Customer Lifetime Value Calculator

LTV is built on margin. Using revenue is the error that ruins the number

Customer lifetime value asks a single question: across the whole relationship, how much money does this customer leave behind? Not how much they pay — how much stays. Those are different quantities, and confusing them is the most common mistake in the entire subject, precisely because nothing about the result looks wrong afterwards. The number just comes out bigger, and bigger numbers do not trigger suspicion.

Take the default scenario on this page. A customer bills $200 a month. At 80% gross margin, $160 of that is gross profit; after $10 a month of support and dedicated infrastructure, the contribution margin is $150. At 4% monthly churn, the correct lifetime value is $3,750. Run the same arithmetic on revenue and you get $5,000 — a third higher, from a single substitution.

That inflation does not stay contained. It flows into the LTV:CAC ratio, which is where it does the damage: $5,000 against a $1,200 CAC reads as 4.2:1, comfortably above the threshold everyone quotes. The honest number is 3.1:1, and once discounted it is 2.5:1. Same company, same month, three different verdicts — and only one of them is calculated on money the business actually has.

This is why the calculator above shows the revenue-based figure permanently, in red, rather than mentioning it in a footnote. Almost nobody knows they are making this mistake. The way to find out is to see both numbers at once.

One practical note on where to draw the line. Gross margin should absorb cost of goods: hosting, payment processing, third-party licences, whatever scales with usage. The separate cost-to-serve field is for real costs that sit outside that line in your accounts — support headcount, customer success, dedicated infrastructure. If those are already inside your gross margin percentage, leave the field at zero. Counting them twice is a quieter error than counting them zero times, but it is still an error.

Lifetime = 1 ÷ churn, and the assumption hiding underneath it

Average lifetime is the reciprocal of the churn rate. At 4% monthly churn the average customer stays 25 months; at 2%, fifty months. The formula is elegant, it is genuinely useful, and it rests on an assumption that is essentially never true: that the churn rate is constant over time.

It is not. Churn is heavily front-loaded. Customers who never onboarded properly, who bought for a project that ended, or who were sold to badly leave in the first few months. The ones still paying after two years have integrated the product into how they work and cancel at a fraction of the rate. Plot churn by cohort age and you get a decay curve, not a horizontal line.

This cuts both ways, and which direction depends entirely on where your input came from:

• Churn measured on new cohorts, extrapolated forever understates lifetime, sometimes by a lot. It takes the worst period of the relationship and assumes it repeats indefinitely.
• Churn measured across a mature base overstates lifetime for the customers you are acquiring today, because that blended rate is being held down by survivors who already passed the risky window.

Neither is a reason to abandon the formula — the alternative is fitting survival curves per cohort, which is the right answer for a company with a data team and overkill for everyone else. It is a reason to treat the lifetime figure as a planning number with a wide interval around it, and to be suspicious of any decision that flips when lifetime moves 30%. If you can only measure one thing better, measure churn separately for customers under twelve months and over twelve months, and run the calculator twice.

Monthly churn and annual churn do not convert by multiplying

This one is pure arithmetic, and it is wrong in an astonishing number of spreadsheets. Monthly churn does not become annual churn by multiplying by twelve. The correct conversion compounds:

annual = 1 − (1 − monthly)¹²

At 4% a month, the annual figure is 38.7%, not 48%. At 5% it is 46.0%, not 60%. At 8% — which sounds catastrophic and is common in SMB self-serve — it is 63.2%, not the impossible 96%. Each month's churn applies to a base that has already shrunk, so the losses do not add up in a straight line. Going the other way uses the same relationship inverted: 30% annual churn is 2.93% monthly, not 2.5%.

The reason this matters more than a rounding error is that the mistake usually shows up at the point of entry, not in the formula. Someone reads "30% annual churn" from a board deck, types 30 into a field labelled monthly, and every downstream number is off by an order of magnitude. The calculator above shows both readings of churn at all times, continuously, for exactly that reason — if the number in the echo line looks absurd, the number you typed is in the wrong unit.

Why there are two lifetime values on this page

Margin ÷ churn treats a dollar arriving in year seven as identical to a dollar arriving this month. For a customer who stays eighteen months that is a harmless simplification. Past roughly three years it stops being harmless.

The discounted version does the financially correct thing: it takes the expected margin in each future period, weights it by the probability the customer is still there to pay it, discounts it back at your cost of capital, and stops at a horizon you choose. On the default scenario that turns $3,750 into $2,959 — about 79% of the simple number.

The gap widens fast with lifetime, which is exactly the point. At 1% monthly churn, the average customer lasts a hundred months and the simple LTV reaches $15,000. Discounted at 10% a year over a ten-year horizon, it is around $7,382 — under half. The longer the contract you are modelling, the more the undiscounted number is telling you about a hypothetical future rather than about a decision you can make today.

The calculator splits that gap into its two causes rather than showing one lump. Part of it is truncation: value beyond your horizon that the simple formula counts and the discounted one does not. The rest is the time value of money. Those are different objections, and you may accept one and reject the other. If the horizon is doing most of the damage, extend it. If discounting is, that is not a modelling artefact — that is your cost of capital, and it is real.

The 3:1 rule: where it came from and what it does not say

An LTV:CAC ratio of at least 3:1 is the most-quoted benchmark in subscription businesses. It is worth knowing what it is: a venture-capital rule of thumb from the early 2010s, shorthand for the observation that gross margin has to cover more than acquisition. It also has to pay for R&D, general overhead, and the customers you win and then lose before they ever break even. Somewhere around three times CAC, there is usually enough left over for all of that. Below roughly 1:1 you are demonstrably destroying value.

What the rule does not say is at least as important. It says nothing about how the LTV was calculated — a 3:1 computed on revenue is often a 2:1 on margin, and the rule assumes margin. It says nothing about whether you can finance the CAC, which is a cash question rather than a profitability one. And a ratio far above 3:1 is not a victory lap: sustained 8:1 usually means a company is underinvesting in growth and leaving a market open, not that it has cracked something.

Treat it as a sanity check with a wide tolerance, and always alongside payback.

CAC payback is the number that decides whether you survive the quarter

LTV:CAC asks whether a customer is worth acquiring. CAC payback asks how long your money is gone before it comes back. For any company that is not sitting on unlimited cash, the second question is the binding one.

A business with 5:1 economics and a twenty-month payback is a business that runs out of money while growing. Every new customer consumes working capital that does not return inside the fiscal year, so growth itself is the thing draining the account. A business at 2.5:1 with a six-month payback recycles the same capital twice a year and compounds off a much smaller balance sheet. The first company looks better on the slide.

The calculator gives two payback figures. The nominal one is CAC divided by margin per period — the standard industry number, which quietly assumes nobody cancels before payback. The survival-adjusted one asks when the expected cumulative margin covers the CAC, accounting for the customers who leave along the way. It is always longer, and if the CAC exceeds the lifetime value entirely it reports that the CAC never pays back at all, rather than printing a large finite number that reads as "slow, but fine". It is not fine.

Filling this in without fooling yourself

• ARPU should come from customers who look like the ones you are acquiring now, not from a blended average dragged upward by legacy enterprise accounts.
• Gross margin is the accounting figure, not the aspiration. If you do not know it, revenue minus hosting, payment fees and third-party licences, over revenue, is close enough to start.
• Churn should be customer churn, not revenue churn, because the model counts customers. If you have meaningful expansion revenue, net revenue retention tells a different and also important story — but do not mix the two inside one formula.
• Discount rate is your cost of capital. For a funded company it is high; equity is expensive. Ten percent is a common floor, and using zero is a deliberate choice to see the undiscounted figure, not a neutral default.
• CAC must be fully loaded: paid media, sales salaries and commissions, the tools that support them, divided by customers actually won. Ad spend alone typically understates it by half or more.

What this tool does not do

It models a single customer segment with a constant churn rate. It does not fit survival curves per cohort, model expansion or contraction revenue, handle multiple price tiers with different retention, or account for referrals from existing customers. Those are real effects, and the tool that models them properly is a data warehouse, not a web page.

Everything runs in your browser as plain arithmetic. Nothing is uploaded, nothing is logged and there is no account — which is worth saying plainly, because ARPU, margin, churn and CAC together describe a company's unit economics about as completely as four numbers can.