Net Retention Rate: How Existing Customers Grow, Shrink and Leave

Net retention answers a single question: taking only the customers you had at the start of a period, are they collectively worth more or less to you at the end of it. Everything they did counts. Upgrades, extra seats, higher usage and second products push the figure up. Downgrades and cancellations pull it down. Customers acquired during the period are excluded, because including them would turn a retention measure into a growth measure.

The metric has become the headline retention figure in subscription businesses because it captures something real that older measures miss. A customer relationship is not binary. Accounts grow and shrink continuously, and a business whose customers steadily expand has a fundamentally different economic engine from one whose customers steadily shrink, even when both keep the same number of logos.

The calculation

Net retention equals opening recurring revenue plus expansion minus contraction minus churn, all divided by opening recurring revenue, times one hundred. Opening recurring revenue is the total for a fixed set of accounts on the first day of the period. Expansion is the increase in recurring revenue from those same accounts. Contraction is the decrease among those that survived. Churn is the recurring revenue lost from those that ended entirely. Nothing from a customer won inside the period appears in any term.

A worked example over twelve months

One hundred customers open the year paying 20,00,000 a month in total. Over the year, 63 of them expand, adding 3,60,000 a month between them. Fourteen contract, reducing spend by 90,000. Six cancel, taking 60,000 with them. Separately, the company wins 47 new customers worth 6,10,000 a month, and bills 4,00,000 in one-off implementation fees.

Net retention is 20,00,000 plus 3,60,000 minus 90,000 minus 60,000, all over 20,00,000, which is 22,10,000 divided by 20,00,000, or 110.5 per cent. The new customers and the implementation fees are excluded. Had the new business been included, the figure would read 141 per cent, which describes company growth rather than customer retention and cannot be compared with anyone else's retention number.

The relationship with gross retention

Gross retention for the same year is 20,00,000 minus 90,000 minus 60,000, over 20,00,000, which is 92.5 per cent. Net retention is 110.5 per cent. The 18 point gap is the expansion rate. That gap is the most informative single number in the whole exercise, because it tells you how much of the headline depends on customers continuing to grow rather than on customers simply staying.

The snapshot basis has a subtle problem worth naming, because it slides so easily into outright survivor bias. If the group is defined as accounts present at both ends of the period, then churned accounts are absent from both the numerator and the denominator, and churn has been silently removed from a retention metric. That single mistake can add many percentage points to the result and is close to invisible in a spreadsheet.

Survivor bias is the largest source of error and the least deliberate. It arrives whenever the underlying query joins opening and closing revenue on accounts that exist in both, which is the natural way to write it. The fix is to build the calculation from the opening cohort and treat missing closing rows as zero rather than as absent.

The second source is reactivation counted as expansion. An account that churned eighteen months ago and returned during the period was not in the opening cohort, so its revenue is not expansion, and adding it inflates both net retention and the expansion rate. The third is one-off fees. Implementation, migration, training and custom development are real revenue but not recurring, and putting them into the expansion term makes retention rise and fall with the professional services calendar.

Good net retention is broad. When you cut it by cohort year, plan tier and industry, the pattern is similar everywhere, which means expansion is a property of how the product is used rather than a property of a few unusual accounts. Gross retention sits close behind it, so the headline does not depend on continuing upsell. And it is stable across periods rather than spiking in the quarters where a price change or a big renewal landed.

Bad net retention can still be a high number, which is what makes it dangerous. The warning signs are concentration, where a small number of accounts supply most of the expansion; a wide gap to gross retention, which means heavy losses are being covered rather than fixed; and expansion driven mainly by renewal price increases rather than by customers using more. The last of those tends to reverse itself, because a customer who paid more without getting more is a customer who is now shopping.

Net revenue churn is the same calculation stated negatively, so net retention of 110.5 per cent is net revenue churn of minus 10.5 per cent. Gross retention removes the expansion term and gains a ceiling. Logo retention counts customers rather than revenue and can move in the opposite direction. Expansion revenue is the positive input rather than a competing metric. Dollar-based net retention and net dollar retention are alternative names for the same thing in currencies other than the rupee, and carry no methodological difference of their own.

Related reading on retention, pipeline discipline and the systems underneath both:what a CRM is, AI CRM, sales automation, CRM versus Excel, lead management software, and features overview.