A KPI Tree Exercise That Exposed Team Misalignment

A KPI tree exercise is supposed to be a straightforward mapping session: pick a top-line metric, decompose it into the levers that drive it, assign ownership. We ran one for a client with growth, retention, and brand teams in the room, and it took about twenty minutes to stop being a mapping exercise and start being group therapy.

Where it broke down

We put 'net revenue retention' at the top of the tree, which everyone nodded along to. The moment we asked each team to name the single metric they personally optimized day to day, growth said new trial starts, retention said logo churn rate, and brand said share of voice. None of those are wrong metrics on their own. None of them, pursued independently without reference to the others, reliably moves net revenue retention — a growth team hitting its trial-start target by loosening qualification criteria can directly damage the retention team's churn number a quarter later.

The KPI tree didn't need to be built. The disagreement about what the tree should look like was the actual finding.

What we actually produced

The deliverable stopped being a clean decomposition diagram and became a documented set of tensions between team-level metrics, with explicit tradeoffs named out loud — for instance, that growth's trial-start target and retention's churn target needed a shared qualification-quality metric sitting between them, owned jointly, or they'd keep working against each other.

• A KPI tree exercise is often most valuable for exposing disagreement, not for producing the diagram everyone expects at the end.
• If every team names a different 'real' metric for the same top-line goal, that's the actual problem to solve before any dashboard gets built.
• We added a shared 'qualified trial' metric jointly owned by growth and retention, specifically to close the gap the exercise surfaced.
• Brand's share-of-voice metric stayed separate on purpose — not every team's work needs to collapse into the same tree, but it does need an honest conversation about where it connects.

If your KPI tree workshop produces a clean, uncontested diagram in under an hour, I'd be suspicious that it didn't ask hard enough questions. The useful ones are a little uncomfortable, because misalignment that's been quietly costing a company money rarely surfaces without someone naming it in a room.