The metrics that matter for product success are the ones that predict future business outcomes, not the ones that describe past activity — and the most important discipline in product metrics is distinguishing between input metrics (things you control) and output metrics (things that result from your inputs), because optimizing for output metrics directly is the path to Goodhart's Law disasters.
Every product team has more data than they know what to do with. The challenge is not measurement — it is selection. Teams that measure everything effectively measure nothing, because when every metric matters, none of them guide decisions. This framework shows how to identify the metrics that actually predict product success.
The Product Metrics Hierarchy
Level 1: North Star Metric (one metric that best captures delivered value)
↓
Level 2: Input Metrics (things the team directly controls)
↓
Level 3: Health Metrics (guardrails — these should not decline)
Level 1: The North Star Metric
The north star metric is the single metric that best captures the value your product delivers to users over time. It should:
- Be correlated with long-term revenue retention
- Be measurable at the user or account level
- Be something the product team can directly influence
- Not be a vanity metric (total signups, app installs, page views)
North star examples by product type:
- Consumer social: Daily active users (or daily active users who create content)
- E-commerce: Orders per buyer per quarter
- B2B SaaS: Weekly active accounts
- Marketplace: Total transaction value (or transactions per active seller)
- Productivity tool: Documents created or tasks completed per active user
Level 2: Input Metrics
Input metrics are the things your product team directly controls that are expected to move the north star. They are leading indicators.
How to identify input metrics: Ask "If we improve X, we expect the north star to move because...". If you can't complete that sentence with a plausible causal story, X is not an input metric — it's a proxy with uncertain connection.
Example: For a B2B SaaS north star of weekly active accounts:
- Input: Onboarding completion rate → more accounts reach first value → more weekly active accounts
- Input: Feature adoption breadth → more value per account → more habituality → more weekly active accounts
- Input: Integration count → more workflow embedding → lower churn → more weekly active accounts
Level 3: Health Metrics (Guardrails)
Health metrics don't predict growth — they detect when something is going wrong. They should be monitored, not optimized.
Common guardrail metrics:
- P95 API response time (technical health)
- Support ticket volume per active user (friction health)
- Error rate (reliability health)
- Feature churn rate (retention health — users who try a feature and stop)
The Vanity Metric Audit
For each metric your team currently tracks, ask:
- Does this metric change if the product delivers more value to users? → If no, it may be a vanity metric
- Would a bad product decision be able to improve this metric? → If yes, it's gameable and dangerous to optimize
- What business decision would change based on this metric? → If you can't name one, stop tracking it
FAQ
Q: What are the metrics that matter for product success? A: The north star metric that captures delivered user value, input metrics that predict north star movement, and health guardrail metrics — organized in a hierarchy where each level serves a different decision type.
Q: What is the difference between input and output metrics in product management? A: Output metrics (revenue, retention, north star) measure what results from the product. Input metrics (onboarding completion, feature adoption, integration count) measure what the team directly controls. Improving input metrics is how you move output metrics.
Q: What is Goodhart's Law and why does it matter for product metrics? A: Goodhart's Law states that when a measure becomes a target, it ceases to be a good measure — people optimize the metric in ways that don't produce the underlying outcome. This is why directly optimizing output metrics like NPS or retention leads to gaming behavior.
Q: What makes a good north star metric for a product? A: It should be correlated with long-term revenue retention, measurable at the user or account level, directly influenceable by the product team, and not a vanity metric like total signups or page views.
Q: How many metrics should a product team actively track? A: One north star metric, 3 to 5 input metrics, and 3 to 5 health guardrail metrics — any more than this and the team cannot hold the full picture in working memory, which defeats the purpose of having metrics.
HowTo: Build a Product Metrics Framework
- Define your north star metric — the single metric that best captures the value your product delivers to users over time, correlated with long-term retention
- Identify 3 to 5 input metrics that are leading indicators of north star movement with plausible causal stories connecting each input to the north star
- Define 3 to 5 health guardrail metrics that should be monitored but not optimized — these detect problems, they don't predict growth
- Run a vanity metric audit on all currently tracked metrics and remove any that do not influence decisions or cannot be linked to business outcomes
- Review the metrics hierarchy quarterly to ensure input metrics still have a causal connection to the north star as the product evolves
- Document the metrics framework in a shared location so all stakeholders use the same definitions — metric definition drift is one of the most common causes of unproductive product strategy discussions