Lifetime Value (LTV)

Lifetime value is the total gross profit a customer is expected to produce before they leave. It sets the ceiling on what acquiring them can rationally cost, which is the only reason to calculate it.

The usable formula is short:

LTV = ARPU × gross margin ÷ churn rate

Each term does work, and each is where the number goes wrong.

Why most LTV figures are overstated

Revenue instead of gross profit. Using revenue ignores the cost of serving the customer: support, infrastructure, the implementation nobody bills for. In B2B SaaS that is rarely small, and skipping it inflates the answer by exactly that amount.

A blended churn rate. Dividing by an average across segments that behave differently produces a number describing none of them. See churn.

The assumption of forever. Dividing by a churn rate assumes it holds indefinitely. Real retention curves flatten, which means early churn is higher and late churn lower than the average implies, and the true figure for survivors is larger while the average is smaller.

Contract length ignored. Annual customers cannot churn monthly. Applying a monthly rate to an annual base measures the renewal calendar.

The practical response is not a more elaborate model. It is to treat LTV as a comparison between segments rather than as a figure to quote, and to publish the assumptions next to it.

LTV against CAC, and why payback matters more

The convention is that LTV should be at least three times customer acquisition cost. The ratio is a sanity check and it hides something important.

Two businesses with an identical 3:1 ratio differ entirely if one recovers its acquisition cost in nine months and the other in four years. The first funds its own growth; the second needs capital to survive the gap. Payback period is the operational number and the ratio is the headline.

What design actually moves

LTV is, structurally, a churn number wearing a revenue costume. The denominator dominates, which is why the levers are mostly retention levers.

Time to value. The strongest single lever, and a design problem before it is a product one. See activation.

The second user. In collaborative products, accounts with one active user churn heavily. Getting a colleague in is worth more to LTV than most pricing work.

Accumulated state. Data, configuration and history that would have to be rebuilt elsewhere. This is why import and migration work outperforms features on this metric and never wins a prioritisation meeting.

Expansion. Seats and usage growing inside an account raise ARPU over time, which is the one place design touches the numerator.

Reading it as a curve, not a figure

The single-number form exists because it fits in a spreadsheet cell. The honest version is a curve, and it changes what you do.

Plot cumulative gross profit per cohort against months since signup. What you get is not a straight line to a ceiling: it rises steeply while early churn takes the weakest accounts, then flattens as the survivors settle.

Two things follow. The average understates the survivors and overstates the leavers, which is why a segment's LTV can be far above the blended figure and still be true. And the payback point is visible on the curve, where cumulative profit crosses acquisition cost, which is the number that decides whether growth is self-funding.

A team that plots this once rarely goes back to quoting the single figure, because the curve makes obvious what the division was hiding.

In practice

A company reports an LTV of about eleven thousand dollars and uses it to justify spending three thousand to acquire a customer.

Recalculated properly the number falls hard. Gross margin is 71%, not the 100% the revenue-based figure assumed. Splitting by segment, the mid-market cohort retains for years while the self-serve cohort loses more than half its accounts inside two quarters. Blended, they had been funding self-serve acquisition with mid-market economics.

The corrected picture is two businesses. Mid-market supports the three thousand comfortably. Self-serve does not support a third of it, and the acquisition spend against that segment had been destroying value for a year.

Nothing about the product changed. The figure had simply been averaging two things that should never have been averaged.

Where teams get it wrong

• Revenue instead of gross profit. Overstated by the whole cost of service.
• One blended number. Two businesses averaged into a description of neither.
• Quoting it to two decimal places. A projection built on an assumption of forever.
• Optimising the numerator. Price rises move it less than churn does.
• Reading the ratio without the payback period. The same ratio can be self-funding or capital-hungry.