पाठशाला Pathshala · हिसाब Hisāb, Unit economics · Lesson 03 · Start

The unit economics of SaaS, with the benchmarks

Three numbers decide whether a software company works: gross margin, CAC payback and net revenue retention. How to compute each, the benchmarks as their authors stated them, and what an Indian SaaS investor expects to see.

Pathshala, The Founder Library · 11 October 2026 · 9 min read

Software is the business model with the best unit economics ever invented and the easiest to misstate on a slide. The product costs nothing to copy, the customer pays every month and the same code serves the thousandth customer as the first. All of that is true, and none of it tells you whether a particular company works. Three numbers do: gross margin, CAC payback and net revenue retention.

This lesson shows how each is computed in a diligence room rather than a pitch, states the benchmarks as the people who published them stated them, and ends with what is specific to an Indian company selling software, which is a cost advantage that is also a higher bar. The tool in the middle runs your own numbers.

Gross margin: what counts as cost of revenue

Gross margin is revenue less the cost of delivering the service. For software that cost is wider than founders like to admit: cloud hosting, the third-party APIs the product calls, payment-processing fees, and the salaries of the people who onboard customers and answer their tickets, because without them the service is not delivered. Since 2023 it has also included the inference bill from any language model inside the product, which for some companies is now the largest single line. A founder who books only servers under cost of revenue and puts customer success in operating expense has produced a margin that will be recomputed in the first week of diligence.

The benchmarks are consistent across sources. Andreessen Horowitz’s 16 More Startup Metrics says software companies should have very high gross margins, in the eighty to ninety per cent range. Bessemer Venture Partners’ Scaling to $100 Million, drawn from its own cloud portfolio between 2010 and 2021, puts the average at 65 to 70 per cent regardless of stage and its top-performing companies at 80 per cent or above. Read those two together: seventy is the floor an investor expects, eighty is what good looks like, and anything under sixty is a services company wearing software clothes, which is not a criticism of services but is a different business with a different multiple.

The common way an Indian company falls under the floor is implementation. A ₹30 lakh annual contract that includes three months of configuration and training by four engineers has a gross margin closer to a consultancy than to software until the configuration is productised. Separate the two on the invoice and in the accounts, report software margin and services margin apart, and tell the investor which one is growing.

CAC payback: the month the customer has paid for themselves

CAC payback is the number of months of gross profit from a new customer needed to recover the cost of winning them. Compute it as acquisition cost divided by monthly revenue per customer times gross margin. Acquisition cost is everything spent on sales and marketing in a period, salaries included, divided by the customers won in that period on paid channels; the lesson on [CAC, LTV and payback](/library/cac-ltv-and-payback-the-three-numbers) has the argument for excluding organic customers from the denominator.

A company in Chennai sells a compliance product to Indian small businesses at ₹25,000 a month. It spends ₹40 lakh a month on a sales team and advertising and signs twenty new customers, so CAC is ₹2 lakh. At 78 per cent gross margin each customer returns ₹19,500 a month in gross profit, and payback is 10.3 months. The same company selling the same product to American small businesses at $500 a month with a $4,000 CAC and the same margin pays back in ten months. Both are inside the line. Neither is remarkable.

The line comes from David Skok’s SaaS Metrics 2.0, published in January 2013, which showed profitability becoming anaemic once the time to recover CAC extended beyond twelve months and observed that many of the best SaaS businesses recover it in five to seven. Bessemer’s data refines it by customer: target under twelve months for SMB, under eighteen for mid-market and under twenty-four for enterprise, and it reports that the average company between $1 million and $10 million of ARR actually took fifteen months. The reason the target shortens as the customer gets smaller is churn. A small business that leaves in month fourteen has repaid an eighteen-month CAC only in the founder’s imagination.

Net revenue retention: whether the base grows on its own

Net revenue retention takes every customer who was paying twelve months ago, adds what they pay today including upgrades and extra seats, subtracts contraction and churn, and divides by what they paid then. New customers are excluded. A figure above a hundred per cent means the existing base grew without a single new sale; below a hundred means the company is refilling a bucket before it can grow it. Skok called the state above a hundred negative churn and named it the ultimate solution to the churn problem.

The benchmarks here are the widest-ranging of the three, because the number depends heavily on whom you sell to. Bessemer’s top performers between $1 million and $10 million of ARR ran at 145 per cent or more; its portfolio average fell from around 140 per cent at that stage to about 120 per cent beyond $10 million; gross retention, which counts only what was lost, sat between 85 and 90 per cent throughout. Those are enterprise and mid-market figures. For a reference point closer to home, Freshworks, founded in Chennai and selling largely to small and mid-sized businesses, reported net dollar retention of 108 per cent for the fourth quarter of 2025 on $838.8 million of annual revenue. An SMB company at 108 per cent is doing well. An enterprise company at 108 per cent has a question to answer.

Compute NRR by cohort, not as one blended number, because a blended figure lets the expansion of three large accounts hide the quiet departure of thirty small ones. A table with acquisition quarter down the side and months since acquisition across the top, each cell the cohort’s revenue as a share of its starting revenue, is the single most informative page in a SaaS data room.

Set the tool to the Chennai company: CAC ₹2 lakh, ₹25,000 a month, 78 per cent margin. Then move churn. At two per cent a month the customer lives fifty months and the lifetime gross profit is almost five times CAC. At five per cent, which is what selling to Indian small businesses often looks like in the first year, life is twenty months and the ratio is under two. Nothing changed except how long customers stay, and the company moved from fundable to not. That is why NRR sits beside payback and not behind it.

What an Indian SaaS investor expects

There is no published, verifiable set of benchmarks specific to Indian SaaS; the reports that circulate are summaries behind forms, and the figures founders quote from them cannot be checked. What can be checked is the structural argument investors make. In a December 2021 McKinsey conversation with SaaSBoomi’s founding partner, Turning India into a SaaS power, the case was put in two numbers: American SaaS companies typically spend 40 to 50 per cent of revenue on sales and marketing, and a hybrid model with an India-based go-to-market team could cut that to 25 to 30 per cent. The same conversation set the growth expectation: Indian SaaS companies growing 50 to 70 per cent a year while losing 10 to 20 per cent of revenue in free cash flow, settling toward 40 to 50 per cent growth past $100 million.

Read what that implies for the three numbers. Gross margin is judged against the global benchmark, because the product competes globally and the hosting bill is in dollars either way. NRR is judged against the global benchmark for the customer segment, because retention is a property of the customer, not the vendor’s address. CAC payback is where the Indian advantage is supposed to show: with sales and marketing at 25 to 30 per cent of revenue instead of 45, an Indian company selling at global prices should pay back faster than its American competitor, not at the same pace. An Indian company with Indian costs and a fifteen-month payback has spent its structural advantage somewhere and the investor will want to know where.

The second thing an Indian investor asks is who the customer is, and it is asked first because it decides the other three. Software sold in dollars to businesses abroad is priced against global alternatives and retains like a global product. Software sold in rupees to Indian small businesses is priced against a WhatsApp group and a spreadsheet, churns when the owner’s nephew builds something, and has to be won at a CAC that the ₹25,000 a month can repay. Both can work. They are different companies and should be modelled as such from the first spreadsheet.

An Indian SaaS company gets the world’s prices and India’s costs. The benchmark is not to match the global payback. It is to beat it.

Three ways the numbers get flattered

Services booked as ARR. Annual recurring revenue is revenue that recurs. A one-time implementation fee, a data migration, a custom integration paid for once are not ARR however they are invoiced, and an investor who finds them inside the ARR line will discount every other number on the page.

Annual prepayment hiding churn. A customer who paid for a year in March and stopped using the product in July is still in the retained base until next March. Count usage alongside revenue; a cohort whose logins have halved will show up in NRR two quarters from now, and the founder should not be the last to know.

LTV computed on revenue. Lifetime value is gross profit over the customer’s life, not revenue. At 78 per cent margin the difference is a fifth of the number; at 55 per cent it is nearly half. The [CAC, LTV and payback](/library/cac-ltv-and-payback-the-three-numbers) lesson shows the same customer passing the three-times test on revenue and failing it on gross profit.

A monthly ritual, in thirty minutes

On the fifth working day of each month, once the books are closed, do four things. Recompute gross margin from the actual cloud, API and inference invoices and the actual payroll of the people who onboard and support customers; do not carry the plan figure forward. Compute CAC for the month on paid channels and divide by monthly gross profit per new customer to get payback. Add a column to the cohort table and read NRR for every cohort older than twelve months. Then put the three on one line with last month’s three beside them.

The rule that follows is simple. If margin is above seventy, payback inside the line for your segment and NRR above a hundred, the company can buy growth and the question is how fast. If payback is inside the line but NRR is below a hundred, the product has a retention problem that more sales will make more expensive; fix it before scaling the team. If margin is under sixty, find out what is in cost of revenue that should not be there, or accept that you are building a services company and price it like one.

Nothing here is legal, tax or investment advice. The benchmarks are stated as their authors stated them and come from portfolios that are not random samples; your investor will carry a version of them anyway, so carry the originals.

Sources

1. David Skok, SaaS Metrics 2.0, forEntrepreneurs, January 2013 — LTV above 3× CAC; profitability anaemic once CAC recovery exceeds twelve months; the best recover in five to seven; negative churn.
2. Mary D’Onofrio and Ethan Ding, Scaling to $100 Million, Bessemer Venture Partners, September 2021 — Gross margin average 65–70 per cent and top performers 80 per cent+; CAC payback targets under 12, 18 and 24 months by segment and a 15-month average at $1–10M ARR; net retention and gross retention ranges.
3. Anu Hariharan, Frank Chen and Jeff Jordan, 16 More Startup Metrics, Andreessen Horowitz, September 2015 — Software gross margins should be in the 80–90 per cent range.
4. McKinsey & Company, Turning India into a SaaS power, McKinsey on Start-ups podcast, December 2021 — US sales and marketing at 40–50 per cent of revenue against 25–30 per cent for a hybrid India model; growth and free-cash-flow expectations for Indian SaaS.
5. Freshworks Inc., Fourth Quarter and Full Year 2025 Results, February 2026 — Net dollar retention of 108 per cent; FY2025 revenue of $838.8 million.