KPI Tree

A KPI tree (also called a metric tree or driver tree) is a hierarchical breakdown of a top-level metric into the sub-metrics and drivers that mathematically or causally determine it — turning one big number (revenue, the NORTH-STAR-METRIC) into the connected web of levers underneath it. Revenue breaks into traffic × conversion × average order value; conversion breaks into its own drivers; and so on down to the daily, actionable metrics a specific team or person actually controls. The tree makes visible how the work at the bottom rolls up to the goal at the top — which is its whole value.

How to build one and what makes it useful: start with the top metric (the one number the business or team is accountable for), then decompose it level by level into its drivers — ideally with real mathematical relationships where they exist (revenue = customers × average revenue per customer; new customers = traffic × conversion rate; the LTV and CAC and retention drivers underneath) and clear causal links where the relationship is directional rather than arithmetic — continuing down until you reach metrics that are specific, owned, and actionable (a team can directly influence them). The two kinds of relationships in a good tree: additive/multiplicative (the arithmetic decomposition — these branches let you actually model how a change in a driver moves the top, the basis of growth modeling) and causal/influencing (the softer relationships — content quality influences conversion, which isn't a clean formula but a real driver). What it does for an organization: alignment (everyone sees how their metric connects to the company's — the support team's resolution time, the content team's organic traffic, the paid team's CAC all visibly rolling up, ending the 'is my work mattering' ambiguity), prioritization (the tree shows which drivers have the most leverage on the top metric — improving a high-leverage mid-funnel driver beats polishing a low-leverage leaf, the same logic as the JOURNEY-ANALYTICS breakpoint hunt), diagnosis (when the top metric moves, the tree tells you WHICH driver caused it — revenue fell, walk the tree to find the branch that broke), and goal-setting (top-level targets cascade into driver-level targets that are concrete and ownable). The disciplines that separate a useful tree from a wall-decoration: it must reflect REAL relationships (a tree of metrics that don't actually drive each other is worse than none — it implies false leverage), it must reach actionable leaves (a tree that bottoms out in metrics nobody controls aligns no one), it must be maintained as the business model evolves (the drivers change), and it must avoid the vanity-metric trap (including metrics in the tree because they're measurable, not because they drive the top — the GOODHART risk of making a tree-metric a target that then gets gamed without moving the real goal). The relationship to its cousins: a KPI tree is the structural backbone that a north-star metric sits atop, that OKRs can cascade through, and that growth models quantify — it's the map of how the metrics connect, which the other frameworks then use.

A KPI tree matters most when an organization or team needs to connect daily work to a headline goal — at planning and goal-setting time (cascading top targets into ownable driver targets), in diagnosis (walking the tree to find which driver moved the top metric), and in prioritization (finding the high-leverage drivers worth investing in). It matters for cross-team alignment (making visible how disparate functions' metrics roll up to one number) and as the backbone beneath north-star metrics, OKRs, and growth models. It matters less as a static diagram (it must reflect real relationships and reach actionable leaves, or it's decoration). The discipline is building the tree on genuine mathematical and causal relationships, decomposing down to metrics teams actually control, using it to find leverage and diagnose movement rather than just to display metrics, maintaining it as the model evolves, and resisting the vanity-metric trap of including measurable-but-non-driving numbers.

The KPI tree formalizes a practice as old as financial analysis - the DuPont decomposition of return on equity into its drivers (1920s) is an early metric tree - and growth and analytics teams adopted it to connect daily, ownable metrics to headline goals; it became the structural backbone beneath north-star metrics, OKR cascades, and quantitative growth models.

Etymology: source.

Search interest for this term over the last five years:

View interest-over-time on Google Trends →

Areas of marketing where kpi tree is a core concern: