The decision
Its worst layer scored 38. It fixed a different one first.
The recording shows the instrument running. This is what came out of it, and
why the obvious move was the wrong one — the judgement is the part the
score exists to produce.
Composite52 / 100
TierAt Risk
Weakest layerExpansion · 38
Rule of 40
44.4Growth is not the problem.
Net revenue retention
108%Expansion is masking churn.
Logo churn
13%The actual constraint.
CAC payback
20 monthsA symptom, not a cause.
The pair worth sitting with is NRR at 108% beside logo churn at 13%.
Both are true at once: a few expanding accounts carry the revenue line while the
account count quietly falls. That is why growth felt more expensive every quarter
without any single number looking alarming.
Two moves were rejected
Fix Expansion first, because it scored lowest.
Expansion multiplies the base it is applied to. Applying it to a base
losing 13% of its logos a year builds a bigger funnel into a leaking tank.
The score was real; the sequencing was wrong.
Attack CAC payback, because 20 months is long.
Payback is arithmetic over a lifetime. Shorten the lifetime and payback
lengthens on its own. Here it was reporting the churn problem rather than a
separate acquisition problem, and cutting spend would have improved the ratio
while shrinking the company.
Chosen: Retention, starting week one. The C-06 churn intervention ranked
first not because retention scored worst — it did not — but because every
other repair compounds against it. Fixing retention makes expansion worth doing,
shortens payback without touching acquisition, and is the only one of the three
whose gain is not erased by the others being left alone.
Northwind Analytics is fictional and exists to show the workflow. Every figure is
calculated from numbers someone typed in; exposure is shown as a range carried at
an assumed multiple. Nothing here is a measured loss.
▶
