Customer lifetime value is the metric that appears on the most slides and changes the fewest decisions. The reason is how it usually gets calculated: revenue per customer, averaged across everyone, once a year. What comes out is a number nobody disputes and nobody uses.
The number becomes useful the moment you read it as a ceiling. It tells you what you may spend to win a customer, how deep a first-order discount may go, and which group never returns the money it cost to acquire. This guide covers which formula fits which stage of data maturity, why contribution margin is the only workable basis, how to calculate the value by cohort instead of as an average, and which five levers move it in retail.
The example calculations below are models built from round numbers, so you can rerun them with your own figures. There are no panel numbers in this post: lifetime value comes out of your own order data, not out of other retailers' mails.
What customer lifetime value actually states
Customer lifetime value, or CLV, is the contribution margin a customer relationship produces over its whole duration, minus the costs that relationship causes, discounted back to today. Three parts of that definition get dropped regularly, and each one shifts the result:
| Element | What it means | What gets calculated instead |
|---|---|---|
| contribution margin | revenue minus cost of goods and variable cost per order | plain revenue, which leaves the margin open |
| whole duration | a fixed window, 24 or 36 months for instance | an endless relationship, which makes the value arbitrarily large |
| discounted | money later is worth less than money today | every year weighted equally, so the value looks bigger than it is |
One more term is needed before the first one decides anything: customer acquisition cost, or CAC, everything you spend to bring a new customer to a first order, divided by the customers won that way. CLV on its own is an observation. CLV against CAC is a decision.
The three formulas and when each one holds
There is no single formula, there are three, matching three stages of data maturity. Use the simplest one your data supports:
| Formula | Calculation | Requires | Good for |
|---|---|---|---|
| Simple projection | contribution margin per order times orders per year times assumed years | one year of order data | a first ceiling, rough channel decisions |
| Retention formula | annual contribution margin times r divided by (1 plus i minus r) | repeat rate r and discount rate i | ongoing budgeting, comparing cohorts |
| Cohort retrospective | actual contribution margin of one entry month after 12, 24, 36 months | two to three years of history | planning, judging acquisition routes |
The middle formula is the one most retailers do well with. It assumes a constant repeat rate and margin, and both are wrong, but wrong in a stable way: the error points the same direction in every cohort, so comparisons stay valid.
Here is a model with round numbers, to make the orders of magnitude visible. Say the average basket is 60 € and gross margin 45 per cent, so 27 €. Shipping, payment fees, packaging and the return share take 9 € off that, leaving 18 € of contribution margin per order. In the first year this customer orders 1.8 times, so roughly 32 €. Of new customers, 30 per cent buy again the following year, and the discount rate is 10 per cent:
- later years: 32 € times 0.3 divided by (1.1 minus 0.3) gives 12 €
- CLV across the whole relationship: around 45 €
- of which earned in the first year: around 32 €
That is the entire calculation, and it has an uncomfortable property: two thirds of the value arrive in year one. Anyone using CLV to argue for bigger acquisition budgets is moving money into a future this calculation only funds to a third.
Why revenue is the wrong input
The most common mistake is not the formula, it is the figure you put into it. Run the same model on revenue and it returns about 150 € instead of 45 €, three times the value. Every decision that follows is generous by that factor.
Discounts show this most sharply. Stay with the 60 € basket and 27 € gross margin: a 20 per cent reduction takes 12 € off revenue, but it does not take 20 per cent off the result. After the 9 € of variable cost, 6 € of contribution margin remain instead of 18 €. The discount costs two thirds of the margin, not a fifth.
From that follows a rule worth applying in every promotion plan: a discount is an advance on lifetime value. It is defensible when it opens a relationship that lasts three years, and expensive when it pulls forward a purchase that was coming anyway. Which discount types do what is covered in the guide to discount strategy in e-commerce, and how deep this market actually goes across the year sits in the discount depth benchmark.
What belongs in the contribution margin is a short list: cost of goods, shipping and packaging, payment fees, the return share of your category, vouchers and discounts. What does not belong: rent, salaries, software. Those are carried by the company, not by the order, and including them puts every customer group at a loss.
The average hides the decision
A CLV across all customers is a number without an addressee. The question at hand is never "are our customers valuable", it is "which route brings customers who come back". So you calculate the value by cohort, grouped by the route the first purchase came through:
| Entry route | Typical pattern | What the calculation says |
|---|---|---|
| through the range | full price, smaller basket, higher repeat rate | acquisition may cost more, the value arrives in years two and three |
| through a welcome discount | larger first basket, lower repeat rate | the first order has to carry most of the value |
| through a deadline promotion | buys on the date, returns for the next date | value hangs on the promotion calendar, not on the range |
| through referral or existing base | highest repeat rate, stable basket | best audience for programmes, worst one for discounts |
Three rules for cohort work:
- The entry month stays the bracket. A cohort is everyone who bought for the first time in one month, and it is followed over the years rather than reassembled.
- Keep discount cohorts separate. Someone who enters at 20 per cent off behaves differently from someone who pays list price. Mix them and you erase the very difference you set out to measure.
- One metric before all others: the rate to second purchase. It is visible after 60 to 90 days and says more about a cohort's lifetime value than any five-year projection.
How to cut those segments without drowning in dimensions is covered in the guide to CRM marketing in e-commerce. The stage after that, the same logic applied to individuals, is in the guide to hyper-personalization in CRM.
CLV against CAC: the number that decides
Lifetime value only becomes a decision once acquisition cost sits beside it. Three readings that belong together in retail:
| Ratio | In the model | What follows |
|---|---|---|
| CLV to CAC | 45 € against 15 €, so three to one | a healthy corridor, growth is fundable |
| first-year payback | 32 € of contribution margin against 15 € CAC | acquisition pays for itself inside year one |
| months to payback | CAC divided by contribution margin per order, here a good eight months | tells you how long growth ties up capital |
The familiar three to one rule of thumb is a reference point, not a target. The decisive line is the second one: as long as acquisition comes back inside year one, growth funds itself out of the business. Draw the ceiling from the full CLV instead and you fund it with confidence in a repeat rate you have not measured.
In practice, set two limits. The hard one is the first-year contribution margin, here 32 €: up to that point a channel may cost without the calculation leaning on the future. The soft one is the full CLV, here 45 €, and it applies only to cohorts whose repeat rate you can evidence across 24 months.
Five levers that move the value in retail
CLV has only four screws to turn, and one of them counts twice. In order of effect:
- The second purchase. Going from none to one repeat is the biggest jump in the arithmetic. Lift the repeat rate in the model from 30 to 40 per cent and CLV grows from 45 € to a good 50 €, without a single basket getting bigger. The measures that carry this are collected in the guide to customer retention in e-commerce.
- Frequency. One extra order a year does more than a bigger basket, because it raises contribution margin in every year of the relationship. The automated flows for it are described in the guide to marketing automation in e-commerce.
- Basket size. The right second item raises margin per order without adding to acquisition cost. The mechanics sit in the guide to cross-selling in e-commerce.
- Discount depth. It works downward, and faster than any other lever works upward, as the calculation above shows. Where your price should sit overall is the subject of the guide to retail pricing strategy.
- Cost per order. Return rate, packaging and the free-shipping threshold are part of the margin and therefore part of CLV. Taking one euro off your shipping burden does the same thing in the arithmetic as adding one euro to the basket.
The lever that counts twice is membership: a programme that ties benefits to an account raises frequency and repeat rate at once. What such a programme has to deliver is in the guide to retail loyalty programmes.
How often to calculate, and how long the result holds
CLV is not a monthly metric. Refresh the cohort calculation quarterly, the repeat rate and the rate to second purchase monthly, the assumptions about relationship length once a year. In between the number stands, even when it is inconvenient.
Two things invalidate a result early: a clear change in range or price level, and a new acquisition route with its own entry offer. Then a new cohort series begins, and the old one stays as a comparison instead of being overwritten.
Checklist
- Start from contribution margin per order, not revenue. Cost of goods, shipping, payment, return share, discount.
- Fix a window, 24 or 36 months, and write it down. No infinite horizon.
- Build cohorts by entry month and entry offer, not by revenue band.
- Report the rate to second purchase monthly. It is the leading indicator of CLV.
- Set two acquisition ceilings: first-year contribution margin as the hard one, full CLV as the soft one.
- Judge every welcome discount against its own cohort's lifetime value, not against the average of all customers.
- Write the assumptions next to the number. Repeat rate, discount rate, window. Without them the value cannot be compared.
Common mistakes
- Calculating on revenue instead of margin. Triples the value in the model and triples the acquisition budget with it.
- Reporting the average across all customers. Hides exactly the differences between acquisition routes that the metric exists for.
- Calculating forever. Without a window and discounting, any channel looks justified.
- Pulling fixed costs in. Rent and salaries belong in the company accounts, not in the order.
- Thinking of discount as a share of revenue. It is a share of contribution margin, and that is the smaller number.
- Projection before retrospective. A five-year extrapolation without two years of history is an opinion with a decimal place.
- Calculating once a year and forgetting it. Without a quarterly rhythm you spot a broken cohort a year late.
How Reyo helps
Lifetime value itself lives in your shop and CRM systems, where the orders are. What is not in there is the reason one cohort repeats more weakly than the one before it: if your market discounts deeper or sends more often in that quarter, your repeat rate moves with no cause visible in your own data. We do not name the shops, this is about the patterns and not the senders. And we see what was sent, not how it performed: opens, clicks and revenue from other retailers are not available to us, their offers and mechanics are.
- Competitor monitoring records the discount depths, conditions and send frequency of your market, the outside forces that act on your margin and therefore on CLV.
- Analytics turns that into series across months and quarters, matching the rhythm of your cohort work.
- Market predictions show when the next discount wave is due in the calendar, so you do not place a welcome offer in the most expensive week of the year.
- The campaign planner records which entry offer ran when, so you can attribute a cohort to the right promotion later.
The annual rhythm of promotions to read your cohorts against is in the Q4 Guide 2026. Basic starts at 69 € per month on annual billing (83 € monthly) for three competitors, with a 14 day trial, plus VAT.
Conclusion
Customer lifetime value is not an award for loyal customers, it is a ceiling on spending. Calculate it from contribution margin, inside a fixed window, discounted, and by cohort according to the route the first purchase came through. Then the number says something: acquisition may cost this much, the welcome discount may go this deep, and this cohort does not carry its budget. The most useful figure that falls out is not CLV itself but the rate to second purchase: it shows up early, it responds to what you do, and it decides whether an order turns into a relationship.
Want to know what your market is putting into inboxes in discounts and conditions right now? Start a 14 day trial, from 69 € per month, or book a demo.
Published on 18 November 2026. The example calculations are models built from round numbers, not market data.