Analytics setup for early-stage startups and MVP validation

Executive overview

Most early-stage teams instrument analytics but never build the habit of acting on it. The result: metrics go unread and the business runs blind.

The solution is a simple three-metric stack — acquisition, retention, and revenue — built on top of a clearly defined funnel, displayed on a shared dashboard, and reviewed daily. This discipline separates data-driven teams from those that stall.

Retention cohorts are the clearest signal of product-market fit: a plateau above 20-30% means people genuinely want the product; a curve that trends to zero means they don't.

The core funnel

• Every product has three stages: acquire, engage (retain), monetise.
• Map your own custom steps within those stages first — tracking is useless without a defined funnel.
• Instrument each stage with named events and relevant properties (e.g. video_played with video_id, duration, progress).
• For revenue, send monthly_recurring_revenue for subscriptions or transaction value for e-commerce.

Acquisition metric: signups per week

• Track net new signups week over week, segmented by invite type (organic vs. referred).
• Organic users are the signal for genuine pull; invited users can inflate numbers artificially.
• Set a weekly numeric goal, display it on a team dashboard, and run experiments toward it.

Retention cohorts

• Group users by signup week and track what percentage returns on weeks 1, 2, 3, 4.
• Products without product-market fit trend toward zero; products with PMF plateau.
• Target plateau: at minimum 20-30% retention at week four.
• Retention period should match natural usage frequency — weekly for social apps, quarterly for travel products like Airbnb.
• Build the cohort report in Amplitude: entry event = signup, return event = core value action.

Revenue metric

• Use a property sum over monthly_recurring_revenue on the subscription upgrade event.
• Plot weekly net new revenue and set monthly growth targets against it.
• This is the primary north star metric once past early MVP stage.

Operating as a data-driven team

• Put all three metrics on a single TV dashboard visible to the whole team.
• Discuss daily: what projects are running, and are they moving the metrics?
• Send a weekly investor/advisor update synthesising what the data shows and what you plan to fix — this forces clarity and gets better advice back.

Recommended tool stack for MVP stage

• Google Analytics — traffic and acquisition source from the web
• Amplitude — feature-level analytics, retention cohorts, event segmentation
• Full Story — session recording to diagnose usability failures (catches issues metrics alone miss)
• Customer.io — behavioural email; send a personalised founder email ~43 minutes after signup
• Intercom — lightweight CRM for early customer list management
• Google BigQuery + Mode — data warehouse and SQL layer for non-technical co-founders (post-MVP)
• Live chat or direct Slack channels with early customers for qualitative signal

Avoiding common failure modes

• Don't agonise over tool choice — Amplitude and Mixpanel produce identical results at early stage.
• Best-in-class tools change roughly every two years; optimise for easy switching, not perfect selection now.
• Don't set up analytics and stop looking — the dashboard only works if the team looks at it every day.
• Don't skip Full Story when metrics look bad — the issue is often a single broken UI element, not a fundamental product problem.

The 43-minute founder email

• Send a short personal email from the founder ~43 minutes after a user signs up.
• Content: introduce yourself, name the next step, offer direct contact.
• Automate with Customer.io using the signup event as the trigger.
• This generates direct replies from confused users — the highest-quality feedback channel at MVP stage.