NRR is a judgement, not a fact
Why one of SaaS’s most quoted metrics depends on choices nobody writes down.
NRR is one of the most misunderstood metrics in SaaS.
I’ve built ARR models, cohort analysis and NRR reporting from scratch. When you’ve built the model yourself, you know where every judgement call sits because you’ve had to make each one.
That’s why I get nervous when I see NRR quoted as though it’s an objective fact.
It isn’t.
The moment you have to calculate it yourself, you realise how much depends on judgement.
Are you netting churn against expansion? When does a new customer enter the cohort? How are upgrades, contractions and renewals treated? Which cohort are you presenting? There are defensible answers to all of those questions, but they aren’t always the same answers.
The same problems come up again and again. NRR is calculated one way in the board pack and another way in the data room. Gross retention looks solid until someone starts pulling on the assumptions. CAC payback is impressive until you understand what’s been included, and what hasn’t.
Most of the time, nobody is trying to game the numbers.
Definitions drift. Different teams build their own versions. Reports get copied. Assumptions become accepted because they’ve always been there. Before long, the metric has taken on a life of its own.
That matters because NRR isn’t just a KPI. It feeds forecasts, valuations and investment decisions.
A small shortcut in a reporting model can end up affecting the value of a business.
I’ve always believed finance should own these metrics from end to end, including the definitions. Not because finance likes process, but because somebody has to make sure everyone is measuring the same thing.
Otherwise the business doesn’t really know its retention.
It just has a number everyone has agreed to believe.
I work as a fractional CFO for SaaS and technology businesses with a raise, a refinancing, a listing or an exit ahead. If that is where you are, I would be glad to talk.