Net revenue retention (NRR) measures how much recurring revenue you keep and grow from existing customers over a period, usually 12 months. Take a group of customers, add their expansion, subtract their contraction and churn, then divide by what they paid at the start. Above 100% means the group is worth more than it was, with no new customers involved. CSMs grow it by protecting renewals first and then helping healthy accounts expand.
The net revenue retention formula
NRR = (starting recurring revenue + expansion − contraction − churn) ÷ starting recurring revenue × 100
Each part has a specific meaning:
- Starting recurring revenue: the annual or monthly recurring revenue of the customers you had on day one of the period.
- Expansion: added recurring revenue from those same customers, such as upsells, cross-sells and extra seats.
- Contraction: recurring revenue lost from customers who stayed but pay less.
- Churn: recurring revenue lost from customers who left.
A worked example
These numbers are made up to show the math. Say the customers you had a year ago were paying $1,000,000 in annual recurring revenue. Over the next 12 months, that same group changed like this:
| Item | Amount |
|---|---|
| Starting recurring revenue | $1,000,000 |
| Expansion | +$150,000 |
| Contraction | −$30,000 |
| Churn | −$70,000 |
| Ending recurring revenue from the group | $1,050,000 |
$1,050,000 ÷ $1,000,000 gives an NRR of 105%. Any customer who signed during those 12 months stays out of the calculation, however large the deal was.
Read NRR next to gross retention
NRR can hide a leak. Strong expansion from a few accounts can cover heavy losses everywhere else, which is why I'd never look at it alone. Gross revenue retention (GRR) uses the same starting number but ignores expansion, so it shows only what you kept. In the example above, GRR is ($1,000,000 − $30,000 − $70,000) ÷ $1,000,000, or 90%.
Here's how two companies with the same NRR can be in very different positions:
| Company A | Company B | |
|---|---|---|
| Gross retention | 96% | 80% |
| Expansion | 9% | 25% |
| Net retention | 105% | 105% |
| What it suggests | Stable base with room to sell more | Selling hard into a base that's leaving |
Company B's number looks fine until expansion slows, and then it falls quickly. It should fix retention before it pushes for more upsell. Company A can put more of its effort into growth.
Calculation mistakes that flatter the number
- Counting new customers. The group is fixed on day one. Revenue from anyone who joined later doesn't belong in it.
- Including one-time revenue. Implementation fees, services and one-off purchases aren't recurring. Usage-based revenue needs a clear, documented rule that you apply the same way every period.
- Counting booked deals as live revenue. Choose whether you measure on signing or on contract start, then stick with it.
- Hiding price increases inside expansion. They count, but report them on their own line so you can see what customers actually chose to buy.
- Treating a win-back as expansion. Agree on how to treat a returning customer, and write the rule down so it doesn't shift between quarters.
- Reporting one average. A company-wide figure can blend healthy enterprise accounts with a struggling small-business segment. Break it down by segment, customer age and CSM portfolio.
As for what counts as a good number, above 100% is the clear milestone, and past that the benchmarks vary a lot with customer size, pricing model and stage. Published surveys also use different definitions. Your own trend, split by segment, will tell you more than someone else's headline figure.
How CSMs grow NRR
A CSM affects all three moving parts of the formula. I work on them in this order, because expansion is hard to sell into an account that isn't getting value.
1. Protect what you have
Churn usually starts long before the renewal date: a slow launch, a lost sponsor, a team that never fully adopted the product. Get customers to a first real result quickly, and track the warning signs on every account. My guide on how to reduce churn covers this in detail, and a solid onboarding checklist prevents many of the problems before they begin.
2. Get ahead of contraction
Contraction is often easy to see coming: seats nobody uses, a department that never rolled out, a reorganization that shrinks a team. Look at your upcoming renewals early and ask which licenses or modules aren't earning their cost. Then there are two options. You can run an adoption push to get them used, or you can agree on a smaller footprint and a plan to grow back. Neither is comfortable, but both are better than a surprise cut at renewal from a customer who feels they've been overpaying.
3. Expand where the customer has already won
The best expansion follows a result. When one team hits its goal, ask who else has the same problem, and let your champion carry that story internally. Good signals include a new business unit, a new use case, teams asking for access and sponsors raising goals you can't yet support.
Expansion is also a team effort. I was part of the account team that closed a $968K net new ACV upsell with OpenAI, and deals of that kind need customer knowledge, commercial skill and technical support working together. Agree early on who owns which part with your account executive. For the step-by-step, read my guide to upselling enterprise customers.
4. Make your reviews about outcomes
Customers buy more and renew faster when they can see what they got for their money. Use a business review to show results against their goals, surface risks while you can still fix them and agree on what comes next.
5. Work from an account-level view
You can't manage a company-wide percentage from a CSM's seat, but you can manage a list of accounts. Keep a simple sheet like this and update it weekly:
For each account: current recurring revenue, renewal date, health and why, biggest contraction risk, next expansion opportunity, next action, owner, date.
Then ask: Which renewals in the next 90 days are at risk? Which healthy accounts have no expansion plan? Which accounts have no next action at all?
That last question often turns up the quiet accounts nobody is watching, which is where surprises tend to come from.
What a CSM can't fix alone
Be honest about what's outside your control. Selling to customers who were never a good fit, product gaps and pricing with no natural path to grow will all limit NRR, however good your account work is. When you see a pattern, bring evidence to the team that owns the problem: which accounts, how much revenue and what they said. That's more useful than a general complaint, and it's how customer success improves the number beyond its own book of accounts.
Want help moving your NRR?
If renewals are slipping or nobody owns expansion in your existing accounts, I can help you protect and grow that revenue.