NRR vs GRR explained: how to read net and gross revenue retention

TL;DR. Net Revenue Retention (NRR) and Gross Revenue Retention (GRR) are two of the most-cited SaaS metrics and the two most-confused. NRR includes expansion (upsells, cross-sells, seat growth); GRR doesn't. A company can post a beautiful NRR of 130% while quietly losing half its logo base — the expansion from the survivors hides the damage. This guide unpacks both metrics, shows how to compute them honestly, and explains what good looks like by stage and segment.

The formulas

Gross Revenue Retention (GRR): (Starting ARR − Downgrades − Churn) ÷ Starting ARR. GRR is capped at 100% — you cannot exceed your starting point because expansion is excluded. It measures how sticky your product is for the customers who started the period.

Net Revenue Retention (NRR): (Starting ARR − Downgrades − Churn + Expansion) ÷ Starting ARR. NRR can exceed 100% — if existing customers expand more than others contract or churn, the cohort is net-growing.

Example: you start a quarter with $1M ARR from existing customers. During the quarter, $80k churns, $30k downgrades, $250k expands (upsells, more seats, plan upgrades). GRR = ($1M − $30k − $80k) ÷ $1M = 89%. NRR = ($1M − $30k − $80k + $250k) ÷ $1M = 114%.

Why most SaaS dashboards report NRR but not GRR

NRR is the metric investors ask for. It's the headline number in every series B and series C deck. A 130% NRR signals capital-efficient growth: existing customers fund acquisition through expansion, so the business compounds without proportional sales spend.

GRR is the metric that should keep operators up at night. It measures whether your product genuinely solves a problem worth keeping. A high NRR with a mediocre GRR means you're winning on price discipline and account management — but the underlying product loyalty is weak. When growth slows or expansion motions saturate, the GRR truth surfaces and revenue collapses faster than the headline suggested.

What good looks like

Public SaaS leaders (Snowflake, Datadog, MongoDB at peak): NRR 130-180%, GRR 90-97%. Expansion motions are real and the underlying product is sticky.

Best-in-class private B2B SaaS: NRR > 120%, GRR > 90%. Aim here if you're building a category leader.

Median B2B SaaS: NRR ~105%, GRR ~85%. This is workable but the runway is shorter — every point of expansion has to be earned through active account management.

Sub-90% NRR / sub-80% GRR: the business is on a treadmill. Acquisition has to fund both growth AND replacing churned customers. The math gets unforgiving fast.

B2C SaaS: the bars are lower — NRR > 100% is good, GRR > 75% is acceptable. Consumer churn is structurally higher.

The divergence trap

Watch the gap between NRR and GRR. A healthy SaaS business shows a divergence of 15-25 points (NRR 115%, GRR 90%). A divergence over 35-40 points (NRR 130%, GRR 85%) is a warning sign: your expansion motion is doing all the work and the customer base is shrinking underneath.

When the gap is wide, ask: is expansion concentrated in a small number of mega-accounts? Is the underlying logo retention masking a churn problem? If you removed your top 5 expanding accounts, would NRR collapse below 100%?

Cohort-based vs blended reporting

Most dashboards report blended GRR — a single number across the entire customer base. This is misleading. New cohorts haven't had time to churn yet; mature cohorts have. Blended GRR systematically over-reports retention.

The honest measure is cohort-based: pick customers who started in a specific quarter and track what percentage of their starting ARR remains 12 months later. A 12-month cohort GRR of 88% means that 88¢ of every dollar from that quarter's acquisitions is still there a year later, ignoring expansion.

Run cohort GRR for the last 8 quarters and overlay them. If older cohorts retain better than newer ones, your ICP has drifted or your product fit is decaying. If newer cohorts retain better, your onboarding or product improvements are working.

Segmenting NRR and GRR

Blended NRR is the most dangerous number in SaaS. Segment by:

Plan tier. Enterprise plans usually retain at 95%+ GRR. SMB tiers often sit at 70-80%. A blended 88% hides both.

Acquisition channel. Inbound-acquired customers retain better than outbound-acquired customers in most B2B SaaS. Paid-acquired customers retain worst.

Cohort vintage. As above.

Industry vertical. If you serve multiple verticals, retention varies wildly. Knowing your healthcare GRR is 95% and your retail GRR is 72% is a strategic input — you might want to double down on healthcare and stop selling into retail.

Action items

1. Compute both NRR and GRR for the last 4 quarters. Cohort-based, not blended.

2. Segment by plan, channel, and vertical. Find your worst-retaining segment.

3. If NRR-GRR gap is > 30 points, run a logo retention analysis and find the churn drivers.

4. Track GRR as your North Star for product-market fit. Track NRR as your North Star for capital efficiency. They're different jobs and need different dashboards.

Read more on the cost side: customer retention vs acquisition cost. For the foundational calculation, see how to calculate customer retention rate.