with no per-event code, at the cost of a larger and noisier event bill.
2
Events land in projects. Lexicon, the data governance layer, is where you name and describe events and properties, hide the junk, and stop the schema turning into 400 half-duplicated event names within a year. Teams that skip this step are the ones who conclude eighteen months later that their analytics are untrustworthy, and the fix is retroactive and unpleasant.
3
Analysis happens in a small set of report types that cover most real questions: Insights for arbitrary event trends and breakdowns, Funnels for conversion between steps, Retention for whether people come back, Flows for the paths users actually take, and Impact for whether a launch moved a metric. Results save into Boards, which is what most of the company will actually look at.
4
Session Replay runs alongside, sampled at a rate you choose, so a replay can be opened from any user in any report. Replays are retained for 30 days from ingestion by default, with longer custom windows available on paid arrangements. Outbound, Mixpanel syncs cohorts to advertising and messaging tools and can pipe raw events to a warehouse or cloud storage, while Warehouse Connectors run the other direction and mirror warehouse tables into Mixpanel.

Feature breakdown

28 features in 5 modules

Core analysis reports

The five report types that answer most product questions without SQL.

Insights: Arbitrary event trends with breakdowns by any property, custom formulas across events, and comparison against previous periods. This is the workhorse report and where most saved charts originate.
Funnels: Multi-step conversion analysis with configurable conversion windows, step ordering rules, and breakdowns to see which segment fails at which step. Drop-off cells link directly to the users and their replays.
Retention: Classic cohort retention curves plus addiction and frequency views, answering whether a given behaviour in week one predicts survival in week eight, which is the question most product teams should be asking first.
Flows: Path analysis showing what users actually did before or after a chosen event, useful for finding the unplanned route through your product that half your users take.
Impact: Measures whether a feature launch or campaign moved a metric, comparing behaviour before and after adoption rather than requiring a formal experiment.
Boards: Saved reports assembled into shareable dashboards with text annotations, scheduled digests, and public sharing links, which is how analytics reaches people who never open the query builder.

Data collection and governance

How events get in and how you stop the schema rotting.

Autocapture for web: The JavaScript SDK can collect page views, scrolls, form interactions, element clicks, dead clicks, and rage clicks with no manual instrumentation. It is opt-in, available on all plans, and sensitive elements such as text inputs, selects, and textareas are excluded by default.
Manual event instrumentation: The recommended path for anything that matters. You name events and attach properties deliberately, which costs engineering time up front and produces analytics that survive contact with a real question two years later.
Lexicon data dictionary: Central catalogue for event and property names with descriptions, ownership, and the ability to hide or drop events. The single highest-leverage hour of setup in the whole product.
Identity management: Anonymous activity is stitched to a known user on signup or login, with identify, alias, and merge calls explicitly excluded from event billing so identity plumbing does not cost you money.
Warehouse Connectors: Mixpanel reads modeled tables out of Snowflake, BigQuery, Redshift, and Databricks, so subscription and billing facts that already live in the warehouse can be analysed alongside SDK events without re-instrumenting them.
Server-side and CDP ingestion: Server SDKs and a raw HTTP ingestion API, plus first-class support for Segment, RudderStack, and mParticle, which is the usual route for teams who do not want a vendor SDK in the client bundle.

Session replay

Watching the drop-off you just found in a funnel.

Cross-platform capture: Replay works on web, iOS, Android, React Native, and Electron desktop apps, with capture also available through Segment and mParticle integrations.
Event-linked playback: Replays are attached to the same user and event stream as the charts, so any funnel drop-off, cohort, or user profile becomes a list of sessions you can watch rather than a number you have to interpret.
Sampling and conditional capture: You set a percentage sampling rate or trigger and end capture manually from code, and UI controls let you adjust the baseline rate and conditional rules without shipping a release.
Privacy masking: Sensitive fields can be masked or excluded, and Mixpanel's guidance is explicit that checkout and payment screens should be protected deliberately rather than assumed safe.
Replay retention controls: Replays are kept 30 days from ingestion by default; paid arrangements allow custom retention configured between 7 and 360 days for teams with longer investigation cycles or shorter compliance windows.

Activation and outbound data

Getting the analysis back out into the tools that act on it.

Cohort sync: Behavioural cohorts computed in Mixpanel push into messaging, advertising, and CRM destinations so a segment defined by product behaviour can be targeted without rebuilding the logic in each tool.
Data pipelines and raw export: Raw events can be exported to cloud storage or a warehouse on a schedule, which is the escape hatch that keeps Mixpanel from being a data roach motel. Historically packaged as a paid add-on rather than included.
Query and export APIs: Documented APIs for ingestion, export, cohort management, and running saved queries programmatically, so reporting can be embedded elsewhere or piped into internal tooling.
Alerts and scheduled digests: Threshold and anomaly alerts on saved reports, delivered by email or Slack, so a broken signup funnel surfaces on its own rather than when someone happens to open a board.
DoubleLoop strategy mapping: Following the October 2025 acquisition of DoubleLoop, Mixpanel is folding in the ability to map shipped work to metric movement, connecting the roadmap to the charts rather than leaving that translation to a slide deck.

Administration, privacy, and scale

The parts a buyer checks after the product team has already fallen in love.

Unlimited seats on every plan: Because billing meters events rather than users, adding the whole company as viewers costs nothing, which materially changes whether analytics gets used or stays with two people.
EU Data Residency Program: Projects can be created against EU infrastructure with dedicated endpoints such as api-eu.mixpanel.com, keeping personal data processing in Europe. The choice is made at project creation, not retrofitted.
Two year default event retention: Since September 2025 events are deleted automatically after two years by default across projects, a change made in the name of data minimisation that also quietly limits very long historical analysis.
GDPR and CCPA tooling: Documented deletion and retrieval endpoints for data subject requests, a published DPA, and opt-out controls in the client SDKs.
Projects, roles, and access control: Separate projects for production and development, role-based permissions on boards and reports, and organisation-level user management for teams that need viewers separated from editors.
SSO and SAML: Single sign-on with SAML is available on enterprise arrangements; Google and Microsoft login cover smaller teams without a contract conversation.

Use cases

4 documented

Seed-stage SaaS founder measuring activation

Signups are growing but nobody knows how many new accounts reach the moment the product becomes useful, and there is no data team to ask.

Six deliberate events cover signup, first project, first invite, and first successful run. A funnel shows 62 percent of accounts never reach the second step, a retention curve confirms those accounts churn, and the free 1 million event tier covers the whole thing at zero cost.

Product manager at a 40-person company

A new feature shipped last quarter and leadership wants to know whether it did anything, but the only evidence is anecdotal support tickets.

An Impact report compares behaviour of adopters against non-adopters, Insights breaks adoption down by plan tier, and Session Replay shows why the second-most-common workflow abandons midway. The answer arrives in a day rather than a sprint.

Growth team running lifecycle messaging

Onboarding emails go to everyone on the same schedule regardless of what the user has actually done in the product.

Behavioural cohorts defined in Mixpanel (signed up, never invited a teammate, active in the last 7 days) sync into the messaging tool, so campaigns fire on product behaviour instead of days since registration.

Analytics-minded engineer at a company with a warehouse

Revenue and subscription facts live in Snowflake, product events live in an SDK, and reconciling the two means exporting CSVs.

Warehouse Connectors mirror the modeled Snowflake tables into Mixpanel so plan tier and MRR band become breakdown properties on every product chart, while data pipelines push raw events back out for the SQL work that Mixpanel cannot do.

Pricing

from $0 (Free plan, 1M events per month), Growth starts at $0 and scales with volume

Usage-based on events ingested per month, with unlimited seats on every plan and session replay metered separately. Three plans: Free, Growth (self-serve with volume discounts), and Enterprise (quoted, optionally metered on monthly tracked users instead of events).

Plan	Price	Includes
Free	$0 per month	Up to 1 million events per month 10,000 session replays per month Unlimited seats and unlimited saved reports Insights, Funnels, Retention, and Flows Autocapture, Lexicon, and the ingestion APIs One of the most genuinely usable free tiers in the category; a small B2B app can live here for a year or more.
Growth	Starts at $0 per month, scaling with event volume	Up to 20 million events per month Up to 500,000 session replays per month Volume discounts as committed volume rises Data governance, group analytics, and advanced report types Self-serve checkout with no sales call Growth plans purchased after February 2025 include the first 1 million events free; the published overage rate at the smallest commitment is $0.00028 per event, falling with volume.
Enterprise	Quoted, from roughly $20,000 per year on MTU-based plans annual contract	Up to 1 trillion events per month Customisable session replay volumes and retention between 7 and 360 days Automated data governance and access controls SSO and SAML, dedicated support Optional monthly-tracked-user billing instead of events The only tier that requires a conversation, and the only one where the meter can be switched away from raw events.

Plan

Price

Includes

Free

per month

Up to 1 million events per month
10,000 session replays per month
Unlimited seats and unlimited saved reports
Insights, Funnels, Retention, and Flows
Autocapture, Lexicon, and the ingestion APIs

One of the most genuinely usable free tiers in the category; a small B2B app can live here for a year or more.

Growth

Starts at $0

per month, scaling with event volume

Up to 20 million events per month
Up to 500,000 session replays per month
Volume discounts as committed volume rises
Data governance, group analytics, and advanced report types
Self-serve checkout with no sales call

Growth plans purchased after February 2025 include the first 1 million events free; the published overage rate at the smallest commitment is $0.00028 per event, falling with volume.

Enterprise

Quoted, from roughly $20,000 per year on MTU-based plans

annual contract

Up to 1 trillion events per month
Customisable session replay volumes and retention between 7 and 360 days
Automated data governance and access controls
SSO and SAML, dedicated support
Optional monthly-tracked-user billing instead of events

The only tier that requires a conversation, and the only one where the meter can be switched away from raw events.

Add-ons

Session replay above plan allowance (Metered per replay beyond the included volume): Free includes 10,000 per month; documentation describes 20,000 included on paid plans with additional volume billed, while the pricing page shows Growth scaling to 500,000. Model your own sampling rate rather than assuming.
Data pipelines (raw event export) (Quoted): Continuous export of raw events to a warehouse or cloud storage; historically an Enterprise-side add-on rather than included in Growth.
Extended replay retention (Quoted): Custom windows between 7 and 360 days replacing the 30 day default.

Billing notes

The meter is events ingested, not seats and not monthly tracked users, so headcount growth is free and product usage growth is what costs money. Enterprise contracts can optionally switch to an MTU meter starting around $20,000 per year.
Not everything counts: identify, alias, merge, and opt-in calls, session recording checkpoints, and user profile updates are excluded from billing, which removes the usual identity-plumbing tax.
Autocapture is the single fastest way to blow up an event bill. Scroll and click events on a busy marketing site can dwarf the deliberate product events you actually analyse, so enable it on a sandbox first and watch the volume.
Events are deleted after two years by default under the retention policy that took effect in September 2025, so year-over-three analysis is not something to assume.
The startup program gives a free first year to companies founded within the last five years with under $8M raised, which for many small businesses removes the cost question entirely for the evaluation period.
Volume discounts on Growth are real and steep at higher commitments; the published per-event overage rate at the smallest tier is a ceiling, not a fair estimate of what a 20 million event account pays.

Value assessment: For a small business the value case is unusually clean. At 10,000 monthly users doing perhaps thirty deliberate events each, you are at 300,000 events and the free plan covers you outright with unlimited seats. At 100,000 monthly users on the same instrumentation discipline you are around 3 million events, which lands in the middle of Growth and, at published entry rates before volume discounting, is a few hundred to roughly a thousand dollars a month depending on how the commitment is structured. The number that decides your bill is not your user count, it is how disciplined your instrumentation is: a team with forty well-chosen events pays a fraction of what an Autocapture-everything team pays at identical traffic. Judged against seat-priced competitors, Mixpanel is cheap for wide access and fair for careful teams, and expensive only for teams who never bothered to decide what to track.

Strengths & limitations

Strengths

The best self-serve analysis interface in the category for non-technical users; a product manager can build a funnel with a breakdown in minutes without SQL or a data team.
Event-based billing with unlimited seats means analytics can be genuinely company-wide rather than rationed to whoever holds the licences.
A free tier of 1 million events and 10,000 replays per month that is a working product, plus a startup program that makes the first year free for young, lightly funded companies.
Session replay is properly integrated with the event data across web, iOS, Android, React Native, and Electron, so a funnel drop-off becomes a video rather than a hypothesis.
Warehouse Connectors and raw export in both directions, which is unusually good data portability for a hosted analytics vendor and reduces lock-in materially.
Sixteen years of continuous operation, a large integration surface, and documentation and SDK coverage that reflect that maturity.
EU data residency available as a first-class program rather than an enterprise-only afterthought.

Limitations

No feature flags, no experimentation engine, no in-app guides, and no error monitoring, so an integrated product stack means three or four vendors instead of one.
The event meter punishes autocapture and high-volume anonymous traffic, and the cost model is genuinely hard to forecast before you have shipped instrumentation and watched real volume for a month.
Session replay is a companion feature rather than a debugging platform; the 30 day default retention and the absence of deep network, console, and error inspection put it well behind LogRocket.
Default event retention is now two years, which quietly caps long-range historical analysis unless negotiated.
EU residency must be chosen when a project is created; retrofitting it to an existing project means a migration rather than a settings toggle.
Data quality is entirely on you. Without deliberate use of Lexicon, the event schema degrades into hundreds of near-duplicate names within a couple of years, and no amount of interface polish rescues that.
Pricing above the free tier is presented through a calculator rather than a rate card, so comparing Mixpanel with a fixed-price competitor requires actually modelling your own volume.

Head-to-head comparisons

4 alternatives

Mixpanel vs Amplitude

from $0 (Free, 2M events per month forever); Plus starts at $0 with the first 2M events free

These two are the closest substitutes in the category and both meter on events with generous free tiers (1 million for Mixpanel, 2 million for Amplitude). Amplitude is the broader platform, bundling feature flags, experiments, guides, surveys, and a CDP, and it is a public company selling upward into enterprise. Mixpanel is the faster, simpler analysis tool with better ad-hoc ergonomics and less to configure. Choose Amplitude if you want flags and experiments on the same bill and expect to grow into a formal analytics function; choose Mixpanel if you want the charts to be excellent and are content to buy the rest elsewhere.

Full Mixpanel vs Amplitude comparison

Mixpanel vs PostHog

from $0 (generous monthly free tiers; pay only past the allowance)

PostHog bundles product analytics, session replay, feature flags, experiments, surveys, error tracking, and a data warehouse into one usage-priced platform with a free tier across every product, and it is open source with an EU cloud. Mixpanel does analytics only, but does it with a friendlier interface for non-technical users. Pick PostHog if you are engineering-led and want one bill for the whole stack; pick Mixpanel if the people asking the questions are product managers and marketers rather than developers.

Full Mixpanel vs PostHog comparison

Mixpanel vs Statsig

from $0 (Developer), then $150 per month (Pro)

Statsig is an experimentation platform that grew analytics, Mixpanel is an analytics platform that never grew experimentation. If your central question is whether variant B beat variant A with statistical rigour, Statsig is built for that and Mixpanel cannot answer it. If your central question is where users fall out of onboarding, Mixpanel is far more pleasant to use. Many teams run both, which is a reasonable outcome given Statsig's free tier includes 2 million events.

Full Mixpanel vs Statsig comparison

Mixpanel vs LogRocket

from $0 (Free, 1,000 sessions per month), Core from around $176 per month at 25,000 sessions

LogRocket is a session replay and frontend debugging platform with analytics attached; Mixpanel is analytics with replay attached. LogRocket meters sessions and captures console logs, network requests, Redux state, and errors, which is what an engineer needs to reproduce a bug. Mixpanel meters events and answers product questions at population scale. Buy LogRocket when the recurring question is why is this broken; buy Mixpanel when it is why are people not converting.

Full Mixpanel vs LogRocket comparison

Implementation & onboarding

Setup time: Under an hour to first chart if you use Autocapture on a web app, or a day or two of engineering to instrument a deliberate event schema across web and mobile. Getting to trustworthy analytics, as opposed to a chart, is a one to two week project because it involves deciding what your product's key actions actually are.
Learning curve: Low for consumers of reports and moderate for builders. The report builder is genuinely learnable by non-technical staff in an afternoon. The harder skill is analytical rather than technical: knowing which funnel is worth building and not chasing a metric that moves for uninteresting reasons.
Onboarding: Fully self-serve including the Growth plan, with in-app guides, a large documentation site, and a sizeable community forum. Enterprise arrangements add named support and implementation help. Nothing about the free or Growth path requires talking to a person.
Migration notes: Historical events can be backfilled through the ingestion API, though very old timestamps have documented import constraints, so plan on a partial rather than complete history import. Coming from Google Analytics 4, the mental model shift from sessions to events is bigger than the technical work. Coming from Amplitude or PostHog, the event and property model maps closely and the real work is renaming to a schema you actually want rather than replicating the old mess. Leaving is comparatively easy because raw export and warehouse pipelines exist.

Platform, API & security

Platforms: Web (JavaScript SDK with optional Autocapture)iOS and Android native SDKsReact Native and FlutterElectron desktop appsServer-side SDKs for Python, Node, Java, Ruby, PHP, and GoHTTP ingestion APIWeb app plus iOS and Android viewer apps
API: Documented ingestion, export, query, cohort, and data-deletion APIs, plus warehouse connectors that read from Snowflake, BigQuery, Redshift, and Databricks and scheduled pipelines that export raw events outward.
Compliance: SOC 2 Type IIGDPR with published DPA and data subject request endpointsCCPAHIPAA available under enterprise arrangements with a BAA
Data residency: US by default, with an EU Data Residency Program offering dedicated European endpoints such as api-eu.mixpanel.com. Residency is selected when a project is created and cannot be switched afterwards without migrating.
SSO: SAML single sign-on on enterprise arrangements; Google and Microsoft social login on all plans.
Security notes: Encryption in transit and at rest, role-based access control across projects and boards, event data deleted automatically after two years by default since September 2025, and session replays retained 30 days by default with configurable windows between 7 and 360 days on paid arrangements.

Support & resources

Channels: In-product chat and email support on paid plansCommunity forum for free usersNamed support and SLAs on enterprise contracts
Documentation: Extensive documentation at docs.mixpanel.com covering SDKs, ingestion, identity management, every report type, session replay, warehouse connectors, and billing mechanics, with changelogs published for pricing and retention changes.
Community: An active public community forum plus Mixpanel-run education content; sixteen years of accumulated third-party tutorials mean most implementation questions have already been answered somewhere.

Company

Founded: 2009
Headquarters: San Francisco, California, United States
Ownership: Venture-backed, privately held
Founders: Suhail Doshi, Tim Trefren
Employees: Not published by the company; third-party trackers estimated several hundred employees in 2026
Funding: Approximately $277M raised across six rounds, the largest being a $200M Series C in November 2021 that valued the company at over $1B.

Funding history

Round	Amount	Year	Notes
Y Combinator	Seed	2009	Summer 2009 batch; one of the earliest analytics companies to come out of YC.
Series A and B	Included in the $277M total	2012 to 2014	Early growth rounds with Andreessen Horowitz among the backers.
Series C	$200M	2021	Led by Bain Capital Tech Opportunities with Sequoia participating, at a valuation above $1B.

Timeline

2009Founded by Suhail Doshi and Tim Trefren in Y Combinator's Summer 2009 batch, built on the argument that page views are the wrong unit and events are the right one.
2014Becomes the default product analytics choice for mobile-first startups as app companies discover that web analytics tools cannot answer app questions.
2019Amir Movafaghi becomes chief executive, beginning a repositioning of the company toward larger customers and a more disciplined product.
2021Raises a $200M Series C at a valuation above $1B and rebuilds pricing around event volume with unlimited seats, a decision that defines how small teams buy the product.
2023Session Replay launches, later extending across iOS, Android, React Native, and Electron, and Warehouse Connectors bring modeled warehouse tables into the analysis layer.
2024Autocapture arrives for web via the JavaScript SDK, giving teams a zero-instrumentation on-ramp at the cost of a noisier and more expensive event stream.
2025Default event retention moves to two years in a September data minimisation change, and Mixpanel acquires DoubleLoop in October to connect shipped work to metric movement.

Integrations

Segment, RudderStack, and mParticle
Snowflake, BigQuery, Redshift, and Databricks via Warehouse Connectors
Slack and Microsoft Teams for alerts and report sharing
Customer.io, Braze, Iterable, and other messaging tools via cohort sync
Google Ads, Meta, and other ad platforms for behavioural audience targeting
Salesforce and HubSpot
Zapier
Figma, Amazon S3, and Google Cloud Storage for export destinations
Ingestion, export, and query APIs for anything not covered natively

Frequently asked questions

10 questions

What is Mixpanel?

Mixpanel is an event-based product analytics platform. You track named actions users take inside your web or mobile app, then answer questions about conversion funnels, retention, feature adoption, and user segments through a self-serve interface that does not require SQL. It also includes session replay, warehouse connectors, and cohort sync to messaging and advertising tools.

How much does Mixpanel cost?

The free plan covers 1 million events and 10,000 session replays per month with unlimited seats. The Growth plan starts at $0 and scales with event volume up to 20 million events per month, with volume discounts and a published overage rate of $0.00028 per event at the smallest commitment. Enterprise is quoted, and MTU-based enterprise plans start at roughly $20,000 per year. Startups under five years old with less than $8M raised get the first year free.

What counts as a billable event in Mixpanel?

Any tracked user action you send, such as Signed Up or Added to Cart. Several things are explicitly excluded from billing: identify, alias, merge, and opt-in calls, user profile creation and updates, and session recording checkpoints. That exclusion matters because identity plumbing would otherwise inflate your bill without producing anything analysable.

Should I use Autocapture or instrument events manually?

Use Autocapture to get moving on a web app in an afternoon, and instrument manually for anything you actually plan to make decisions with. Autocapture collects page views, scrolls, clicks, form interactions, dead clicks, and rage clicks with no code, but it produces high volume that costs money and generic event names that are hard to reason about later. Mixpanel also advises against it on sites handling protected health information. Most successful implementations use both: autocapture for exploration, deliberate events for the metrics that matter.

How long does Mixpanel keep my data?

Since a change effective September 2025, events are deleted automatically after two years by default across all projects. User profile data is retained for the life of an active subscription. Session replays are kept 30 days from ingestion by default, with custom retention between 7 and 360 days available on paid arrangements. If you need multi-year historical analysis, export raw events to your own warehouse.

Does Mixpanel offer feature flags or A/B testing?

No. Mixpanel has no feature flag system and no experimentation engine. It can measure the effect of a launch through its Impact report, but it cannot randomise users into variants or compute statistical significance for you. Teams that want flags, experiments, and analytics from one vendor should look at PostHog or Statsig; teams already on Mixpanel commonly pair it with a dedicated flagging tool.

Can Mixpanel host my data in the EU?

Yes. The EU Data Residency Program processes and stores personal data in Europe using dedicated endpoints such as api-eu.mixpanel.com, with SDK configuration documented for JavaScript, Python, Node, Android, and Swift among others. The important caveat is that residency is chosen when you create a project and is not a setting you can flip later, so decide before you instrument.

How good is Mixpanel's session replay compared with dedicated tools?

It is good enough to explain a funnel drop-off and not built to debug a production incident. Replays are tied to the event data across web, iOS, Android, React Native, and Electron, with sampling controls and privacy masking, which is exactly what a product manager needs. What it lacks is the deep frontend forensics that LogRocket, FullStory, or OpenReplay provide: console output, network waterfalls, state inspection, and error correlation. If replay is the reason you are buying, buy a replay tool.

Can I get my data out of Mixpanel?

Yes, and this is one of its better qualities. There is a raw export API, scheduled data pipelines that push events to a warehouse or cloud storage, and Warehouse Connectors that run the other direction. Exit costs are genuinely lower than with analytics tools that treat your event history as a retention mechanism, though pipelines have historically been packaged as a paid add-on rather than included.

Who owns Mixpanel and is it a stable vendor?

Mixpanel is a privately held venture-backed company founded in 2009 in San Francisco by Suhail Doshi and Tim Trefren, with roughly $277M raised including a $200M Series C in 2021 at a valuation above $1B. It has been operating for sixteen years, is led by chief executive Amir Movafaghi, and acquired DoubleLoop in October 2025. It is one of the more established independent vendors in a category where several competitors have been absorbed into larger platforms.

Editorial verdict

Mixpanel is the safest first analytics purchase a small software company can make. The free tier is a working product rather than a teaser, billing on events with unlimited seats means the whole team gets access, and the analysis interface is the one non-technical people actually use rather than abandon. Buy it if your questions are about conversion, retention, and feature adoption, and if you are comfortable buying flags, experiments, and error monitoring from other vendors. Do not buy it as a session replay tool, do not buy it to measure a marketing site, and do not switch on Autocapture without watching the event meter for a week first. The single thing that determines whether Mixpanel is cheap or expensive is instrumentation discipline, and that is entirely within your control.

Written by the SaaSTracker editorial team. Awards, when shown, are judged against the published criteria in our methodology.