Net Revenue Retention (NRR): How to Calculate and Improve It
By Newtown Analytics · · 6 min read
For startup founders
Net revenue retention tells you how much of last year's recurring revenue you still have this year from the same customers, after counting upgrades, downgrades and cancellations, and it is one of the clearest signs of whether a subscription business can grow from its existing base. A figure above 100 percent means customers are spending more in total than they did a year ago, even if some of them left. A figure below 100 percent means the base is shrinking and new sales must replace it. This guide gives the formula, works through a cohort example with checked arithmetic, separates net from gross retention, shows how to segment the metric, covers what drives it and how to improve it, and lists common calculation errors. It belongs to unit economics and budgeting and forecasting, builds on annual recurring revenue and MRR vs ARR, and leads to gross revenue retention and the Rule of 40.
The formula
NRR = (Starting recurring revenue + Expansion - Contraction - Churn) / Starting recurring revenue
Measure it over a fixed period, usually twelve months, for the group of customers that existed at the start. New customers acquired during the period are excluded, even if they expand later. This is what makes the metric a measure of the existing base.
A worked cohort
Take the 100 customers who were active on January 1, with recurring revenue of $1,000,000. During the next twelve months (hypothetical):
| Movement | Amount |
|---|---|
| Expansion (upgrades, added seats, price increases) | +$150,000 |
| Contraction (downgrades) | -$30,000 |
| Churn (cancelled customers) | -$70,000 |
Revenue from the same customers at the end of the year is 1,000,000 + 150,000 - 30,000 - 70,000 = $1,050,000.
NRR = 1,050,000 / 1,000,000 = 105.0 percent.
Gross revenue retention (GRR) counts only the losses, not expansion: (1,000,000 - 30,000 - 70,000) / 1,000,000 = 900,000 / 1,000,000 = 90.0 percent. GRR cannot exceed 100 percent, and it shows how well you keep what you have. The difference between the two, 15 points here, is the contribution of expansion. See gross revenue retention.
The cohort grew by $50,000 without any new customer. If the company had also added $300,000 of new-customer ARR during the year, total ARR would have risen from $1,000,000 to $1,350,000, and only a part of the growth would come from outside the base. Compare this with a business at 85 percent NRR: with the same $300,000 of new ARR, it would end at $850,000 + $300,000 = $1,150,000, and it would need to win $150,000 of new business just to stay flat.
Why a company with high NRR is in a different position
NRR compounds. Starting from $1,000,000, a cohort with 105 percent NRR is worth $1,102,500 after two years (1,000,000 × 1.05 × 1.05), and one at 90 percent is worth $810,000 (1,000,000 × 0.9 × 0.9). Over five years, 105 percent gives about $1,276,000 and 90 percent gives about $590,000. The numbers are straightforward powers: 1.05 to the fifth power is 1.2763, and 0.90 to the fifth power is 0.5905. A business with high retention needs less new business to grow, so its sales and marketing spend is more efficient and its forecasts are more predictable. This is why investors ask for NRR early.
What drives net revenue retention
| Driver | Effect | Typical levers |
|---|---|---|
| Churn of whole customers | Lowers NRR | Onboarding, support, product fit, contract length |
| Downgrades | Lowers NRR | Value realization, usage monitoring, plan design |
| Seat or usage growth | Raises NRR | Land-and-expand motion, usage-based pricing |
| Upsell to higher tiers or add-ons | Raises NRR | Packaging, product roadmap, account management |
| Price increases | Raises NRR | Value-based pricing, communication. See how to raise prices |
| Cross-sell | Raises NRR | Product breadth, integration |
A price increase counts as expansion in NRR, so a rise can lift the figure without any improvement in customer success. Look at NRR with and without price effects to understand what is happening.
Segment before you conclude
A single company-wide NRR can hide important differences. Calculate it by customer size, plan, industry, acquisition channel, region and customer age. A typical finding is that large customers expand and small customers churn, so the total hides two different businesses. Another is that customers who were onboarded with a particular process retain much better than others. The segments that show the weakest results point to where to invest.
| Segment | Starting ARR | Ending ARR | NRR |
|---|---|---|---|
| Large customers | $600,000 | $690,000 | 115.0% |
| Mid-size customers | $300,000 | $294,000 | 98.0% |
| Small customers | $100,000 | $66,000 | 66.0% |
| Total | $1,000,000 | $1,050,000 | 105.0% |
Check: 690 + 294 + 66 = 1,050 (in thousands); 690 / 600 = 115.0 percent, 294 / 300 = 98.0 percent, 66 / 100 = 66.0 percent. The total of 105 percent masks a segment losing a third of its revenue. A firm might decide to change how it serves small customers (a self-serve or lower-touch model) or to stop acquiring them.
Monthly versus annual measurement
NRR is most often reported on a trailing twelve-month basis for the cohort that existed twelve months earlier. For businesses with monthly contracts, a monthly NRR can also be computed, but it is hard to compare with annual figures: monthly NRR of 99.0 percent compounds to 0.99 to the twelfth power, which is about 88.6 percent a year. Always state the period. Do not annualize a monthly figure by multiplying by 12.
Improving NRR
- Reduce early churn. Focus on the first 90 days: onboarding, first value, training. Early churn is often the most expensive.
- Monitor health. Track usage, support contacts, engagement and payment behavior, and intervene early with at-risk accounts.
- Design plans for expansion. Price on a metric that grows with the customer's success.
- Build an expansion motion. Assign ownership for growth in existing accounts and give them tools and incentives.
- Improve renewals. Start renewal discussions early, use multi-year terms where they suit, and learn from every cancellation.
- Listen to the reasons. Record a reason for each downgrade and cancellation, and review them as a group.
- Do the unglamorous product work. Reliability and fit usually matter more than clever retention programs.
Common calculation errors
- Including new customers in the numerator.
- Using the wrong starting base, such as customers at the end of the period.
- Mixing revenue and customer counts. NRR is a revenue measure. See revenue churn versus logo churn.
- Counting pilots and one-time fees.
- Treating reactivated customers inconsistently. Decide whether a returning customer is new or part of the cohort.
- Hiding price increases inside expansion.
- Excluding churned customers that you consider "bad fits".
- Choosing the period after the fact to show a better number.
Using NRR in the model
For forecasting, apply the NRR assumption to the opening recurring revenue to get the revenue expected from the existing base, then add new customers. A forecast that assumes 110 percent NRR on a base that has historically retained 95 percent is optimistic by 15 points on the largest part of revenue. Test the plan against a range, such as 90, 100 and 110 percent. See budgeting and forecasting.
FAQs
What is net revenue retention?
The percentage of recurring revenue from customers who existed at the start of a period that remains at the end, after expansion, contraction and churn.
How do you calculate NRR?
Take starting recurring revenue, add expansion, subtract contraction and churn, and divide by starting recurring revenue. Exclude new customers.
What is a good NRR?
It depends on your market, customer size and pricing model. Above 100 percent means the existing base is growing on its own. Compare against your own history and similar businesses, and look at the segments.
What is the difference between NRR and GRR?
NRR includes expansion, so it can exceed 100 percent. GRR counts only downgrades and churn and cannot exceed 100 percent.
Can price increases improve NRR?
Yes, because they count as expansion. Separate price effects from usage and upsell effects to see what is driving the number.
Keep reading
This article is for general information. It is not accounting, tax, legal or investment advice, and examples use illustrative numbers rather than client data. Confirm decisions that carry financial or legal consequences with a qualified professional.