How to build a product metrics framework for an early-stage startup requires choosing one north star metric that captures core value delivery, three leading indicators that move before the north star, and two counter-metrics that prevent local optimization — and resisting the pressure to track more until you have answered why your current metrics aren't where they should be.
Early-stage startups face two opposite mistakes in metrics: measuring nothing (running on intuition) or measuring everything (drowning in dashboards). The second is more common and more damaging, because it creates the illusion of data-driven decision-making while making every decision harder.
A metrics framework is not a list of metrics. It is a hierarchy — a small number of metrics organized so their relationships are clear and every team member knows which number to improve this sprint.
The Metrics Hierarchy
Level 1: The North Star Metric
The north star metric captures the core value your product delivers to customers. It is the single number that, if it increases consistently, means your product is working.
Criteria for a north star metric:
- Measures value delivered, not output produced (weekly active collaborators, not logins)
- Leads revenue rather than lagging it
- Is meaningful to both the product team and leadership
- Can be improved by the product team's work
North star examples by business model:
| Business model | North star example | |---------------|-------------------| | Consumer subscription | Monthly active subscribers who watch >3 sessions | | B2B SaaS | Accounts with weekly active users in core workflow | | Marketplace | Successful transactions per week | | Developer tools | Weekly active developers using the API | | Freemium | Activated free users (reached the aha moment) per month |
H3: The North Star vs. Revenue
Revenue is not a north star metric for early-stage startups, even though it feels like the most important number. Revenue lags product usage by 30–90 days (subscription delay, contract cycles). By the time revenue drops, the product problem that caused it happened months ago.
A north star that leads revenue gives you the early warning system that revenue alone cannot.
According to Lenny Rachitsky's writing on north star metrics, the most common mistake early-stage startups make is treating monthly revenue or ARR as the north star. "Revenue is a trailing indicator. If your north star drops in February, your revenue will drop in April. If you use revenue as the north star, you've lost 60 days of reaction time."
Level 2: Leading Indicators (3 metrics)
Leading indicators are the upstream behaviors that predict the north star. They move earlier and give the team actionable signals.
How to find leading indicators:
- Map the user journey from acquisition to north star moment
- Identify the 3–5 steps that most correlate with north star completion
- These become your leading indicators
Example (B2B SaaS, north star = accounts with weekly active users in core workflow):
| Leading indicator | Why it predicts the north star | |-----------------|-------------------------------| | Activation rate (users completing setup in first session) | Activated users are 5x more likely to reach weekly active state | | Team invitation rate (users who invite a teammate in first week) | Team accounts have 3x higher 90-day retention | | Core workflow completion rate | Users who complete the core workflow in week one retain at 2x the rate |
Level 3: Counter-Metrics (2 metrics)
Counter-metrics prevent the team from improving the north star through local optimization that damages the product long-term.
Example counter-metrics:
| North star | Counter-metric | Why | |-----------|---------------|-----| | Activations per month | 30-day retention of activated users | Prevents cheap activation tactics that don't retain | | Transactions per week | Transaction quality score | Prevents fraud or forced transactions | | Free-to-paid conversions | Paid churn rate | Prevents overpromising in trial to convert users who won't stick |
Common Early-Stage Metrics Mistakes
Mistake 1: Tracking Vanity Metrics
Vanity metrics look impressive but don't predict outcomes:
- Total signups (vs. activated users)
- Page views (vs. task completion rate)
- Social media followers (vs. traffic from social)
- Press mentions (vs. referral from press)
The vanity metric test: If the metric can improve while the business is failing, it's a vanity metric. Signups can grow while activated users stagnate. Track the activated metric.
Mistake 2: Metric Proliferation
Every new stakeholder wants a new metric. Over time, the dashboard grows to 40 metrics and no one knows which 5 matter.
The counter-protocol: For every new metric proposed, ask "what decision will this metric inform that our current metrics don't?" If there's no answer, don't add it.
H3: The "So What?" Test
According to Shreyas Doshi on Lenny's Podcast, every metric in the dashboard should pass the "so what?" test — what action would you take if this metric changed? If the answer is "I don't know" or "nothing immediately," the metric doesn't belong in the active dashboard. "A dashboard that can't drive a decision is a reporting artifact, not a decision tool."
Building the Framework
Step 1: Define the north star
Get team agreement on the single metric that captures core value delivery. This requires a conversation, not a spreadsheet.
Step 2: Map the leading indicators
Run a retention cohort analysis to find the early-stage behaviors that predict north star completion. Document the relationship: "users who do X in their first week reach the north star at 3x the rate."
Step 3: Add counter-metrics
For the north star and each leading indicator, ask: what metric should not go down as a result of trying to improve this metric? Add those as counter-metrics.
Step 4: Build the minimum viable dashboard
Three levels, 6–8 total metrics. One screen. Updated weekly.
According to Gibson Biddle on Lenny's Podcast, the most valuable metrics frameworks he observed at Netflix were the simplest ones — one north star, three leading indicators, two counter-metrics, reviewed weekly by the whole team. "The teams with 40-metric dashboards were always the slowest to react to problems. The teams with 6 metrics were always the fastest. Simplicity is not a limitation — it's a forcing function for clarity."
FAQ
Q: How do you build a product metrics framework for an early-stage startup? A: Choose one north star metric that captures core value delivery, identify three leading indicators from retention cohort analysis, add two counter-metrics to prevent local optimization, and limit the active dashboard to 6-8 metrics reviewed weekly.
Q: What is a north star metric and how do you choose one? A: A single metric that captures the core value your product delivers to customers, leads revenue rather than lagging it, and can be moved by the product team's work. Examples include weekly active users in the core workflow, successful transactions per week, or activated free users per month.
Q: What are vanity metrics and how do you avoid them? A: Metrics that can improve while the business is failing — total signups, page views, social followers. Avoid them by using the vanity metric test: if the metric can grow while the business fails, track the activated version instead.
Q: How many metrics should an early-stage startup track? A: 6 to 8 in the active dashboard: one north star, three leading indicators, two counter-metrics, and one business metric like MRR. Tracking more creates confusion; tracking less misses leading signals.
Q: What is a leading indicator in a product metrics framework? A: An upstream behavior that predicts north star completion, moves earlier than the north star, and gives the team time to intervene. Found through retention cohort analysis identifying which first-week behaviors predict long-term retention.
HowTo: Build a Product Metrics Framework for an Early-Stage Startup
- Define the north star metric as the single number capturing core value delivery that leads revenue and can be improved by product team work
- Run a retention cohort analysis to identify which first-week user behaviors most correlate with north star completion — these become your leading indicators
- Add two counter-metrics that should not decrease as a result of optimizing the north star — to prevent local optimization that damages long-term product health
- Apply the vanity metric test to every metric: if it can improve while the business is failing it is a vanity metric; track the activated version instead
- Build a minimum viable dashboard with 6 to 8 total metrics reviewed weekly by the whole team on one screen
- Apply the so-what test to every proposed new metric — if no one can describe the decision it would inform, do not add it to the active dashboard