What is a customer actually worth?

Almost every acquisition decision rests on this figure and almost nobody computes it. What gets used instead is the revenue on the first job, because it is available and feels concrete. In a business with any repeat purchase or referral, that understates a customer substantially. In one with high servicing costs or frequent disputes, it overstates them. Both errors send the marketing budget in the wrong direction with apparent precision.

The version worth computing is not complicated. Take the margin — not revenue — on everything a customer buys across the relationship, subtract what it costs to serve them beyond the direct cost of the work, and you have a figure you can hold an acquisition cost against. Margin rather than revenue matters more than any refinement: a customer generating substantial revenue at low margin can be worth less than one generating half as much profitably, and a business optimising for revenue will happily acquire more of the wrong one.

Repeat purchase is the term that changes the answer most in service businesses and is the hardest to establish, because it requires knowing how long customers actually stay rather than how long you assume. That is answerable from records most businesses have: how many customers who bought two years ago bought again, and how many times. It will usually be lower than the impression and higher than the first-job figure, and being between the two is what makes it useful.

Referral value belongs in the calculation and is routinely omitted because it is hard to attribute. It does not need precision to matter — if a meaningful share of new customers say they were recommended, then the customers doing the recommending are worth more than their own purchases, and that shifts what you can afford to spend acquiring the kind of customer who recommends. Recording how each new customer heard of you, which is one field, is enough to see the shape.

Segmentation is where this becomes actionable rather than interesting. A single blended figure across all customers describes nobody and hides the finding, which is nearly always that customer worth varies enormously by type, source and service. The segment that repeats and refers may be worth several times the one that buys once and leaves, and if both are acquired through the same channel at the same cost, the business is overpaying for one and underpaying for the other simultaneously.

The practical use is a ceiling rather than a target. Knowing what a customer is worth tells you the most you can rationally spend to acquire one, which is the number that resolves most marketing arguments — whether a channel is too expensive, whether a lead price is defensible, whether a retainer pays. Without it, those arguments are conducted entirely on intuition, and the loudest intuition wins.