You’re growing thirty percent year-on-year. So is the company that will be worth three times as much as yours at exit.
The difference is NRR. You’re probably not measuring it the right way — and if you are, you’re probably treating the wrong variable as the problem.
MRR tells you what happened last month. NRR tells you whether your business is compounding or leaking. A company growing thirty percent with eighty-five percent NRR is filling a leaky bucket. Most of its growth is replacing customers it already lost. A company growing thirty percent with one hundred and ten percent NRR is compounding. Same headline number. Completely different businesses underneath.
The Leak Nobody Has a Word For
Most SaaS companies have a churn taskforce. Almost none have a contraction policy.
Contraction is when a customer downgrades — drops from your growth plan to starter, reduces seats, moves from annual to monthly. It doesn’t show up in your churn rate because the customer didn’t leave. It doesn’t show up in your NRR dashboard because nobody built an alert for it. It just quietly erodes your revenue base while every metric you watch looks approximately fine.
A five hundred dollar monthly downgrade has the same NRR impact as a five hundred dollar monthly cancellation. The cancellation triggers a save attempt, a post-mortem, a review of what went wrong. The downgrade gets logged as a field update in your CRM and forgotten.
Here’s how it plays out at scale. You add fifteen new customers this quarter. Three customers downgrade. Eight churn. The topline grows four percent. You run a churn retrospective on the eight who left. Nobody looks at the three who stayed but got smaller. Next quarter the same thing happens. The bucket leaks from two holes and you’re only watching one of them.
NRR makes both leaks visible. That’s why it’s the right metric. Not because it’s more sophisticated than churn rate — because it’s more honest.
The formula: (starting MRR + expansion MRR − contraction MRR − churned MRR) ÷ starting MRR × 100. Run it for the last twelve months. If you’ve never done this, the number will tell you something your other metrics have been hiding.
The benchmarks: median NRR for B2B SaaS at three to twenty million ARR is one hundred and four percent. Top decile is one hundred and eighteen. Companies above one hundred and twenty percent trade at a sixty-three percent valuation premium at exit. NRR isn’t just a growth metric. It’s a valuation multiplier.
Here’s where you should be at each stage. Under one million ARR: ninety percent or above is acceptable — you’re still learning what makes customers stay. One to five million ARR: one hundred percent should be the floor — existing customers should not be contracting your revenue base. Five to twenty million ARR: one hundred and ten percent is achievable and expected by investors who understand SaaS unit economics. Above twenty million ARR: one hundred and twenty percent is the top-quartile benchmark. The companies that reach it at this stage built the expansion and contraction systems years earlier. If your NRR is below the floor for your stage, that’s not a product problem — it’s a customer operations problem. And it’s fixable before the next funding conversation.
The Compounding Math Nobody Runs
One hundred and three percent monthly NRR compounds to one hundred and forty-three percent annually. Ninety-seven percent monthly NRR compounds to seventy percent annually.
Three percentage points either direction feels like noise month-to-month. Over three years it’s the difference between a business growing from its existing base and one that treads water regardless of how many new logos it adds. The gap between those two businesses isn’t sales team size or marketing budget. It’s whether the existing customer base is growing or shrinking.
This is the math that changes where you invest next quarter. Moving NRR from ninety-four to one hundred and eight percent over eighteen months is worth more than doubling new customer acquisition — without the proportional CAC. The compounding works permanently. Every month that passes at one hundred and eight percent, the distance between you and the company that chose acquisition instead grows wider.
One thing NRR can hide. A company with one hundred and twenty percent NRR and seventy-five percent GRR is losing a quarter of its revenue base every year and covering it with expansion from the customers who stayed. That’s not a healthy business. It’s a business with a serious churn problem that looks fine on the NRR dashboard because a small group of expanding customers is carrying everyone else. GRR and NRR together tell the real story. NRR alone doesn’t.
Where to Start
Three things. In order.
Fix your measurement first. Separate contraction MRR from churned MRR in your reporting. They’re not the same problem and they don’t get fixed the same way. If you can’t make that separation in your current metrics setup, that’s the first afternoon’s work. You can’t manage the leak you can’t see.
Build a contraction intervention. Find every customer who downgraded in the last six months. Find the date they downgraded. Go back thirty, sixty, ninety days and look at what their product usage looked like at each point. The signal was there. Build the same early warning system you’d build for churn — because contraction is churn in slow motion. A customer who downgrades today churns at a significantly higher rate within twelve months than a customer who doesn’t. The downgrade is the leading indicator you’ve been ignoring.
Make expansion systematic, not opportunistic. At fifty million ARR, expansion contributes fifty-eight percent of new ARR at the median. The companies that get there build the expansion system before they need it. Not a QBR slide — a triggered motion. Usage hitting eighty percent of tier limit. Three users added in a week. A milestone completed that opens an adjacent use case. Each signal routes to the right person before the customer thinks to ask. Expansion that connects to what the customer has already achieved converts. Expansion that arrives as a scheduled upsell call doesn’t.
We built this on MatrixFlows — structured customer records with contraction flags, expansion signals, and milestone tracking in one place. Contraction triggers an intervention. Expansion signals route to CS. The system watches the whole NRR picture, not just the half that shows up in your churn report.
What to Do This Week
Run the NRR formula for the last twelve months. Write the number down. Then pull every customer who downgraded in the last six months — not churned, downgraded. Calculate what that contraction cost you in MRR. Compare it to what your churn cost you.
For most companies that number is a surprise. Contraction is costing more than churn and getting a fraction of the attention.
That asymmetry is where most NRR problems hide. And it’s the one that’s easiest to fix — because the customers are still there.
If the expansion side of NRR is where you want to go next, the signal capture system is here. The health score system that catches at-risk accounts before they downgrade is here. MatrixFlows is free to start.
.svg)
.svg)
.svg)
