पाठशाला Pathshala · हिसाब Hisāb, Unit economics · Lesson 01 · Build

CAC, LTV and payback: the three numbers that decide whether you should grow

What it costs to win a customer, what that customer returns in gross profit, and how many months until the first number is repaid. Get these three honest before you spend a rupee on growth.

Pathshala, The Founder Library · 10 October 2026 · 9 min read

Every growth decision is a bet that a customer returns more than they cost to win. Three numbers settle the bet: the cost to acquire a customer, the gross profit that customer returns over their life, and the month in which the second repays the first. Most founders can recite them. Far fewer compute them in a way an investor would accept.

This lesson sets out the arithmetic as it is done in a diligence room, the two places Indian companies most often get it wrong, and the cohort method that stops the numbers flattering you. The tool in the middle lets you run your own figures.

CAC: paid, not blended

Customer acquisition cost is the total spent on sales and marketing in a period divided by the number of new customers won in that period. The total includes salaries of the sales and marketing team, agency fees, advertising, tools, referral payouts and discounts given to close. A company spending ₹30 lakh a month on all of that and adding 250 customers has a CAC of ₹12,000.

The trap is the denominator. If 150 of those 250 customers arrived through word of mouth, search or a founder’s network, dividing by 250 gives a blended CAC that says nothing about what the next customer will cost. Divide the paid spend by the paid customers instead. Andreessen Horowitz’s 16 Startup Metrics puts it plainly: investors consider paid CAC more important than blended CAC in judging whether a business is viable, because blended CAC hides how well the paid channels actually perform. Anu Hariharan, then at Y Combinator, gave the same warning to Startup School in 2019 with a sharper example: a blended CAC of twelve could be as high as seventy on the paid channel alone.

Report both if you like. Decide on paid. Organic customers are a gift that does not scale on command; the paid number is what you are proposing to pour money into.

LTV: gross profit, not revenue

Lifetime value is what a customer returns before they leave. The simplest version is monthly revenue per customer multiplied by the number of months they stay, and the number of months is one divided by monthly churn. A customer paying ₹1,500 a month with four per cent monthly churn stays twenty-five months on average and returns ₹37,500 in revenue.

That figure is wrong for the purpose, and the error is the most common one in Indian pitch decks. Revenue is not what repays acquisition. Gross profit is. Serving the customer costs something every month: cloud, payment gateway fees, support staff, delivery, the goods themselves if you sell goods. Multiply revenue by gross margin first. At seventy per cent gross margin the same customer returns ₹1,050 a month in gross profit and ₹26,250 over their life, not ₹37,500. Bill Gurley made this point in 2012 in The Dangerous Seduction of the Lifetime Value Formula: it is critical to bundle all future variable costs of supporting the customer, and the formula is at best a good guess about how the future will unfold. The a16z piece goes further and defines LTV as the present value of future net profit from the customer, and names revenue-based LTV as a mistake.

Churn deserves the same scepticism. Four per cent a month sounds small and is thirty-nine per cent a year. A company six months old does not know its churn; it knows that nobody has left yet. Use the oldest cohort you have, assume it will get worse before it gets better, and never let a projected lifetime run past three years. Few customers in India are that loyal to a product that is still being built.

The 3× rule, and where it came from

The benchmark every investor carries is that LTV should be at least three times CAC. Its most cited origin is David Skok’s SaaS Metrics 2.0 on forEntrepreneurs, first published in January 2013, which observed that the best SaaS businesses have an LTV to CAC ratio higher than three, sometimes as high as seven or eight. Skok called them guidelines, not laws, and they were drawn from American software companies. They became the rule of thumb anyway, and you will be measured against it.

The logic behind three is worth knowing so you can argue with it. One times means the customer only ever repays what they cost, with nothing left for product, overheads or profit. Two leaves a thin margin that one bad quarter of churn erases. Three means that for every rupee spent on acquisition two come back to run the company, and it leaves room for the LTV estimate to be a third too optimistic, which it usually is. A ratio above five or six is often a sign of the opposite problem: the company is under-investing in growth and could spend more.

Run the example above. At ₹12,000 CAC and a revenue-based LTV of ₹37,500 the ratio is 3.1 and the slide looks fine. At the gross-margin LTV of ₹26,250 it is 2.2 and the slide is a problem. Same customer, same spend; the only difference is whether gross margin was applied. That gap is the whole lesson.

The curve is cumulative gross profit from one customer. It starts below zero by the CAC and climbs as monthly gross profit arrives, with each month weighted by the share of customers still alive. Churn shows up as the curve flattening, which is why a customer who should repay in eleven months on simple arithmetic crosses the line nearer sixteen when churn is counted. Move gross margin from seventy to forty per cent and watch the crossing point disappear off the right edge.

Payback: the number that decides the cash

LTV to CAC is a ratio and ratios do not pay salaries. Payback is the number that decides whether you can afford to grow: the month in which the cumulative gross profit from a customer crosses the CAC that bought them. Simple payback is CAC divided by monthly gross profit; honest payback weights each month by churn, as the tool does.

The benchmarks differ by model, and here the sourcing is thinner than founders assume. For software sold to businesses Skok’s data showed profitability becoming anaemic once payback stretched past twelve months, with the best companies recovering CAC in five to seven. Twelve months is the line most seed investors draw for SaaS. For consumer subscription the rule of thumb investors repeat is six months or less, because consumer churn is higher and lifetimes shorter, so the window to recover the spend is narrower. For marketplaces payback is computed on take rate, not transaction value, and a marketplace that has to show it must show it on the gross profit of its commission. Treat the six-month and the marketplace rules as what experienced investors say, not as published data; the twelve-month SaaS line has the data behind it.

Why payback matters more than the ratio at the build stage: a customer who returns three times their CAC over four years still has to be funded for the first eighteen months. If payback is eighteen months and you add a hundred customers a month at ₹12,000 each, you are lending your customers ₹12 lakh a month and getting it back a year and a half later. That is a working-capital business, and it needs equity or debt sized to it. A payback of eight months with the same ratio needs less than half the capital to grow at the same rate. Payback is the bridge between unit economics and runway.

LTV to CAC tells you whether to grow. Payback tells you whether you can afford to.

The India-specific trap: GMV is not revenue

Indian marketplaces, aggregators and commerce companies have a decade-long habit of reporting gross merchandise value as if it were revenue, and of computing unit economics on it. Between 2013 and 2015 the largest e-commerce companies in the country ran on GMV targets and said so in public, and the sector spent the years after 2016 unlearning it. The habit survives in seed decks.

The a16z metrics piece puts it in four words: GMV does not equal revenue. Revenue is the part of GMV the platform keeps. Hariharan used Airbnb as the illustration: of the money that moves through the platform, the company probably makes twelve per cent, and that twelve per cent is the net revenue on which every ratio should be built. A marketplace moving ₹10 crore a month through its platform at a twelve per cent take rate has ₹1.2 crore of revenue, and after payment gateway, logistics and support perhaps ₹60 lakh of gross profit. A CAC that looks like three weeks of GMV is in truth eight months of gross profit. Build LTV on the gross profit of the take rate, not on the transaction, and say so on the slide, because the investor will recompute it either way and would rather you had done it first.

The same discipline applies to discounts and cashbacks. A ₹200 cashback to close a first order is CAC, not marketing. A subscription sold at ₹999 for the first year and ₹2,999 thereafter has an LTV built on ₹2,999 only if customers actually renew at ₹2,999, which the first cohort has not yet proven.

Cohort payback: the method that keeps you honest

A single CAC and a single LTV describe an average customer who does not exist. The method investors use, and the one you should run monthly, is cohort payback: take every customer acquired in one month, record what it cost to acquire them in total, and then track the cumulative gross profit from that group month by month until it crosses the acquisition cost. Do the same for the next month’s cohort, and the next. The result is a table with acquisition months down the side and months-since-acquisition across the top, and a diagonal line where each cohort paid back.

The a16z follow-up, 16 More Startup Metrics, describes what good looks like in such a table: retention in each cohort stabilising after six or twelve months rather than sliding to zero, and newer cohorts performing progressively better than older ones. Add payback to the same table and two further questions become answerable. Is payback getting shorter as the product improves, or longer as you push into channels that convert worse? Did the cohort acquired during the festive-season spend pay back at all, or did the discount buy customers who left in January?

Cohort payback also settles the blended-versus-paid argument by itself. Split each month’s cohort by channel and the paid cohorts will show their own payback line. If it is twice as long as the blended one, the blended number was decoration.

A monthly ritual, in thirty minutes

On the fifth working day of each month, once the previous month’s books are closed, do four things. First, compute paid CAC for the month just ended: paid sales and marketing spend over paid customers won, with discounts and cashbacks counted as spend. Second, update the cohort table with one more month of gross profit for every cohort, and mark which ones crossed payback. Third, recompute gross margin from the actual cost of serving customers, not from the plan. Fourth, put the three numbers into the tool above or your own sheet and read the ratio and the payback month.

Then apply the rule. If LTV to CAC is above three on gross profit and payback is inside twelve months, growth spend can rise, and the question becomes how fast. If the ratio is above three but payback is beyond eighteen months, you have a working-capital problem, not a model problem: grow at the pace your cash allows or raise money sized to the gap. If the ratio is below three on gross profit, do not scale spend. Fix the ratio first, through price, margin, churn or a cheaper channel, and that is a product and pricing conversation, not a marketing one.

None of this is financial advice. It is arithmetic, with the benchmarks stated as the people who coined them stated them. Read Gurley’s essay before you believe any LTV, including your own.

Sources

1. Jeff Jordan, Anu Hariharan, Frank Chen and Preethi Kasireddy, 16 Startup Metrics, Andreessen Horowitz, August 2015 — Paid versus blended CAC, LTV as net profit, GMV does not equal revenue.
2. Anu Hariharan, Frank Chen and Jeff Jordan, 16 More Startup Metrics, Andreessen Horowitz, September 2015 — Cohort analysis and what a healthy cohort table looks like.
3. Bill Gurley, The Dangerous Seduction of the Lifetime Value (LTV) Formula, Above the Crowd, September 2012
4. David Skok, SaaS Metrics 2.0, forEntrepreneurs, January 2013 — The origin most cited for LTV above 3× CAC and payback inside twelve months.
5. Y Combinator, Startup School Week 3 Recap: Anu Hariharan and Adora Cheung, September 2019 — Do not blend CAC; GMV is not revenue; measure unit churn for consumer and revenue churn for SaaS.
6. Itika Sharma Punit, Indian e-commerce dumps its once-favourite baby GMV, Quartz India, February 2017