Dalal Street Blogs provides information for educational purposes only and does not constitute investment advice. Please consult a SEBI-registered investment adviser before making any investment decisions.
CAUTION: Dalal Street Blogs publishes market news and educational content only. We do not promote, recommend or endorse any stock, and nothing here is a solicitation to buy or sell securities. Please consult a SEBI-registered adviser before investing.
Results

Net Revenue Retention: How to Read the Software Metric Behind Growth

3903
By Dalal Street Blogs EditorialUpdated 9 min read

Net revenue retention answers one blunt question about a subscription business: if the company signed no new customer for a year, would its revenue still grow? Take every customer it had twelve months ago, add up what they pay today, and compare that with what the same group paid back then. Above 100%, the existing base is growing on its own. Below 100%, the company has to sign new customers just to stand still.

The metric comes out of the software-as-a-service (SaaS) world, and it turns up in results calls, investor decks and the KPI pages of offer documents. The definitions and methods below follow the SaaS Metrics Standards Board’s NRR standard and its companion gross revenue retention standard. All the rupee figures in our worked examples are invented to keep the arithmetic easy, and don’t belong to any real company.

In this article
  1. What net revenue retention measures
  2. Method one: the cohort calculation
  3. Gross revenue retention: the number to read next to NRR
  4. Method two: the formula calculation
  5. Same NRR, very different businesses
  6. Reading NRR in results and investor presentations
  7. Net revenue retention in IPO documents and valuation
  8. A quick checklist for net revenue retention

What net revenue retention measures

The Standards Board defines NRR, which some companies call net dollar retention (NDR), as the percentage of recurring revenue a business keeps over a period. Two forces pull in opposite directions inside that one number.

  • Losses: customers who leave (churn), customers who cut usage, and customers who drop to a cheaper plan (contraction, or down-sell).
  • Gains: existing customers paying more through up-sells, cross-sells, price increases or heavier usage (expansion).

The word “net” is there because the losses are netted against the expansion. That’s the whole point, and also the catch, as we’ll see. A company can lose a painful chunk of customers and still print a healthy-looking NRR if the ones who stay are spending a lot more.

The standard notes that NRR is typically above 100% but can fall below it when revenue from existing customers is shrinking or expansion is thin. New customers are kept out of the calculation entirely. NRR is about the base you already had, nothing else.

Method one: the cohort calculation

The Standards Board calls the cohort method the most accurate approach and the preferred one in most situations. You freeze a group of customers at the start of the period, usually a year back, and track only them.

The formula is simple:

NRR = recurring revenue today from the customers you had at the start ÷ recurring revenue from those same customers at the start

The standard asks for recognised revenue, not bookings, billings or cash. Monthly recurring revenue (MRR) is the usual input, though enterprise software firms with long contracts can use annual recurring revenue (ARR) or contracted ARR instead.

Worked example: Company S and its five customers

Say Company S, a hypothetical Indian software firm, had five customers a year ago. Here’s their annual recurring revenue then and now, in ₹ lakh:

CustomerARR a year agoARR todayWhat happened
A4052Added seats
B3030Flat renewal
C250Left
D6051Moved to a cheaper plan
E4572Bought a second product
Total200205

NRR is 205 ÷ 200, or 102.5%. The base grew 2.5% without a single new logo.

Now suppose Company S also signed customer F during the year, who pays ₹35 lakh. Total ARR is ₹240 lakh, up 20% from ₹200 lakh. But F stays out of the NRR calculation. Of the ₹40 lakh of growth, only ₹5 lakh came from the old base. The other ₹35 lakh came from a new sale. That split matters, because new sales cost money to win, and the growth from old customers mostly doesn’t need a fresh sales pitch from scratch.

The standard’s own sample works the same way: ten customers whose MRR goes from $5,000 to $5,100 over a year, for an NRR of 102%, even though some of them churned and some contracted.

NRR tells you what the old customers are worth today. It says nothing about how many of them you lost along the way.

Gross revenue retention (GRR) uses the same cohort but throws away the good news. For each customer, you take the lower of the starting revenue and today’s revenue. Expansion is capped out; only churn and contraction remain. So GRR can never go above 100%.

Back to Company S. Cap each customer at its starting figure: A counts as 40 (not 52), B as 30, C as 0, D as 51, E as 45 (not 72). That adds up to 166. GRR is 166 ÷ 200, or 83%.

Read the two together and the story gets sharper. Company S lost 17% of its starting revenue to churn and down-sells, then won back 19.5 percentage points through expansion at two accounts. NRR of 102.5% looks fine. The 83% GRR says the base is leaky and leaning on two big customers to fill the gap.

In the Standards Board’s sample, the same customers that give 102% NRR give only 76% GRR. That’s a wide gap, and it’s exactly the kind of gap a single headline number hides.

Method two: the formula calculation

The second method skips customer-by-customer tracking and works from four totals:

NRR = (beginning recurring revenue − churned revenue + expansion revenue − contraction revenue) ÷ beginning recurring revenue

The standard’s sample uses beginning MRR of $100,000, churn of $9,000, expansion of $11,000 and contraction of $500. Plug those in and you get $101,500 ÷ $100,000, which works out to 101.5%.

According to the standard, this method suits companies with smaller contracts, shorter sales cycles and higher churn, and fast-growing businesses where many customers are less than a year old. It is often run over a month or a quarter, and it reacts faster to churn among new customers than the cohort method does.

The out-of-cycle problem

The formula method has a known weakness. Some of the churn or expansion that happens during the period may come from customers who weren’t in the beginning revenue at all. The standard gives a clean example: a customer who signs up after the start date and leaves before the end. Their churn hits this period’s NRR, even though their revenue was never in the denominator, so retention is understated. In the next period, their revenue sits in the beginning figure but the churn has already been counted, so NRR is overstated. The longer the measurement window, the bigger this error gets, which is why the standard calls the formula method a poor fit for annual numbers.

Same NRR, very different businesses

Here’s where investors trip up. Take two hypothetical companies, T and U, each starting the year with ₹200 crore of ARR.

₹ croreCompany TCompany U
Beginning ARR200200
Churned1440
Contraction610
Expansion4070
NRR110%110%
GRR90%75%

Both report 110% NRR. T keeps 90% of its starting revenue before any expansion. U keeps 75% and needs ₹70 crore of upsell to cover a ₹50 crore hole.

Now run a bad year. Say a slowdown halves expansion at both: T gets ₹20 crore, U gets ₹35 crore, and churn and contraction stay where they were. T’s NRR falls to 100%. U’s falls to 92.5%, so its base is now shrinking. The company with the stickier customers absorbs the hit. The one that was papering over churn with upsell doesn’t.

When expansion dries up, GRR is the floor you land on. A company with high NRR and weak GRR has a long way to fall.

Reading NRR in results and investor presentations

When a company puts NRR in a quarterly deck, a few questions sort the useful disclosures from the decorative ones.

Which method, and over what period?

The standard treats a year as the most intuitive and easiest period to benchmark. A shorter period can be annualised by raising the result to a power: the fourth power for a quarter and the twelfth for a month. A quarterly NRR of 103% compounds to about 112.6% a year (1.03 to the fourth power). A monthly 99% compounds to roughly 88.6% a year. Small monthly slippage turns into a big annual leak, so check which figure you’re looking at before you compare two companies.

Is usage revenue in or out?

The standard flags a live debate over whether usage above a customer’s committed level (overages) belongs in NRR. If variable usage is large and included, NRR starts behaving more like a revenue growth figure than a measure of how stable recurring revenue is. The standard’s view is that, if it is included, it should be clearly highlighted and applied consistently. For businesses that earn mostly from usage, it describes a “double back” variant: compare the last twelve months of revenue with the prior twelve, for customers who had revenue 24 months ago, to soften the ramp-up spike that new usage customers create.

When does a lost customer count as lost?

Many companies define a win-back period. The standard says most use 30 to 90 days: if a customer who didn’t renew comes back within that window, the revenue is treated as a renewal, not a new sale. A long win-back window can flatter retention a little. It’s worth knowing what the company uses.

Is GRR disclosed alongside?

This is the big one for us. A company that shows NRR quarter after quarter but never GRR is showing you the flattering half. If you can only get one, NRR is still useful. If you can get both, the gap between them tells you how much of the retention story depends on upsell.

Does it match the revenue line?

NRR should broadly square with what you see in the reported numbers. If a company claims 120% NRR and also says new customers are a big part of growth, yet total revenue is up only 10%, something isn’t adding up. Our guide on how to analyse quarterly results of a company walks through checking the revenue line against management’s commentary.

Net revenue retention in IPO documents and valuation

Indian offer documents disclose key performance indicators in support of the basis for the issue price, and SEBI’s ICDR Regulations (as amended up to March 21, 2026) set rules for them in Schedule VI. Every KPI and the terms used in it must be defined consistently and precisely in the offer document’s “Definitions and Abbreviations” section, in simple English. Each KPI must be approved by the company’s audit committee and certified by a peer-reviewed chartered accountant or cost accountant, and the company has to explain how management has used it to track performance. Our explainer on KPI disclosure in IPOs goes through that section in more detail. If NRR appears in an RHP, go straight to that definition: cohort or formula, MRR or ARR, overages in or out.

Retention also feeds into how we’d think about valuation. A business whose existing customers spend more every year needs fewer new customers for the same top-line growth, and the Standards Board describes companies with high NRR as more predictable, with higher growth potential. If you’re comparing subscription businesses on price to sales, retention is one of the things to look at before deciding whether a gap in multiples makes sense. It can also be the first metric to crack when customers start trimming budgets.

There’s a parallel in telecom, where ARPU tracks how much each user pays. ARPU is an average across all users. NRR follows a fixed group of customers. The two answer different questions, and neither replaces the other.

A quick checklist for net revenue retention

  1. Find the definition. Cohort or formula? Annual, quarterly or monthly? If it isn’t annual, compound it before comparing.
  2. Look for GRR. The gap between NRR and GRR is the share of retention that depends on expansion.
  3. Ask about overages and win-backs. Both can lift the number without making the base any stickier.
  4. Track the trend. One quarter’s NRR says little. A slide from 115% to 105% over four quarters says a lot.
  5. Square it with revenue. NRR, new customer growth and reported revenue growth should tell the same story.

NRR is one of the most useful numbers a subscription company can give you, precisely because it strips out new sales. Just don’t read it alone. Pair it with GRR and the revenue line, and you’ll see whether the base is really growing or just being propped up.

Disclaimer: This article is for educational purposes only and is not investment advice. The companies and figures in the examples are hypothetical. Investments in securities are subject to market risks; please consult a SEBI-registered investment adviser before making any investment decision.

Leave a Reply

Join the conversation

Commenting is open to registered readers. Sign up with just your email — no password needed, and it takes under a minute.

Sign up to comment Log in