Net Dollar Retention
The percentage of opening live annual recurring revenue you retain and expand from existing customers over a period, after churn and downgrades and after adding upsells — calculated on revenue that is billing today, not on contracted bookings.
◆ Percentage
Formula
Built from
What it measures
The net dollar movement in your existing customers' live annual recurring revenue over a period — churn and downgrades pulling down, upsells and price increases pushing up — expressed as a percentage of where that base started. Only customers active at period start are counted; new logos are excluded entirely. Because it runs on live ARR (the revenue that is billing today, already recognized on the P&L), NDR shows whether the revenue you actually collect from your existing base is growing or shrinking, independent of contracted bookings not yet recognized.
Why it matters
NDR is the cleanest read on whether your existing customer base is winning or losing on the revenue you actually bill. It separates the health of that base from new sales: when ARR is up 50%, the board's first question is "how much is expansion versus new logos?" — and NDR answers it on a live-revenue basis. It is dollar-weighted, so losing one large customer moves it far more than losing many small ones, which is exactly the signal finance and CS want. Because NDR sits on live ARR rather than contracted bookings, it is the metric that tracks what is hitting the income statement right now — making it the truth check against contracted retention (CNDR/NRR), which can run ahead of live revenue when recognition, refunds, or deferrals lag. NDR above 100% is the compounding engine of a land-and-expand business: you grow the recognized base even if you stop acquiring.
How to read it
Read NDR as a trend, never as a single month, and always alongside Gross Revenue Retention. NDR of 102% means your live base grew 2% from existing customers alone; 100% is flat; below 100% is shrinking. The gap between NDR and GRR is your net expansion — if NDR is 110% but GRR is 85%, upsells are masking a base that is leaking 15%, and you need to fix retention even while the headline looks healthy. Watch the month-over-month direction obsessively: 110% that was 118% last month is decelerating expansion or accelerating churn, regardless of how good the absolute number reads. The gap between NDR (live) and CNDR/NRR (contracted) is its own signal — when contracted retention runs hot but NDR lags, you have recognition headwinds, refunds, or a slow bookings-to-live transition draining the base before it reaches revenue.
What good looks like
Good
NDR comfortably above 100% — your live base is expanding faster than it churns, so you grow recognized revenue even before new sales. A sign of strong product-market fit and a working land-and-expand motion.
Watch
NDR hovering just above 100% — the live base is growing only modestly. Check whether upsell velocity is slowing, churn is creeping up, or you're landing lower-value segments before the cushion disappears.
Bad
NDR below 100% — the existing live base is shrinking and new bookings are masking a leaky bucket. Unsustainable; dig into churn root causes and downsell patterns immediately.
Watch-outs
- Reading NDR without GRR. A 115% NDR can sit on top of an 85% GRR, meaning 15% of the live base is leaving and only heavy upsell is hiding it. Always split NDR into GRR plus net expansion before you call retention healthy.
- Treating the absolute number as the signal. NDR of 110% this month versus 118% last month is a decline the headline flatters — expansion is decelerating or churn is accelerating. Track the month-over-month trend obsessively, not the snapshot.
- Comparing NDR directly to contracted retention and drawing wrong conclusions. NDR (live ARR) and CNDR/NRR (contracted CARR) measure the same motion on different revenue bases; a gap usually means refunds, deferrals, or recognition timing, not a data error. Always read both to see where revenue is truly landing.
- Counting new-logo revenue in the numerator. NDR is existing customers only; if a new customer's ARR leaks into upsell, you inflate retention and lose the ability to separate base health from acquisition.
- Ignoring customer concentration. One whale's large upsell can lift NDR to 120% while quietly making that account a third of your base. Segment NDR by cohort, size, and territory so a single contract doesn't disguise rising concentration risk.
Worked example
Hypothetical
You open March with $500K in live ARR. During March two customers churn for $40K total, one downgrades by $10K, and three expand their plans for $60K. Net dollar retention is ($500K + $60K − $10K − $40K) ÷ $500K = $510K ÷ $500K = 102% — the live base grew 2% in the period from existing customers alone. Had those upsells been only $30K, NDR would be $480K ÷ $500K = 96%, signaling real base erosion despite the expansion.
Variants & windows
The same metric re-expressed by a mechanical transform — a trailing window, a growth rate, a per-unit scaling, or a book/segment cut. Each is computed from Net Dollar Retention above.
- NRR Growth Rate (MOM) Growth rate
- NRR (T3M) Trailing 3-month
- NRR Growth Rate (T3M) Growth rate · Trailing 3-month
- NRR (TTM) Trailing 12-month
- NRR Growth Rate (TTM) Growth rate · Trailing 12-month