The Metrics Your Startup Should Actually Track

There are two types of metrics problems in startups: not tracking enough, and tracking too much of the wrong thing.

The first is common in very early-stage products. The second is common in products that have added analytics reactively — one event for each feature as it shipped, with no coherent framework.

Here's a framework that scales from MVP to Series A — and the order in which to build it.

The five categories that matter

Acquisition: How do people find you?

• Organic search, paid ads, direct, referral, social
• Source breakdown for sign-ups specifically — not just visits

Activation: Do new users reach a meaningful experience?

• % of sign-ups who complete your defined activation milestone
• Time from sign-up to activation

Retention: Do users come back?

• Day 1, Day 7, Day 30 retention (% of users who return after each period)
• Feature-level retention (users who use feature X in their first week are often 2–3x more likely to be retained at 30 days)

Revenue: Are people paying?

• Monthly Recurring Revenue (MRR)
• Conversion rate from free to paid
• Churn rate

Referral: Are users bringing others?

• Viral coefficient (sign-ups generated per existing user)
• NPS, measured in-product or via email

Defining your activation milestone

The activation milestone is the most important definition in your analytics setup. It's the moment after which users are significantly more likely to become retained customers.

To find it: look at your most retained users (still active at Day 30). What did they all do in their first session that users who churned didn't do? That's your activation milestone.

Common patterns:

• Collaboration tool: created a project AND invited a teammate
• Analytics tool: installed tracking code AND saw first data
• Finance tool: connected a bank account AND ran a report

The milestone usually involves completing the core workflow, not just signing up or browsing.

Before you have enough data to find this empirically, make a hypothesis and start tracking. You'll refine it as you learn.

Retention: the metric that predicts everything

Retention is the most predictive metric for long-term success. A product with good retention can be grown through acquisition. A product with poor retention can't be fixed by acquisition — you're filling a leaky bucket.

Measure cohort retention: of all users who signed up in week X, what % are still active in week X+4?

As a rough benchmark:

• Below 20% at 30 days: you have a product problem that more acquisition won't solve
• Above 40% at 30 days: you have something worth growing

If retention is low, no amount of marketing will fix it. The work is in the product.

The dashboard to review weekly

The metrics dashboard should have:

• One primary metric for each of: acquisition, activation, retention, revenue
• Trends over time (not just point-in-time numbers)
• Cohort-based retention chart
• Funnel visualisation for your core conversion flow

What it shouldn't have:

• Vanity metrics (total sign-ups without activation context, pageviews in isolation)
• Metrics with no clear owner or action trigger
• Raw event counts without normalisation

Review this weekly with the founding team. Align on which metric matters most for your current stage and orient product decisions around moving it.

The order of instrumentation

Tell your developer to build this in this sequence:

1. First: user lifecycle events (signed up, activated, upgraded, churned)
2. Second: core feature events for your 3 most important features
3. Third: funnel events for your primary conversion flows
4. Later: engagement signals, feature discovery, NPS surveys

Don't try to instrument everything at once. Get lifecycle events right, then add as you need to answer specific questions. A metrics system that's 70% of the way there but maintained and used is more valuable than a comprehensive system nobody checks.

Building a product with proper instrumentation from day one? Let's talk →