Analytics
LTV to CAC ratio calculator for early-stage startups
By Charles Summers · Updated · Free, no signup
Short answer
Before you have a cohort old enough to have lived through the lifetime you are claiming, the ratio is a number with an error bar wider than any decision it could inform: two cancellations out of forty customers put monthly churn somewhere between roughly 1.4 and 16.5 percent at ordinary confidence, which is a customer lifetime anywhere from six months to six years and a ratio anywhere from 0.9 to 11 on identical data. Report retained gross profit at the age your oldest cohort has actually reached, and say how old that cohort is.
Use the ltv to cac ratio calculator
Why early stage startups need a different approach
There is a specific point at which this ratio starts to mean something, and it is later than almost every founder quoting it believes. The formula turns a churn rate into a lifetime by dividing one by it, which magnifies whatever uncertainty is in that rate rather than averaging it away. A rate estimated from a small number of customers over a short window has a great deal of uncertainty in it, and the division converts a modest range of plausible churn rates into a range of lifetimes spanning years.
That is not an argument for ignoring unit economics early. It is an argument for reporting the parts you have measured and being explicit about the part you have not. A company nine months old has genuinely observed nine months of retention, nine months of gross profit per account, and whether any cohort has yet returned what it cost to acquire. Those are facts. A thirty-four-month lifetime derived from them is an extrapolation carrying a decimal point it has not earned, and it tends to become a commitment the moment it enters a deck.
It runs entirely in your browser. Nothing you type is sent to a server, no account is required, and there is no usage limit, because there is no cost per run to control.
The error bar, worked through, on numbers a seed-stage company would recognise
Take forty paying customers and two cancellations in a month. The point estimate is five percent monthly churn, and nobody would blink at it in a board update. Put a standard 95 percent interval around a proportion that size, and the plausible range for the underlying rate runs from about 1.4 percent to about 16.5 percent. That is not a technicality, it is the honest width of what forty customers can tell you about a rate.
Now push both ends through the formula. At 190 dollars a month and 72 percent margin, an account contributes about 137 dollars of monthly gross profit. At 1.4 percent churn the implied lifetime is 72 months and lifetime value is close to 9,900 dollars; at 16.5 percent it is six months and roughly 830 dollars. Against a 900 dollar acquisition cost that is a ratio of 11 to 1 at one end and 0.92 to 1 at the other, from one dataset that nobody measured incorrectly. The point estimate in the middle is 3.04 to 1, which looks reassuringly like the conventional target and is the least informative thing in the paragraph.
Notice what the two ends recommend. One says spend everything you can raise; the other says stop selling until the product changes. When a metric's confidence interval contains both of the available decisions, the metric has not yet earned the right to make either of them, and quoting the midpoint does not resolve that, it conceals it. Sample size is the fix, and there is no shortcut: the interval narrows roughly with the square root of the number of customers observed, so quadrupling the base halves the width.
Two more distortions specific to the first cohorts, both pointing the same way
The first is censoring. Your oldest customers are seven months old, so no observation of month twenty exists, and the accounts that will cancel at month fourteen have not had the opportunity to yet. Any lifetime longer than your oldest cohort is being asserted rather than measured, and the assertion sits precisely where the majority of the claimed value lives, since a low early churn rate front-loads almost nothing and back-loads almost everything.
The second is who those customers are. The first cohort of a young company is rarely representative of the market it intends to sell to. It contains design partners, people from the founders' networks, discounted or free-converted accounts, and users with unusual tolerance for a rough product because they wanted the problem solved badly enough to put up with it. Those customers churn differently in both directions, sometimes staying through things a normal buyer would not, sometimes leaving abruptly when a favour has been repaid, and their pricing distorts the revenue term as well as the retention one.
Acquisition cost has a matching problem at this stage. Founder-led selling is the dominant cost and rarely appears in the number, which makes early acquisition look almost free; and because it does not scale, the figure is not a forecast of what customer forty-one will cost. Either you count founder time at a loaded rate, in which case the ratio looks worse than the one everyone else quotes, or you exclude it and label the figure as excluding founder time. What you should not do is exclude it silently and then plan a hiring round on the result, because the first sales hire will not close at the founder's cost or the founder's rate.
What to put in the board update instead, and how to say it
Replace the single ratio with three things you have actually observed. Retained gross profit per acquired customer at the age of your oldest meaningful cohort, which the capped-horizon rows in this calculator give you directly: pick the horizon closest to your real cohort age and quote that number with the age attached. Whether any cohort has yet returned its acquisition cost in cash, which is a binary fact rather than a model. And the direction of travel between cohorts, meaning whether customers acquired in month eight are retaining better at day 60 than customers acquired in month three.
That third one is the most useful signal a young company has, and it is the one the ratio destroys by averaging. Improvement between cohorts is evidence the product and the targeting are getting better; a flat sequence across six cohorts is evidence they are not, and no aggregate lifetime value figure will surface either pattern. It is also robust to small numbers in a way a rate is not, because you are reading a shape rather than a level.
For the investor conversation, the framing that works is the one that shows you understand the limits of your own data. A founder who says the observed twelve-month retained gross profit is 1,258 dollars against a 900 dollar acquisition cost, that the oldest cohort is seven months old, and that everything beyond that is extrapolation, is describing a company under control. A founder who quotes 3.04 to 1 without qualification is one diligence question away from having to explain why the churn rate came from forty customers. The second conversation is much worse than the first, and it happens after the term sheet has been drafted rather than before.
Numbers worth knowing
| Metric | Typical | What it means |
|---|---|---|
| Interval on two cancellations from forty customers | roughly 1.4% to 16.5% monthly churn | A standard 95 percent interval on that proportion. It corresponds to a customer lifetime between six and 72 months, which is the whole span of plausible answers rather than a refinement of one. |
| Resulting range for the ratio | about 0.9:1 to 11:1 | Same customers, same revenue, same margin, same acquisition cost. When the interval contains both available decisions, the midpoint is not a compromise, it is a guess presented as a finding. |
| How fast the interval narrows | with the square root of customer count | Quadrupling the observed base roughly halves the width. There is no analytical shortcut that substitutes for having sold to more people for longer. |
| What to quote instead | retained gross profit at your oldest cohort age | A measurement with a date attached, taken from the capped-horizon rows, beats an infinite sum every time in a diligence conversation. |
Mistakes that quietly cost you results
- Projecting a rate observed over one quarter across a lifetime nobody has lived
- Use the capped horizon closest to your oldest cohort age and label it with that age. If the oldest cohort is seven months old, a 36-month figure is a forecast, and it should be presented in the tone forecasts deserve.
- Quoting the ratio to two decimal places on forty customers
- The precision implies a confidence the sample cannot support. Give the range the data actually permits, or give the observed retained figure instead, which is a number rather than an estimate.
- Leaving founder selling time out of acquisition cost without saying so
- It is usually the largest component early and it does not survive the founder stepping back. Either price it at a loaded rate or state the exclusion clearly, because the first sales hire will not reproduce either the cost or the win rate.
- Building the base cohort out of design partners and friendly accounts
- Discounted, hand-held and network-sourced customers retain unlike the market you intend to sell to, in both directions. Report them separately from customers acquired through a repeatable channel, even when that leaves an uncomfortably small sample.
What does the output look like?
This is the exact output the tool produces from the example inputs. It is generated by the same code that runs when you click the button, so what you see here is what you get.
Frequently asked questions
How many customers and how much history do we need before this ratio means anything?
Enough that the interval around your churn rate is narrower than the decision you want to make, and a cohort at least as old as the horizon you intend to quote. Forty customers is nowhere near it: two cancellations put the rate anywhere between roughly one and seventeen percent. The width shrinks with the square root of the sample, so meaningful improvement takes multiples more customers rather than a few more weeks of the ones you have.
Investors are asking for LTV to CAC and we are eight months old. What do we send them?
Send the capped figure with its age attached, plus the cohort-over-cohort direction. Something like: twelve-month retained gross profit per acquired customer of 1,258 dollars against 900 dollars of acquisition cost, oldest cohort seven months, everything past that horizon modelled not observed. Investors who have seen the failure mode read that as competence. The ones who wanted a single confident number will ask you where it came from within one diligence call anyway.
Cohort one churned badly and cohort four looks excellent. Which do we use?
Neither on its own. Report the sequence, because the movement between cohorts is the real signal and it is the thing an aggregate destroys. Rising retention across successive cohorts is evidence the product and the targeting are improving, which is more valuable information than any level. A single flattering recent cohort is also the most likely to be too young to have reached the point where its predecessors failed.
Does the churn rate get better or worse as we grow, and which way does that bias the early number?
Usually it improves, because early cohorts absorb the roughest version of the product, the least focused targeting and the least practised onboarding. That biases the early ratio downward, which sounds like good news until you notice the countervailing distortion: early cohorts also contain hand-held design partners and network-sourced customers who stay for reasons that will not repeat at scale. The two effects work in opposite directions and neither is measurable yet.
What should we track instead while the ratio is unusable?
Three things. Whether contribution per account is positive at all, which a surprising number of early companies fail while their modelled ratio looks healthy. Whether any acquisition cohort has returned its cost in cash, which is a fact rather than a model. And the retention curve by cohort, read as a shape rather than a rate, because shapes are legible at small samples in a way percentages are not.
Related free tools
- LTV to CAC Ratio Calculator Analytics
- LTV to CAC Ratio Calculator for ecommerce Analytics
- LTV to CAC Ratio Calculator for b2b sales led Analytics
- A/B Test Significance Calculator Analytics
- A/B Test Sample Size Calculator Analytics
- UTM Campaign Tagger Analytics
- Abandoned Cart Sequence Writer Ecommerce
- Cold Calling Scriptwriter Outbound
Some links on this site are affiliate links, which means Hacking Demand may earn a commission if you buy through them at no extra cost to you. This does not influence which tools are listed. The tools on this page are free and have no affiliate relationship of any kind.