Startup growth diagnostic: the method
A growth diagnostic is not another dashboard. A dashboard answers "where are we?"; a diagnostic answers "what is stopping us going faster, and how do we know?". A diagnostic, unlike a dashboard, has to be able to be wrong.
Start with what nobody can show you
Day one of a diagnostic isn't spent in the data: it's spent asking for five numbers and writing down which ones arrive. A number you get in two minutes and a number that takes a week say different things about the company, even when they're identical.
Three possible reasons for an absence, and each calls for a different plan: the event isn't instrumented, it is but nobody computes it, or it's computed three different ways that don't agree. The third is the most expensive and the most common — and it's the only one that looks like success from a distance.
The order to look in
The five stages multiply: ten percent better at four of them is worth roughly forty-six percent more at the end, not forty. That's also what makes the order decisive — fixing acquisition on a product nobody keeps is filling a leaky bucket faster.
The practical rule: work backwards. Look first at whether the people who pay stay, then at whether the people who sign up reach value, and only then at where the signups come from. That's the reverse of the acronym's order, and it's the order in which fixes pay off.
Measuring practices, not results
Comparing your churn to the median in a public report is nearly always false precision: the median covers companies whose segment, definition and window you don't know. Two companies can publish the same rate while counting two different things.
Practices, on the other hand, do compare: "do you track your activation rate?" has the same answer everywhere, whatever definition is used. That's why a diagnostic of practices holds up from one company to the next where a diagnostic of results does not — and it's what the fifteen questions measure.
End on an action, not a finding
A diagnostic that stops at the finding changes nothing. The last page has to carry ONE action, the one that unlocks the weakest stage — not a list of twelve recommendations nobody will know where to start with.
And that action has to be checkable within a short window. "Improve retention" isn't one; "instrument an activation event and read next week's cohort" is, because in eight days you'll know whether it happened.
The five stages, one page each
Want the list itself, ready to work through? The 15-point growth audit checklist
The method's fifteen questions, in three minutes, with the stage that's holding you back named and one action to take: