Analytics for Indie Hackers: What Actually Matters at 0→1

When you’re building a product from zero, analytics feels like a problem for later. You’ve got one user — you. You know exactly what’s happening because you built the thing.

Then you launch.

Ten people sign up. You have no idea what they’re doing. You added Google Analytics in a panic at 2am and now you’re staring at a dashboard full of “bounce rate” and “sessions” that tells you nothing useful.

Here’s what actually matters.

The Only Metric That Matters at 0→1

Did they come back?

That’s it. Retention is your whole job at zero to one. Not pageviews. Not time on site. Not conversion rate. Those come later when you have enough signal to optimize anything.

A user who returns twice tells you more than a thousand users who bounced. Build toward that.

What to Actually Track

At launch:

Did they sign up?
Did they come back in 7 days?
What page did they land on?

After 10 customers:

Which signup source retains better? (LinkedIn referrals vs. HN vs. direct)
Where do people drop in your onboarding flow?
What does a “successful” user do in week one that a churned user didn’t?

After 100 customers:

Which plan tier retains best?
What’s your payback period by acquisition channel?
Where do refunds and cancellations cluster by cohort?

Most analytics tools front-load you with funnels, goals, and segments before you have the data to fill any of them. Start with three metrics. Add complexity when the data demands it.

The Tool Problem

Every analytics tool is designed for a different stage of company.

Google Analytics 4 is built for enterprise marketing teams with dedicated analysts. It’s free and comprehensive, but you’ll spend 80% of your time learning the tool and 20% learning about your users.

Plausible, Fathom, Umami are built for privacy-conscious developers who want a simple dashboard. They’re good at traffic. Weak on user behavior and anything agent-readable.

Mixpanel, Amplitude are built for growth teams with product analysts. Powerful, but overkill before you have product-market fit.

None of them were built for what’s actually happening in 2026: your users aren’t just humans anymore.

The New Problem: Agents Are Users Too

If you’re building something developer-facing — an API, a developer tool, an MCP server — your “users” increasingly include AI agents. These agents don’t have sessions. They don’t click around a UI. They make API calls, sometimes dozens per minute.

Traditional analytics tools don’t understand this. They’re built around pageviews and sessions. An agent making 50 API calls in 30 seconds looks like a bot attack, not a power user.

What you actually want to know:

Which MCP tools are being called most?
Which agent frameworks are integrating with my product? (Claude, Cursor, GPT)
Are agent users more or less retained than human users?
What does agent traffic tell me about product-market fit with AI-native workflows?

This isn’t a future problem. If you have an API or MCP server, you have agent traffic today. Most founders have no idea how much.

What We Built

Measure.events tracks both kinds of users — human and agent — in one place. One script tag for web traffic. API for agent-side events. A /summary endpoint that any AI tool can query in plain English.

Ask your own analytics: “What did my users do last week?” Get a real answer.

It’s $29/mo with a 14-day trial. If you’re at the 0→1 stage and don’t want to spend your first year learning dashboards, it’s worth trying.

The Real Point

Most founders over-invest in analytics tooling too early and under-invest in actually talking to users. The best analytics setup at 0→1 is the one you’ll actually use.

Pick something simple. Track the one thing you need to know. Don’t let the tool become the work.

Measure.events is privacy-first analytics built for developers. Start your free trial →