Are later cohorts retaining and expanding more than earlier ones?
№ 004 · NRR cohort heatmap
Definition
MetricNet revenue retention
Unit% of year-ago cohort revenue
Each row is an acquisition cohort, each column a month after acquisition. Green cells signal expansion above 100%; red signals contraction. Triangular because newer cohorts have less elapsed time. Build by dividing each cohort's MRR at month N by their month-zero MRR. Watch newer cohorts trending greener earlier than older ones.
- NRR
- Net revenue retention. Cohort revenue this period divided by the same cohort revenue a year ago, including expansion, contraction and churn. New logos excluded.
- MRR
- Monthly recurring revenue. Sum of every active subscription normalised to a monthly value, excluding one-off fees.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| 95% or below | 101-102% | 110-115%+ |
Private B2B SaaS, $1M+ ARR, blended across ACV bands, 2023-2025 survey data; NRR defined as MRR/ARR from year-ago customer base including upsell, cross-sell and price increases, measured over 12 months · Definitions matter: SaaS Capital measures Dec-over-Dec MRR from the year-ago base; some companies report trailing cohort averages. Median has compressed from ~105% (2021) to ~101% (2024) per Benchmarkit and Maxio data. The blended median hides a 21-point spread between enterprise and SMB, so benchmark within ACV band (see category split).
| Category | Bottom 30% | Median | Top 30% |
|---|---|---|---|
| SMB, ACV under $25k | 90% or below | 97% | 105%+ |
| Mid-market, ACV $25-100k | 97% or below | 102-108% | 111%+ |
| Enterprise, ACV above $100k | 103% or below | ~118% (top quartile of survey base) | 118-120%+ |
When it looks bad
Diagonal darkening: each newer cohort row turns below 100% earlier than the one above it, meaning recent vintages contract faster than old ones ever did.
2022 cohort held 104% at month 12; 2023 cohort 99%; 2024 cohort is at 96% by month 9. The product or ICP has quietly changed and every new cohort is worth less.
What to do about it
- Diagnose whether the heatmap darkens from churn or contraction by overlaying the GRR heatmap (card 005); if GRR is stable and NRR falls, the expansion motion broke, which is a pricing and packaging fix, not a CS fix.
- Add a usage or consumption pricing component for the expansion ceiling: usage-based companies reported 125% NRR vs 115% for subscription peers in OpenView's survey data; even a hybrid base-plus-usage model reopens expansion in flat-seat accounts.
- Target one NRR band of improvement, worth ~5 points of company growth (SaaS Capital); sequence: fix top 2 churn reasons in the newest cohorts first, since cohort-level fixes compound across all future vintages.
- If newer cohorts underperform older ones, audit ICP drift in the last 12 months of closed-won: win-at-all-costs deals during a slow quarter are the usual culprit and show up 9-12 months later on exactly this chart.
Sources
- SaaS Capital Median NRR 102% across survey; ACVs above $25k show median NRR of at least 103%; top quartile above $100k ACV at 118-120% saas-capital.com ↗
- SaaS Capital ACV $25-50k: median NRR 102%, top quartile 111%, bottom quartile 97%; NRR rises with ACV saas-capital.com ↗
- KeyBanc Capital Markets and Sapphire Ventures Net retention ~101%, stable sapphireventures.com ↗
- KeyBanc Capital Markets and Sapphire Ventures NRR remained above 100% through 2025 with modest improvement expected sapphireventures.com ↗
- Benchmarkit NRR benchmarks filterable by ACV; NRR most correlated with ACV among company attributes benchmarkit.ai ↗
- ChartMogul NRR is the key 2024 metric; retention and expansion now carry growth as new business slowed chartmogul.com ↗
- ICONIQ Growth Net dollar retention settling into a healthy ~110-120% range for their growth-stage universe iconiq.com ↗
- McKinsey Public companies with NRR 120%+ traded at median 21x EV/revenue vs 9x below that mark mckinsey.com ↗
- High Alpha and OpenView Public SaaS NDR steady at 110% for three quarters after declines from 2022 highs highalpha.com ↗