Net User Retention

What it measures

How your existing user base moves in a period once you net retention, expansion, contraction, and churn together. It isolates the health of the cohort you already had — new acquisition is deliberately excluded, so a high number means current users are finding more value, not that sales filled the gap.

Why it matters

Net User Retention is the user-count twin of Net Revenue Retention, and it answers a question revenue can hide: is the core base healthy? You can post strong NRR while quietly bleeding users if the survivors pay more — great for margin, a red flag for product-market fit. Investors read NUR as a durability signal: a base that expands on its own compounds, while a base that needs constant acquisition to stay flat is a leaky bucket no sales engine outruns. It moves before revenue does, which makes it an early warning.

How to read it

Read NUR as a ratio centered on 1.0, always as a trend. Above 1.0, your existing users net-expanded — retention plus upsell beat churn plus downsell. At exactly 1.0, expansion is perfectly offsetting losses; the base is flat, not safe. Below 1.0, the cohort is contracting and acquisition is masking the leak. Decompose every move into its four parts: NUR can hold at 1.0 with tiny churn and tiny expansion (stable) or with huge churn barely covered by huge expansion (volatile). Same number, opposite stories — so never read it without the components.

What good looks like

Watch-outs

• Mixing the periods. The denominator must be Starting Users (last period's close). Pair a monthly numerator with a quarterly base and the ratio is meaningless — keep numerator and denominator on the same clock.
• Dropping expansion or contraction. If you only compute (Starting − Churned) ÷ Starting you've built Gross User Retention, not NUR, and you lose the entire signal about whether surviving users are buying more or less.
• Reading the ratio without its parts. NUR = 1.0 from low churn and low expansion is a calm base; NUR = 1.0 from heavy churn barely covered by heavy expansion is a volatile one. Always decompose the four components.
• Comparing cohorts of different ages. A one-month-old cohort churns harder than a year-old one, so their NUR differs even when product health is identical — control for cohort age before comparing.

Worked example