What is a Metric Tree?
A metric tree (also called a driver tree or KPI tree) is a visual model that breaks a top-level metric into its component sub-metrics in a hierarchical structure. Each level shows how the metrics below it combine to produce the metric above.
For example, Monthly Recurring Revenue breaks into: Number of Customers x Average Revenue Per Customer. Number of Customers breaks into: Existing Customers + New Customers - Churned Customers. Each branch continues until you reach metrics that individual teams can directly influence.
Why Metric Trees Matter
Without a metric tree, teams optimize local metrics that may not move the needle on what matters. A growth team might celebrate a 20% increase in signups while the business metric (revenue) stays flat because activation and retention are broken.
Metric trees make the connections visible. When the team sees that signups flow through activation, retention, and monetization before becoming revenue, they can identify the bottleneck and focus there.
How to Build a Metric Tree
Start at the top with your company's north star metric or primary business outcome (revenue, DAU, etc.).
Decompose mathematically. Revenue = Customers x ARPU. Customers = New + Retained - Churned. ARPU = Price x Usage. Keep going until each leaf metric is actionable by a specific team.
Identify leading and lagging indicators at each level. Leading metrics predict future performance (trial starts, activation rate). Lagging metrics confirm past performance (revenue, churn rate).
Assign ownership. Each leaf metric should have a team that owns it. The onboarding team owns activation rate. The growth team owns new signups. The retention team owns churn. This prevents gaps and overlaps.
Metric Trees in Practice
Spotify uses a metric tree rooted in "time spent listening." This decomposes into sessions per user, tracks per session, listening duration per track, and content variety. Each squad owns a piece of the tree and optimizes their segment.
Amplitude (the analytics company) publicly shares their metric tree framework. They decompose customer value into: acquisition, activation, engagement, retention, and monetization. Each stage has specific metrics and specific teams responsible.
Common Pitfalls
- Too deep. A tree with 8 levels is hard to use. Aim for 3-4 levels. Deeper analysis can happen within teams.
- Ignoring interactions. Improving one metric can hurt another. Increasing signups by lowering quality hurts activation. Track the tree holistically.
- Static trees. Business models evolve. Review and update the tree quarterly.
- No ownership. A tree without assigned owners is just a diagram. Each leaf metric needs a team that watches it and acts.
Related Concepts
Metric trees connect to the north star framework for selecting the top-level metric. They are populated with leading and lagging metrics and inform OKR target-setting. Product analytics tools provide the data that fills the tree.