Net revenue retention: the number enterprise buyers ask us about
What a net revenue retention benchmark really measures, and the cuts a buyer checks before trusting yours.
A net revenue retention benchmark only helps when you know the cohort behind it. In the latest SaaS Capital survey of more than 1,000 private B2B SaaS companies, median net revenue retention is 101% and median gross retention is 91%. Enterprise buyers ask for both numbers. Then they ask how you built them.
One net revenue retention benchmark is really three numbers
When a founder writes "our retention is 110%" in an entry, we read three possible numbers. Logo retention counts customers kept. Gross revenue retention counts revenue kept from the starting cohort, with no credit for upsell. Net revenue retention adds expansion back in.
Each one answers a different question. Logo retention tells a buyer whether peers stay. Gross retention tells them whether the product holds its price. Net retention tells them whether customers buy more once the product is installed.
Most retention figures we see in technology entries arrive as one percentage with no label. The jury rarely argues about whether the figure is good. It argues about which of the three it is.
Gross retention, net retention and the expansion that flatters both
Take an illustrative cohort of 100 customers, each paying 10,000 a year. Starting revenue is 1,000,000. Over the next twelve months, eight customers leave and five downgrade by 2,000 each.
That is 90,000 lost. Gross retention is 91%. Ten customers then upsell by 3,000 each, adding 30,000. Net retention is 94%.
Now add one enterprise customer who expands by 100,000. Net retention jumps to 104%. Gross retention stays at 91%. One deal moved the headline by ten points and hid the same churn underneath.
This is why buyers read the two together. Expansion should never touch gross retention. SaaS Capital's formula caps each customer at their starting revenue, so gross retention cannot exceed 100%. If a deck shows gross retention above 100%, expansion has leaked in. Price rises need the same care, because they lift net retention without any new value delivered.
What an enterprise buyer infers from a retention curve
A careful buyer compares you with companies that sell at a similar price. SaaS Capital makes the same point: contract value is the best starting cut for retention benchmarks. The table below summarises its latest edition by annual contract value (ACV).
| ACV band | Median NRR | NRR 25th percentile | NRR 75th percentile | Median GRR |
|---|---|---|---|---|
| Less than $12k | 98% | 90% | 106% | 90% |
| $12k to $25k | 103% | 98% | 115% | 91% |
| $25k to $50k | 102% | 97% | 111% | 91% |
| $50k to $100k | 104% | 96% | 110% | 90% |
| $100k to $250k | 102% | 94% | 109% | 91% |
| More than $250k | 106% | 102% | 110% | 95% |
Source: SaaS Capital B2B SaaS retention benchmarks, latest edition (14th annual survey, more than 1,000 private B2B SaaS companies). ACV bands exclude companies under $1M ARR.
Three readings follow. First, gross retention near 90% is the floor. SaaS Capital calls a gross figure of at least 90% "table stakes" for parity with peers. Second, net retention of 110% or more sits with faster growth, while companies below 100% grow slower than the survey median. Third, larger contracts show higher medians and tighter ranges, so a high-ACV seller at 100% looks weaker than a small-ticket seller at the same figure.
The curve matters as much as the point. A buyer wants to see each cohort fall early and then flatten. A curve that keeps sliding after the second year suggests the product never became part of daily work.
The cohort maths most decks skip
The headline number depends on choices nobody prints. These are the cuts we check first.
Who sits in the starting cohort
Net retention follows only customers who were paying at the start of the window. New logos signed during the window belong in bookings. When they creep in, the figure measures growth.
The window and the revenue basis
A trailing twelve-month figure on recurring revenue is the common basis. One strong month annualised, or bookings instead of billed revenue, will flatter the result. Currency swings can do the same for companies billing in several markets.
Blended or cohort-weighted
A blended figure lets one large, happy account carry many small leavers. Ask for each annual cohort separately. If customers on a discontinued plan were removed from the base, the remaining figure suffers from survivorship.
The three retention artefacts we ask an entry to attach
Our judging process scores customer value on evidence, and retention data carries weight there. For entries in the SaaS awards category, three attachments settle most questions.
- A cohort table. Starting revenue per annual cohort, then churn, contraction and expansion in separate columns.
- A churn reasons log. Every lost or shrinking account with a one-line reason. Patterns here tell the jury more than the percentage.
- A contract list. Customer, start, term and value, with names masked if needed. It lets an assessor sample the cohort.
Our winners directory lists 161 technology winners across 46 countries. Read their descriptions and you will see few mention retention at all. Most cite users, clients or deployments instead. Entries that show the cohort table stand out for that reason alone.
Every nominee can enter one programme in one category on the Free plan. Paid plans add categories and promotional benefits; see the fee page. No plan, package or payment influences the jury or the result. You can browse every award category before choosing, and our awards FAQ covers eligibility. For how the wider Golden Tree network judges entries, see Golden Tree Awards.