Data Analyst & Analytics Interview · Product & Business Metrics · Lesson 1 of 2
Defining metrics, north-star, funnels & retention
Analysts are judged less on math than on choosing and defining the right metric. A good metric is specific, tied to value, and hard to game.
- North-star metric: the one measure that best captures delivered value (e.g. weekly active users, nights booked). Everything ladders up to it.
- Guardrail metrics: things you must not harm while moving the north star (latency, churn, refunds).
- Funnel: ordered steps toward conversion (visit → signup → activate → purchase). Analyze step-to-step conversion to find the leak.
- Retention / cohorts: group users by join period and track how many return over time. The retention curve that flattens (rather than hitting zero) signals product-market fit.
- Vanity vs actionable: total registered users is vanity; active users or activation rate drives decisions.
Define before you countHalf the battle is nailing the definition. 'Active user' — active how? Any event, or a core action? Per day or 28-day window? De-duplicated per user? Interviewers probe exactly here; state your definition explicitly before computing.
Cohorts are class yearsA cohort is like a graduating class: you track the January signups separately from the February signups. Mixing them hides that new users churn fast while old ones are loyal — the retention curve only makes sense per cohort.
◆ Lock it in
- Pick a north-star tied to value, protected by guardrail metrics.
- Funnels expose where users drop; cohort retention curves expose whether they stay.
- Always define a metric (window, dedup, which action) before you compute it.
Feynman drill — say it out loudExplain the difference between a north-star metric and a guardrail metric, and why you track retention by cohort.
Step 1 rate your confidence · Step 2 pick your answer
Which is the strongest north-star metric for a note-taking app?
Step 1 — how sure are you?