How do I calculate customer lifetime value?

Lifetime value implies you know how long a customer will keep buying. You do not, and a store that is three years old has no way to find out.

Use a fixed window instead. Twelve months is the usual choice because it covers a full seasonal cycle and matches how most businesses plan. Some categories work better at six months, others at twenty four.

The method is the same either way. Pick the customers who placed their first order in a given month, far enough in the past for the window to have closed. Add up all revenue from that group since, including their repeat orders. Take out the variable costs: product cost, fees, shipping, packing, refunds. Divide by the number of customers in the group.

That is a real number. It contains no assumptions and no decay curve, and you can show anyone how it was produced.

Use profit rather than revenue. A revenue-based figure looks impressive and cannot be compared with acquisition cost, which is the only comparison the number exists for.

Once you have one window measured, repeat it for several cohorts. The trend across cohorts is where the information lives, because it shows whether the customers you are acquiring now are better or worse than the ones from last year.

Underneath the arithmetic sit four inputs, and it helps to see them separately because each has a different owner.

Average order value. What each order is worth. Moved by bundles, thresholds and add-ons, and the fastest of the four to change.

Gross margin. What is left after variable costs. Moved by pricing, supplier terms, shipping arrangements and return rates, and usually the slowest to change but the largest in effect.

Purchase frequency. How often a customer buys inside the window. Moved by replenishment timing, subscriptions, range extension and simply reminding people at the right moment.

Retention. What share of customers are still buying at all as the window progresses. Moved by product quality, delivery experience and whether there is any reason to come back.

Multiply the first three and apply the fourth and you have the shape of the calculation. More usefully, look at which of the four is weakest in your store, because that is where the value is being lost.

Most stores discover the weak input is frequency rather than order value. Customers are happy and simply forget, or they were never given a reason to return at the moment they needed one.

A single store-wide figure is a headline. The segmented versions are what change what you do.

By acquisition channel. Customers from different sources behave differently long after the first order. A channel with a higher acquisition cost can be the better investment if its customers return, and the cheap channel can be the expensive one once the year is counted.

By first product. The item someone starts with predicts a great deal. Some products introduce people to the range and lead to a second order. Others are bought once and never again, often as a gift. Knowing which is which changes what you advertise and what you put in front of new visitors.

By discount at first order. Customers acquired with a heavy discount often show a lower value across the window, because the offer attracted the offer rather than the brand. Comparing full price and discounted cohorts is one of the more uncomfortable and useful things you can measure.

By geography and by device where volumes allow, since delivery cost and experience differ enough to matter.

Do the segmentation on cohorts of reasonable size. Slicing a small customer base into ten segments produces ten noisy numbers and a lot of confident wrong conclusions.

The value of the figure is in the decisions it unlocks, and there are four worth naming.

It sets what you can pay to acquire a customer. With a proven twelve month profit figure and a payback window your cash can survive, the maximum acquisition cost stops being a guess.

It prices retention work. If a post purchase programme moves customers from one and a half orders a year to two, the increase in value per customer multiplied by your customer count is the size of that prize. Usually it dwarfs what people expect, which is why retention deserves budget rather than leftovers.

It informs which products to promote. Promote the products that start relationships, not only the ones with the best first order margin.

And it changes how you talk about a bad month. A month with a higher acquisition cost is not automatically a bad month if the customers acquired look better on the early indicators.

One caution. Do not use the number to justify spending you cannot fund. A twelve month value is money arriving over twelve months, while the acquisition cost is paid today. Many stores that failed had excellent lifetime value on paper and ran out of cash in month four.