Calculate and predict customer lifetime value

Work out what your customers are actually worth, and what separates the valuable ones from the rest.

LTV is arithmetic until you ask which customers land where and why. The engine handles both: the calculation, then the model that explains it.

Run this with your AI assistant

meelu-analytics-mcp is a free, open source MCP server that runs on your machine. One command installs it and configures your assistant.

meelu-analytics-mcp View on GitHub All open source

The problem

Why this is hard to answer today

A single blended LTV figure is an average of customers who have almost nothing in common, and averages hide exactly the thing you need to see.

One company-wide LTV tells you nothing about which customers are worth defending.
Recalculating it by hand means it is out of date by the time anyone reads it.
Without knowing which traits predict high value, acquisition spend is aimed at whoever is cheapest to reach rather than whoever is worth reaching.

The solution

What the engine does instead

Compute value per customer, group customers by it, then ask the harder question — what do the valuable ones have in common that the others do not.

Value computed per customer and saved back alongside your data.
Natural value groupings found in the data rather than imposed by round numbers you picked.
A model that ranks which traits actually predict value, tested rather than assumed.
Every step keeps its method on the record, so the number survives being questioned.

Built on Feature engineering, one of nine tool categories in the server.

How it works

Step by step

Step 1

Export customers with their revenue history

One row per customer, with what they have spent and how long they have been around. Add acquisition cost if you have it and the LTV:CAC ratio comes along free.

customers.csv

customer_id,signup_date,months_active,total_spent,monthly_spend,plan,acquisition_cost,source
CUST_001,2023-06-15,15,45000,3000,enterprise,7500,outbound
CUST_002,2024-01-20,8,12000,1500,team,4200,paid_search
CUST_003,2023-11-02,11,3300,300,starter,900,organic

Step 2
Install the MCP server
Install meelu-analytics-mcp once, then restart your assistant. The install page has the one command and the README covers the rest.

Step 3

Ask for value, then for the explanation

Two questions, because the second is the one that changes what you do next.

In your assistant

Using meelu, load ~/Downloads/customers.csv, work out the
lifetime value of each customer, and group them by value.

Then: what distinguishes the highest-value group from the
lowest? Which acquisition source produces the best ones?

Example output

What comes back

Shape of an answer — illustrative numbers, real structure.

Value groups found in the data (4 natural groups)

  Group   Customers   Median LTV   Median LTV:CAC
  A            38       $118,000           9.4:1
  B           142        $34,000           5.1:1
  C           390         $9,600           2.7:1
  D           211         $2,100           0.8:1

What separates group A from group D, ranked:
  1. seats at signup
  2. acquisition source (outbound over paid_search)
  3. months to first expansion

Trust: high — 781 rows.
Caveat: group D skews recent. Customers who have not had
time to be valuable yet are not the same as customers who
never will be.

That caveat is a real trap in LTV work — recent signups drag the average down simply by being recent, and reading it as a quality problem sends you fixing the wrong thing.
Group D having an LTV:CAC ratio below 1 is the finding worth acting on: those customers cost more to acquire than they return.

In practice

Who asks this

Growth lead

Two channels, similar cost per acquisition, and no idea which brings better customers.

Compare value by source. Cost per acquisition is only half the equation and usually the less interesting half.

Customer success

A fixed number of hours and hundreds of accounts.

Spend them on the group where the value is, and let the rest have a good automated experience.

Founder setting pricing

You suspect the entry tier is not paying for itself.

Compare value against acquisition cost by tier. If the ratio is under one, the tier is a marketing expense — which may be fine, but should be a decision.

FAQ

Common questions

What is customer lifetime value?

Customer lifetime value, or LTV, is the total profit you expect from a customer across the whole time they stay with you. It is the number that says how much you can afford to spend acquiring one. Quoted as revenue it flatters; quoted as gross profit it is useful, because the cost of serving a customer comes out of the same pot.

How do you calculate LTV?

The simplest formula is average revenue per customer per month, multiplied by gross margin, divided by monthly churn rate. Charge £100 a month at 80% margin with 4% monthly churn and LTV is £2,000. The alternative is historic LTV: add up what past customers actually paid over their whole time with you. That is less elegant and more trustworthy, and it needs enough history for a meaningful share of customers to have finished.

What is a good LTV to CAC ratio?

Three to one is the usual benchmark for a subscription business, meaning each customer returns three times what it cost to acquire them. Below one the business loses money on every sale. Far above three often means underinvestment in growth rather than excellence. Pair the ratio with payback period — how many months until acquisition cost is recovered — because a healthy ratio that takes three years to pay back still starves cash flow.

What is the difference between historic and predictive LTV?

Historic LTV measures what customers have already paid you, so it is accurate but backward-looking and says nothing about people who are still active. Predictive LTV estimates what current customers will go on to pay, using their behaviour and traits. Use historic to judge past acquisition decisions, and predictive to decide where to spend next.

How do you increase customer lifetime value?

There are only three levers: keep customers longer, charge more, or sell them more. Retention is usually the largest of the three, because LTV is inversely proportional to churn and halving churn doubles the value. After that, expansion revenue from existing customers is cheaper than raising prices, and both are cheaper than buying new customers.

Keep going

Related use cases

Customer segmentation

RFM segments and natural groupings found in your data

Churn prediction

Work out what predicts a customer leaving — and who is at risk now

Cohort retention

See which signup months held up, and find the month things changed

Try customer ltv on your own data

Install meelu-analytics-mcp once and every use case on this site is available to your assistant — it is one server with forty-five tools, not six separate products.