How to Get Started with Amplitude Analytics

What Amplitude Does

Amplitude tracks user behavior as a series of events. An event is any action a user takes in your product: signed up, clicked a button, completed a purchase, opened a feature. Every event carries properties — the plan the user is on, the country they are in, the page they were visiting. These events feed into charts that answer the questions most product and ops teams need: where users drop off, who retains, and what behaviors predict conversion.

The platform has four core chart types. Event Segmentation shows how often an event fires over time. Funnel Analysis measures the percentage of users who complete a series of steps in sequence. Retention Analysis tracks how many users return after performing an initial action. User Sessions shows visit patterns and session depth. That set of four covers the majority of product analytics work.

Creating Your Account and Choosing a Plan

Go to amplitude.com and create an account. The free Starter plan supports up to 50,000 monthly tracked users with unlimited events. It also includes session replay, feature flags, and web experiments at no cost. No credit card is required to start.

During signup you will be asked about your product type (web app, mobile app, or both) and your team size. The answers configure your initial project settings. You can change them later under Project Settings.

Connecting Your Product

Amplitude needs to receive events from your product. There are three main paths.

SDK integration. Amplitude ships client-side and server-side SDKs for JavaScript, React, iOS, Android, Python, Go, Node.js, and more. For a web app, install the Browser SDK with a single npm package and initialize it with your API key. You then call amplitude.track('Event Name', { property: value }) wherever a meaningful user action happens.

Autocapture. If you want a no-code starting point, Amplitude's Autocapture automatically tracks page views, clicks, form submissions, and URL changes without manual instrumentation. You install a JavaScript snippet and events start flowing immediately. Autocapture is limited compared to fully custom instrumentation, but it gets you to real data within minutes. It is the fastest path to a working funnel for teams that do not have an engineer available.

Third-party connectors. If your product already sends events to Segment, RudderStack, or another CDP, you can connect Amplitude as a destination. Events route through without any additional code changes.

Planning Your Events Before You Track

One common mistake is tracking everything and ending up with thousands of vague event names. Before you instrument anything, define the five to ten user actions that actually matter for your core use case. For a SaaS product that might be: Account Created, Onboarding Completed, First Report Generated, Integration Connected, Subscription Started.

Each event name should describe a completed action, not a page view. "Dashboard Viewed" is weaker than "Report Generated." Completed actions are what drive business outcomes.

Give each event a consistent naming format. Most teams use either snake_case or Title Case. Pick one and stick to it across the whole product.

Building Your First Funnel

Funnels are where Amplitude shows its value immediately. A funnel measures how many users progress through a defined sequence of steps.

In Amplitude, click New Chart and select Funnel Analysis. You will see an Events Module on the left side. Add your first event — for example, Account Created — then add the next step — Onboarding Completed — and continue adding steps. Click Update to run the chart.

Amplitude shows the percentage of users who completed each step within a conversion window you define (default is 30 days). If 1,000 users created an account and 310 completed onboarding, your activation rate is 31%.

You can break the funnel down by user properties — plan type, country, device — to see where different segments drop off. You can also set the conversion window to one day instead of 30 to see how urgency affects behavior.

The most useful follow-on action is to click into the "did not convert" group at any step. Amplitude shows individual user paths. Looking at five to ten of those paths often reveals a consistent pattern: a missing email confirmation, a required field users skip, or a page that fails to load on mobile.

Measuring Retention

Funnel analysis shows where users drop off. Retention analysis shows whether users come back.

In Amplitude, click New Chart and select Retention Analysis. You set a "birth event" (the action that starts the cohort, typically Account Created or Subscription Started) and a "return event" (what counts as coming back, typically any meaningful engagement action).

Amplitude groups users by the week or day they performed the birth event and tracks what percentage returned in each subsequent interval.

A healthy SaaS retention curve flattens above zero within the first four to six weeks. If the curve drops to near zero by week two, the product has a retention problem that better onboarding alone will not fix. If it flattens at week four around 20 to 30 percent, the core product is working but activation may need attention.

Amplitude's own data from 2025 shows the median day-7 retention rate across B2B SaaS customers is 22 percent. Breaking the retention chart by user segment — paid versus free, enterprise versus SMB — usually points to where the gap is concentrated.

Event Segmentation for Day-to-Day Monitoring

Funnel and retention charts are for deep dives. For day-to-day monitoring, Event Segmentation is what teams actually live in.

Event Segmentation shows a time series of event counts. Select any event — Feature Used, API Call Made, Export Completed — and the chart shows daily, weekly, or monthly volume. Add a group-by property and the chart breaks the line into segments.

A practical use: track your core activation event (the first action that strongly predicts paid conversion) as a daily metric. If that number drops two days in a row, something changed in the product or the top-of-funnel source. You can add multiple metrics to the same chart. Tracking signups and activations on the same axis makes it easy to spot whether a signup spike translated into an activation spike or not.

Setting Up a Dashboard

Amplitude Dashboards let you pin charts to a shared view your team can bookmark. Click Dashboards in the left nav and create a new one. Add your most important charts: your primary conversion funnel, your day-7 retention chart, and your core activation event segmentation.

Share the dashboard link with stakeholders. Amplitude dashboards update in real time, so there is no need to email screenshots.

Using Amplitude AI in 2026

Amplitude added AI Agents in 2026 that automatically surface trend changes and behavioral anomalies in your data. When a metric shifts unexpectedly, Amplitude flags it proactively and proposes a segmentation or path analysis to explain the change.

Amplitude also added an MCP integration in 2026, which means you can query your Amplitude data directly from Claude or Cursor using natural language. If you ask "what is my 30-day retention for users who completed onboarding this month?" Amplitude responds with the data without you opening the Amplitude UI.

If your product is not yet instrumented and you want to analyze existing user data now, VSLZ can take a CSV export of your event log and produce funnel and retention outputs directly — no SDK setup required — while you work on the full integration.

What to Track First

For a new Amplitude setup, start with five events:

1. Account Created
2. Your core activation event (the first action that predicts whether a user stays)
3. Your primary conversion event (first paid action, or first time the product delivers clear value)
4. A daily engagement event (whatever users do when they are getting ongoing value)
5. Churn or cancellation

Build one funnel from event 1 to event 3. Build a retention chart using event 1 as the birth event and event 4 as the return event. That is enough to tell you whether your product is working, and it gives you a baseline to measure any product change against.