NRR vs GRR
Net and gross revenue retention answer different questions — what each one hides, and why you need to read them together.
What's the difference?
Both metrics follow the same cohort of existing customers over a period and ask how much of their revenue you kept. The difference is one term: NRR counts expansion, GRR does not. GRR can never exceed 100% — it only measures what survived. NRR can, because upgrades are allowed to offset losses.
That makes them answers to different questions. GRR: how durable is the revenue base if nobody ever upgrades? NRR: left alone, does the existing book grow or shrink? A 105% NRR on a 90% GRR means expansion is papering over a 10-point leak.
How to calculate NRR and GRR?
Start from the same cohort MRR. GRR = (starting MRR − churned MRR − contraction MRR) ÷ starting MRR: a $20,000 base losing $1,200 to cancellations and $800 to downgrades retains 90%. NRR adds expansion back in: with $3,000 of upgrades, ($20,000 − $2,000 + $3,000) ÷ $20,000 = 105%.
The gap between them — the gross-to-net spread, 15 points here — is precisely how much expansion is contributing. Both numbers must be computed on the same cohort and period, or the comparison is meaningless.
Should you watch NRR or GRR?
Both — for different failure modes. NRR is the headline investors quote, but it can be flattered by a handful of big upgrades while the base quietly erodes; two companies at 105% NRR are very different businesses if one holds a 95% GRR and the other an 85%. GRR is the floor: it tells you what happens the quarter expansion stalls.
The practical reading order: GRR tells you whether retention needs fixing, NRR tells you whether the growth model works. A GRR below 85% is a retention problem no expansion motion should be allowed to hide; an NRR below 100% on a healthy GRR is an expansion problem — a pricing model with no axis to grow along.