What is LTV?

A store with a A$120 average order, a 45% gross margin, 2.4 orders a year and an average relationship of 2.5 years: 120 × 0.45 × 2.4 × 2.5 = A$324 of contribution per customer. Stated as revenue the same customer looks like A$720, which is the figure most often quoted and the one that cannot be spent. The A$324 is the ceiling every acquisition decision has to fit under.

Lifetime value asks how much a customer is worth across the whole relationship rather than at the first purchase. It exists because acquisition decisions made on the first order alone are systematically wrong in businesses where customers come back, and because the advertiser willing to pay the most for a customer usually wins the customer.

The reasoning is simple. If a customer produces A$300 of contribution over three years, the business can pay considerably more than a first-order profit to acquire one and still be ahead. A competitor who only counts the first order will bid lower, lose the auction, and be right about the arithmetic in front of them and wrong about the business.

That is the whole of the strategic case for the metric, and it is a strong one. The difficulty is that lifetime value is a forecast, and forecasts are where optimism enters an otherwise arithmetic discipline.

The single largest error in this metric is stating it as revenue.

A customer who spends A$720 across three years has not handed the business A$720 to spend on advertising. Cost of goods, shipping, payment processing, returns and the support they consume all come out first. What is left — the contribution — is what can pay for acquisition, overhead and profit.

The gap is not marginal. At a 45% gross margin, a revenue-based lifetime value is more than twice the contribution-based one, and an acquisition ceiling set from it will be more than twice as high as the business can carry. Accounts run this way look healthy for exactly as long as it takes for the cash to run out.

A workable formula, stated in contribution:

LTV = average order value × gross margin × purchases per year × years retained

Every one of those four terms is measurable, which is the point. A figure assembled from four observable numbers can be checked. A figure produced by a model nobody can decompose cannot.

Customer value is skewed in almost every business. A minority of customers produce a large share of the contribution, and the rest cluster well below the mean. A single company-wide figure therefore describes almost none of the actual customers, and sits above the median by a margin that varies with how skewed the base is.

Averaging across cohorts compounds it. Customers acquired with a 30% discount code behave differently from customers acquired at full price: they order less often, return more, and churn faster. Customers acquired from branded search were frequently already intending to buy. Customers acquired from a broad prospecting audience were not. Blending them produces a number that supports a bid nobody should make on either group.

Segmenting is what turns this metric from a board slide into an instruction. Lifetime value by acquisition channel, by first product bought, by discount level and by cohort month are all computable from data most businesses already hold, and each of them can carry its own acquisition ceiling.

The word "lifetime" invites a projection longer than the evidence.

A company with eighteen months of trading history cannot observe a five-year customer relationship. It can model one, and the model will rest on a retention curve extrapolated well past the data. Small changes in assumed annual retention compound into very large differences in the fifth year, which is precisely the part of the forecast that is doing the most work in the number and resting on the least evidence.

The practical answer used by most disciplined teams is to stop using lifetimes and start using windows. Twelve-month contribution and twenty-four-month contribution are less impressive figures and far better ones: they are observable, they can be compared cohort to cohort, and a forecast made this year can be checked next year.

A discounted calculation, which weights near-term contribution more heavily than distant contribution, is the formal version of the same instinct. Cash next month is worth more than cash in three years, and a business funding its own growth feels that acutely.

Lifetime value does nothing on its own. It enters decisions in two places.

The first is the ratio against acquisition cost. A customer worth three times what they cost to acquire is a common rule of thumb for a healthy business, with the surplus covering fixed costs and profit. A ratio close to one means the business is buying customers at cost and living on hope; a very high ratio often means the account is under-spending and leaving growth on the table.

The second is payback period, and it is the one that binds in practice. A customer worth A$900 over three years who costs A$400 to acquire looks excellent at a ratio of better than two to one, and still requires the business to fund that A$400 for however many months it takes to come back. Growth is limited by cash, not by ratios, and a long payback with a healthy ratio is the shape of most companies that grow themselves into trouble.

Three of the four terms in the formula are not advertising settings.

Average order value responds to merchandising, bundling, shipping thresholds and pricing. Purchase frequency responds to lifecycle marketing, replenishment timing and whether the product gives anyone a reason to come back. Retention responds to the product and the service around it more than to anything a campaign can do.

Gross margin responds to procurement, freight and returns, and a percentage point here is worth more than most bid changes, because it applies to every customer the business will ever have.

Acquisition affects the fourth term indirectly, by determining which customers arrive. Channels, offers and audiences that bring in customers who come back raise lifetime value without anyone touching the product, and channels that bring in one-off discount seekers lower it. This is the one lever the advertising account genuinely owns, and it is worth far more than it usually gets credit for.

• State it as gross margin, never revenue. Revenue cannot pay for advertising, and a lifetime value quoted at full price will sanction an acquisition cost the business cannot carry.
• Compute it by cohort rather than as one company-wide figure. Customers acquired by discount behave differently from customers acquired at full price, and the average across them describes neither.
• Cap the horizon at something the business can see. A five-year projection from eight months of data is an assumption wearing a number, and one year of observed contribution is worth more than five years of extrapolated contribution.
• Subtract the cost of serving, not just the cost of goods. Shipping, returns, payment fees and support are all real and all scale with the customer.
• Read it beside payback period. A high lifetime value with a two-year payback is a cash-flow problem no matter how healthy the ratio looks.