The Three Questions That Matter

In the first 90 days, your analytics should help you answer three questions:

1. Are users activating? (Do new users experience the product's core value?)
2. Are users retaining? (Do activated users come back?)
3. Why are users leaving? (Where is the leakage and what's causing it?)

Everything else is secondary. Revenue metrics, virality, advanced funnel analysis — these matter, but they're not what you need to optimize for in the first 90 days. Get activation and retention right first.

Activation: The Most Actionable Early Metric

Activation is when a user first experiences the core value of your product. Your activation event should be specific and behavioral — not "completed onboarding" but "created their first [core object] and used it for [core action]."

For a project management tool: "Created a project and added at least one task that was marked complete." For an AI writing tool: "Generated at least 200 words and copied or exported the output." For a marketplace: "Completed a first transaction."

To measure activation: track the percentage of new users who reach this event within their first session, first day, and first week. Benchmark against these:

• First session activation: above 30% is strong
• 7-day activation: above 60% of your first-session activators is solid

Retention: The Signal That PMF Exists

Retention is the most honest signal of product-market fit. If users keep coming back, the product has value. If they don't, no amount of acquisition will build a sustainable business.

Measure retention in cohorts: of users who signed up in Week 1, what percentage came back in Week 2, Week 4, Week 8?

The shape of the retention curve matters more than the number. A curve that drops steeply and then flattens (even at 15–20%) suggests a retained core. A curve that continuously declines toward zero suggests no one finds enough ongoing value to stick.

In the first 90 days, your sample sizes will be small. Don't over-index on precise percentages. Look for direction and shape.

The Metrics That Can Mislead You

Daily Active Users (DAU): Meaningless in absolute terms with small user counts. A DAU of 50 might represent 50 very engaged users or 1,000 users at 5% daily engagement. Context matters.

Signups: Signups are a measure of acquisition, not product value. A viral landing page can get thousands of signups for a product nobody uses. Track signups relative to activation, not in isolation.

Time on site / session duration: High time on site can mean high engagement or high confusion. It's not a reliable signal without context from user interviews.

Bounce rate on content: Irrelevant for the product analytics question. That's a marketing metric.

Session Analysis: Qualitative Complement to Metrics

Numbers tell you what is happening. Sessions recordings (via LogRocket, FullStory, or PostHog) tell you why.

When you see a high drop-off at a specific step in your funnel, watch 10–15 session recordings of users who dropped off there. You'll often see patterns: users confused by a UI element, attempting the same action multiple times, or landing on a state they clearly didn't expect.

This qualitative complement to your quantitative data is one of the highest-leverage activities in your first 90 days. It takes 2–3 hours per week and generates specific, actionable product improvements.

The Weekly Analytics Ritual

Set up a weekly analytics review. For a small team, this takes 20–30 minutes:

1. Activation this week: What percentage of this week's new users activated?
2. Retention cohorts: How are the cohorts from weeks 2, 4, 8 ago holding?
3. Top drop-off point: Where are users leaving the funnel?
4. Any anomalies? Unusual spikes or drops in any key metric?
5. One hypothesis to test: Based on what you saw, what change will you make and measure next week?

This ritual turns analytics from passive reporting into active learning. The goal isn't to have perfect metrics — it's to make data-driven decisions faster than your competition.

Setting Your Baseline

In the first 30 days with real users, don't optimize — observe. Your metrics will be noisy because your sample size is small and your cohorts haven't matured. Use this period to understand patterns and set baselines.

At the 30-day mark, you should be able to answer: what is our current activation rate, what does our retention curve look like, and what is the single biggest obstacle between signup and activation? The answers to these questions should drive your product roadmap for days 31–90.