Growth

Product Analytics

See what signed-in users do inside the product — which steps they finish, where they drop off, and whether they come back.

Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗

The real choice

All-in-one platform or dedicated analytics tool. All-in-one platforms put event analytics next to session replay, feature flags and experiments, billed per product. Dedicated tools go deeper on funnels, retention and cohorts, with replay and flags added around them. Every option needs events wired into your code.

Pick by situation

Tap the ones that are you — the tools that fit light up below.

You want analytics, session replays and feature flags in one tool, with a free allowance on eachPostHog
Your product data must stay on your own serversPostHog
You mostly need funnels, retention and cohort reports that the whole team can buildMixpanel Amplitude
You expect to send more than 1 million events a month and want to stay on a free planAmplitude

The contenders

Grouped by the side of the choice they answer, not ranked. Open a row for when to use it, the trade-off and what makers say.

ToolBest for

All-in-one platform1

PostHogplatform · open source · free tierOne tool for analytics, session replay, feature flags, experiments and surveys, each with its own free monthly allowance.

321+177 open source$0Pay-as-you-go11.5M/wk+108% vs npm

Use it whenYou want to see what users do in the app and test changes without adding more vendors.

Trade-offEvents tied to identified users cost several times more than anonymous ones past the free million; many products in one UI take time to learn.

llms.txt MCP CLIIn Lovable

Used by Acedit, Agenta, Agentcard, Alai, Almanac and 316 more · in 177 open-source projects.

PostHog in full →Plans and calculator →

We switched to PostHog from another analytics tool and we've been super impressed - we rely on Session Replay daily and love being able to quickly spin up new charts and dashboards

Arrange, the makerSep 2026 ↗

Dedicated analytics2

MixpanelserviceMature funnel, retention and cohort analysis on event data.

108+18 open source$0Free

Use it whenProduct decisions depend on funnel and retention reports that non-engineers build themselves.

Trade-offThe free plan stops at 1 million events and 5 saved reports per user; group analytics and data pipelines are paid add-ons.

Used by ADYOUNEED, Agiled, AI Test Engineer by BlinqIO, Articos, Assembly and 103 more · in 18 open-source projects.

Mixpanel in full →Plans and calculator →

Mixpanel helps us understand how users interact with Zown in real time. It reveals drop-off points, highlights high-intent users, and powers our product decisions with data we trust.

Zown, the makerSep 2026 ↗

AmplitudeplatformFunnel, retention and cohort analysis, with session replay and experiments on every plan, including free.87+17 open source$0Free

Use it whenYou want event-based product analytics with a large free event allowance and unlimited seats.

Trade-offBuilt for larger product teams; many advanced features sit on custom-priced Growth and Enterprise plans.

In Replit

Used by ADPList, ADYOUNEED, Agiled, Andsend, Attributer and 82 more · in 17 open-source projects.

Amplitude in full →Plans and calculator →

Notion - the place where all our documentation lives Bluedot - for meeting and screen recordings Amplitude - that’s where we spend lots of time looking at our customer’s activity.

Bluedot, the makerDec 2023 ↗

Cost: the cheapest plan that fits events 1M events, from list prices. Try your own numbers →

Cost as you grow

Each contender's cheapest usable plan as usage rises.

Amplitude Mixpanel PostHogx: events per month (event) · cheapest usable plan at each point, list prices · try your own numbers

Before you choose

How to approach it

Add product analytics once you have users to learn from. Track sign-up, activation and payment as events first, then add more only when a question needs it. Identify users when they sign in so events from one person link up across visits and devices, and send the same events from the server for anything that must be counted exactly, like payments.

Common mistakes

Tracking every click from day one, so the event list is noise and the free allowance runs out before you have a question to answer.
Naming events inconsistently ("Signed Up", "signup_completed"), which splits one step into several and breaks funnels.
Counting payments and sign-ups only in the browser, where ad blockers drop a share of events; send those from the server.
Sending emails, names or other personal data as event properties without listing the vendor in your privacy policy.