Finance, metrics & runway · 2026-09-14

NRR vs GRR for the board pack

NRR asks whether existing customers grow revenue. GRR asks how much you keep with zero expansion. Put both on the board slide with cited bands. Do not mask a leaky base.

GRR answers how much recurring revenue you keep from a starting cohort with zero credit for expansion. NRR answers whether that same cohort’s revenue grows after expansion, contraction, and churn. Boards need both. NRR alone can hide a leaky base. Not legal, tax, or investment advice.

Two different board questions

MetricBoard questionExpansion?Ceiling
GRR (gross revenue / dollar retention)How much revenue would we keep with zero expansion?No100%
NRR (net revenue / dollar retention)Is revenue from existing customers growing overall?YesNone (can exceed 100%)

Do not compare your GRR to someone else’s NRR. Same cohort, same period, same publisher bands.

Formulas and a worked example

NRRGRR
Formula(Starting MRR + Expansion + Reactivation − Contraction − Churn) / Starting MRR(Starting MRR − Contraction − Churn) / Starting MRR
Counts expansion / reactivationYesNo
Can exceed 100%YesNo
SourceChartMogul NRR · updated 8 Sep 2026ChartMogul GRR · updated 8 Sep 2026

Labeled ChartMogul-style example (not your company). Start with $770 MRR from four paying customers. Later, the same cohort is at $800 MRR after expansion, churn, and contraction → NRR = 800 / 770 = 103.9%. Excluding expansion, remaining MRR from that cohort is $630 → GRR = 630 / 770 = 81.8%. Same cohort. NRR sits above GRR because expansion only ever adds.

Why boards need both

Keith Wallington (quoted on ChartMogul’s GRR page): chart GRR, NRR, and growth on one page for a board meeting. One page then shows the full customer journey: new business, retention, and upsell.

ChartMogul’s masking warning: 70% GRR with 115% NRR means a leaky base covered by expansion. Investors check for it in diligence, and your board should too. Fix GRR leaks before you celebrate NRR.

Bessemer’s Atlas churn piece: net dollar retention should be 100%+. Bessemer names net retention as the customer success North Star. That sets the NRR bar; you still need the GRR leak check.

Cited benchmarks (do not average)

SaaS Capital Research Brief 28 (2023)

Survey of private B2B SaaS; retention cuts exclude companies with less than $1M ARR unless noted.

CutMedian NRRMedian GRR
All surveyed SaaS (>$1M ARR)102% (unchanged vs 2022)91% (unchanged vs 2022)
ACV <$12k~100%~90%
ACV $12–25k~102%~90%
ACV $25–50k~103%~92%
ACV $50–100k~105%~93%
ACV $100–250k~107%~93%
ACV >$250k~110%~93%

Source: SaaS Capital, Research Brief 28: 2023 B2B SaaS Retention Benchmarks (Figure 1 and conclusions).

Other SaaS Capital lines from the same brief:

Bessemer segment aims (State of the Cloud 2019)

Aims by customer segment, not a survey median table.

SegmentACVGRR aimNRR aim
SMB<$12k70–80%80–100%
Mid-market$12–50k80–90%90–120%
Enterprise$50k+>90%>100%

Source: Bessemer, State of the Cloud 2019 (6 Feb 2019).

Bessemer Scaling to $100 Million: strong retention often cited as ~85%+ gross and ~120%+ net across lifetimes. Net retention ranges they publish include ~105–145% at $1–10M ARR and ~105–125% at $100M+; middle 50% still >100%. Average net ~140% at $1–10M then ~120% at $10–100M+. Mindbody IPO ~109% NRR on ~$2k ACV vs Okta 123% on $50k+ ACV. Different framing from SaaS Capital’s 2023 medians. Show both; do not blend.

ChartMogul Dec 2025 (separate sample)

CutNRR
Median B2B SaaS (~3,500 software companies analysis)82%
Top quartile B2B SaaS97%

Source: ChartMogul NRR page (updated 8 Sep 2026; cites Dec 2025 analysis). This is a different sample than SaaS Capital’s private B2B survey. Do not average 82% with 102%.

Best-in-class GRR (ChartMogul): over 86% at any stage (lose ~14% gross revenue per year). From the ChartMogul GRR page / guide. SaaS Capital’s peer-parity floor (~90%) is a different publisher’s bar. Label which one you mean on the slide.

What to put on the board slide

  1. Period. Usually trailing 12 months. If you use a quarter, say so and keep it consistent.
  2. Cohort. Starting customers only. No new logos mid-period. Say whether you use MRR or ARR.
  3. GRR and NRR side by side for that cohort, plus growth (Wallington).
  4. Trend. Last 4–8 quarters if you have them. One heroic month is noise.
  5. ACV or segment label so the board picks the right Bessemer / SaaS Capital band.
  6. Publisher footnotes. “SaaS Capital 2023 median” is not “ChartMogul Dec 2025 median.” Never average them into one target.

Labeled board sketches

Sketch A: healthy mid-market. Starting cohort MRR $1,000k. Contraction $40k, churn $50k, expansion $120k, reactivation $0. GRR = (1000 − 40 − 50) / 1000 = 91%. NRR = (1000 + 120 − 40 − 50) / 1000 = 103%. Near SaaS Capital’s overall medians. Board talk: hold GRR above ~90% and push expansion without buying it with discounts that later contract.

Sketch B: masked leak. Starting $1,000k. Contraction $80k, churn $220k, expansion $350k. GRR = 70%. NRR = 105%. Looks “fine” on net. ChartMogul’s 70% / 115%-style warning applies. Board talk: stop celebrating NRR until GRR recovers.

Sketch C: enterprise expansion engine. Starting $1,000k. Contraction $30k, churn $40k, expansion $200k. GRR = 93%. NRR = 113%. Sits near SaaS Capital’s higher-ACV medians and above Bessemer’s enterprise NRR floor (>100%). Board talk: protect GRR while expansion compounds.

Mistakes that waste a board meeting

Showing only NRR. Expansion can paper over churn. Bring GRR.

Mixing GRR and NRR labels. GDR/NDR are the same metrics under dollar names. Compare like with like.

Including new logos in the retention cohort. That inflates both numbers and measures acquisition instead of retention.

Averaging SaaS Capital 102% with ChartMogul 82% into “aim for 92%.” The samples differ. Pick the band that matches your ACV and say which source.

Treating Bessemer 2019 segment aims as 2023 survey medians. Aims ≠ medians.

Declaring victory at 100% NRR while GRR sits at 75%. Wallington’s three-line slide (GRR, NRR, growth) surfaces that faster than a single vanity net number.

Sources

Your board will read retention closely. Founders who have walked GRR and NRR through a raise or a board pack will pressure-test your cohort definitions in a FounderNexus session.