A value per customer, not one average
What each individual customer is likely to be worth to you from here on, based on how often they buy and what they spend — not one blended number that describes nobody.
You want to know what you can afford to pay for a customer, and the usual answer is one average that describes nobody. Upload a CSV of your orders — customer id, order date, amount — and get a value for each customer going forward, which of them have quietly stopped buying, and which groups cost more to win than they give back. It runs in your browser, so the file never leaves your machine.
Drop a transactions CSV here, or
One row per order: customer id, order date, amount. Up to 200,000 rows. Extra columns — source, plan, acquisition cost — get used, they do not get in the way.
The file is read and modelled in this browser tab. It never reaches our server, so there is nothing for us to keep.
What each individual customer is likely to be worth to you from here on, based on how often they buy and what they spend — not one blended number that describes nobody.
In a shop nobody tells you they have left. You get the odds each customer is still buying at all, so a fortnightly buyer gone quiet for six months is treated differently to a twice-a-year one.
Your customers grouped by what they are worth, each group next to what it cost to win them. A group giving back less than it costs is the finding worth acting on.
Any extra columns in your file — source, plan, first order value — ranked by how well they tell your best group apart from your worst. Useful for deciding where to spend.
If your file is too thin or your customers do not behave the way this assumes, you get what they have already spent instead, clearly labelled. A confident wrong number is the worst thing it could give you.
The per-customer numbers and the group summary as CSVs. Your file is read inside your browser tab and never reaches us. Free, no sign-up.
One row per order, with a customer id, an order date and an amount. A Shopify, WooCommerce or Stripe export already has all three. If you also know what each customer cost to win, or anything else about them, bring those columns too.
The columns are picked out of your headers and everything happens inside your browser tab. Nothing is uploaded and there is no account to make.
The first panel tells you whether you are looking at a forecast or at money already spent, and why. Then work down: the value groups, how concentrated your revenue is, and what sets the groups apart.
If you sell subscriptions you know exactly when somebody left, because they cancelled. In a shop or a marketplace nobody tells you. A customer who has not ordered for four months might be gone for good, or might buy twice a year and be perfectly happy. You cannot tell those two apart from the date alone — only from how often each of them used to buy.
That is what your order history is for. A fortnightly buyer who has gone quiet for six months gets written down hard. A twice-a-year buyer quiet for the same six months barely moves. Same silence, different conclusions — and a calculator with three boxes in it cannot make that distinction.
The formula in every article is:
LTV = average order value × purchase frequency × customer lifespan
It is not wrong. It is an average, and averages are fine when you want one number for a slide. You get it here alongside everything else, so you can see the gap between the two.
One blended LTV:CAC for a whole business is an average of customers who have nothing in common. It is the number that gets put on a board slide and the number that hides the finding. If your ratio is a healthy 3.4:1 overall and your bottom group is at 0.6:1, you are funding one group of customers with another and nobody has decided to.
So you get the ratio for each value group instead. If your file has a column for what each customer cost to win, that gets used. If it does not, type one figure into the box — less precise, still far more useful than a single company-wide number.
Read a group below 1:1 as a decision rather than a failure. An entry tier that does not pay for itself might be a deliberate marketing expense that feeds the tier above it. The point is that it should be a choice you made rather than something you found out about eighteen months later.
Customer lifetime value is the total profit you expect from a customer across the whole of their relationship with you, not just their first order. It tells you what a customer is worth, which in turn tells you what you can afford to spend winning one. People write it as LTV or CLV; the two mean the same thing.
The standard formula is average order value × purchase frequency × customer lifespan, then multiplied by gross margin to get profit rather than revenue. A customer spending £60 an order, four times a year, for three years at a 40% margin is worth £288. For subscriptions the shortcut is average revenue per account ÷ churn rate, so £50 a month at 2% monthly churn gives £2,500. Both assume the past repeats, which is why predictive models are used when the stakes are higher.
3:1 is the widely used benchmark — each customer should return about three times what it cost to acquire them. Below 1:1 you are losing money on every sale. Much above 5:1 usually means you are underspending on growth rather than running a brilliant business, and a competitor with a looser wallet will take the market. Use gross-margin LTV, not revenue, or the ratio flatters you.
LTV:CAC tells you whether a customer is profitable in the end; payback period tells you how long your cash is tied up before that happens. A business can have a healthy 4:1 ratio and still run out of money because payback takes 30 months. The usual target is under 12 months for SMB and under 18-24 months for enterprise. Look at both, because they fail in different ways.
Predictive LTV forecasts what a customer will spend in future rather than totalling what they have already spent. It fits a model to each customer's purchase history — how often they buy, how long since the last order, how much they spend — and produces an expected value along with the probability they are still an active customer. It is more useful than the historical average because it separates a quiet customer who will return from one who has silently left.
There are only three levers: get customers to spend more per order, buy more often, or stay longer. Staying longer is usually the biggest of the three and the cheapest to move, because retention compounds through the whole calculation. Practical work means better onboarding, sensible cross-sells, subscription or replenishment options, and recovering failed payments. Raising prices lifts LTV directly but only while it does not push churn up more than it lifts margin.
Profit. Revenue-based LTV counts money that never reaches you: cost of goods, payment fees, delivery, returns and the support cost of serving the account. A 20%-margin ecommerce business using revenue LTV will overstate customer worth by five times and cheerfully overspend on ads. Multiply by gross margin, and subtract ongoing service costs if they are material.
There is no absolute number, because LTV only means something next to your acquisition cost. £200 is excellent if customers cost £40 to win and terrible if they cost £300. Compare it to CAC, compare it to your own figure last year, and compare cohorts against each other. A rising LTV in recent cohorts is a better sign than any benchmark from another company.
This page does one thing with one file, in your browser, and then stops. It cannot join your orders to your support tickets, it cannot answer the question you think of after reading the table, and it cannot tell you why group D looks the way it does.
For the follow-up questions, run the same analysis inside your AI assistant with meelu-analytics-mcp. It is free and open source, it runs on your machine, and your data stays there for the same reason it does here. See the LTV use case.