Customer lifetime value, written as CLV or LTV, is the total revenue or profit a customer generates over their entire relationship with you, from the first invoice to the last, rather than what they pay in the month they sign.

It's the figure behind most growth decisions that involve money not yet received: how much you can spend to win a customer, which segments deserve your sales team's time, and whether each new cohort of customers is shaping up better than the one before. In any business with renewals or repeat orders, the first deal is a small part of what a customer is worth.

The catch is that lifetime value is a forecast, and the method you pick changes the answer a lot. A one-line formula, a discounted version and a count of what past customers actually paid give the same company three different figures, and a webshop can't use the same method as a software company at all.

TL;DR

• The simple formula is annual revenue per account × gross margin ÷ annual churn: €7,200 × 70% ÷ 15% = €33,600 for our Zwolle example. Billing tools often skip the margin, and Stripe assumes a 60-month lifetime when nobody churned in the period.

• Discounting future years brings it down. At 15% churn and a 10% discount rate, the same customer is worth €5,040 × 1.10 ÷ 0.25 = €22,176, a third less than the undiscounted figure.

• The cohort method adds up what one group of customers really paid. Zwolle's 2021 signings produced €21,357 of gross profit per original account in five years, about 14% more than the formula expected for those years, because the accounts that stayed churned less and grew.

• Retention rises with cohort age because high-churn customers leave early, a sorting effect Fader and Hardie describe. Probability models (shifted-beta-geometric for subscriptions, BG/NBD plus gamma-gamma for repeat purchases) project it forward, and pay off at volume.

• Subscription CLV starts from an observed cancellation and is measured per account, with expansion included. E-commerce CLV has to infer whether a silent customer has left, and builds value from order frequency, order value and margin after returns and shipping.

• Use CLV as the ceiling on acquisition spend and to rank segments. In Gupta, Lehmann and Stuart's study of five listed firms, a 1% gain in retention raised customer value by 3 to 7%, against about 1% for margin and 0.02 to 0.3% for acquisition cost.

Three ways to put a number on customer lifetime value

Every CLV calculation asks how much a customer will have been worth once they're gone, and there are three broad ways to answer it.

The historical method looks backwards. Take the customers who signed in the same period, add up everything they paid, and divide by the number who started. It contains no guesses, but it stops at today.

The formula method looks forwards with a few averages: what a customer pays, the margin you keep and how fast customers leave. It's what most finance sheets and billing dashboards show, and it treats every customer and every year as identical.

The predictive method fits a probability model to each customer's history and projects how long they'll stay and what they'll spend. Most B2B companies end up using the first two together, the cohort numbers to check the formula and the formula to fill in the years the cohorts haven't reached.

The simple CLV formula, on a Zwolle scheduling company

Take an illustrative software company in Zwolle that sells maintenance-scheduling software to installation firms. Its average account pays €600 a month, billed annually, so €7,200 a year. Its gross margin, after hosting, support and onboarding, is 70%. Each year it loses 15% of its recurring revenue to cancellations and downgrades, which is the same as saying its gross revenue retention is 85%.

The formula has three inputs:

CLV = annual revenue per account × gross margin ÷ annual churn rate

Each account earns €7,200 × 70% = €5,040 of gross profit a year. Dividing by churn turns that into a lifetime: 1 ÷ 0.15 means the average account stays about 6.7 years. So CLV is €5,040 ÷ 0.15 = €33,600.

Drop the margin and you get €7,200 ÷ 0.15 = €48,000, which is the revenue version. That's the one most billing tools report. Stripe's support article on subscriber lifetime value, for example, defines it as average revenue per subscriber divided by subscriber churn rate, and when no subscriber churned in the period it assumes a 60-month lifetime. Fine for a dashboard, but paste it into a plan and you're planning on a number the tool made up.

Which churn you divide by matters too. A logo count weighs a €50 account like a €5,000 one, so if sizes vary, use revenue churn from gross revenue retention, and keep the period consistent: annual revenue with annual churn, monthly with monthly. The customer churn rate entry covers how monthly and annual churn convert, which isn't by multiplying by twelve.

Discounting the years that haven't happened yet

The simple formula treats €5,040 in year six as worth the same as €5,040 today. It isn't, because money now can be reinvested and a forecast six years out is shakier. Gupta, Lehmann and Stuart, in their 2004 paper on valuing a firm's customer base, define a customer's value as the expected sum of discounted future earnings, given a retention rate and a margin.

With the first year paid at signing and the same margin every year after, that sum comes out as:

CLV = annual gross profit × (1 + i) ÷ (1 + i - r)

where r is the retention rate and i the discount rate. The paper notes that finance texts suggest an annual discount rate between 8% and 16%, and the authors used 12%. Run Zwolle at 10%: €5,040 × 1.10 ÷ (1.10 - 0.85) = €5,040 × 4.4 = €22,176. Set i to zero and the same formula gives back €33,600, so the gap of €11,424 is entirely the cost of waiting.

Discounting matters most when lifetimes are long or when you're valuing a customer base for investors. For internal comparisons, the undiscounted figure is fine as long as you always use the same one.

What a 2021 signing cohort really paid

The formula assumes the company loses 15% of its revenue every year from every group of customers, for ever. Real customer groups don't. Here are the 40 accounts Zwolle signed in 2021, all starting at the €7,200 average, followed through five full years:

Divide the cumulative €1,220,400 by the 40 accounts that started and each original account has paid €30,510. At 70% margin that's €21,357 of gross profit per account, and nothing in it is an estimate.

Now compare it with what the formula expected for the same five years. Year one counts in full, then 85% of it, then 85% of that, and so on: €5,040 × (1 + 0.85 + 0.7225 + 0.614125 + 0.52200625) = €5,040 × 3.70863125 = €18,691.50. The cohort beat the formula by €2,665.50, about 14%, and it isn't finished. The 29 accounts still on the books pay €230,400 a year between them, or about €7,945 each, which is more than they paid at signing.

A single churn figure hides two things here. Losses shrank fast after year one, and the accounts that stayed added seats and modules, which is expansion revenue doing what the simple formula leaves out. The net revenue retention entry covers how to measure that growth on the base without new customers muddying it.

The falling churn has a known explanation. Peter Fader and Bruce Hardie show in their paper on projecting customer retention that retention rates usually rise with the age of a cohort even when no individual customer becomes more loyal. Customers who were always likely to leave do so early, and the ones left are those who were always likely to stay. They call it a sorting effect. So a blended churn rate overstates how fast your oldest customers will leave and understates first-year losses.

Predictive CLV models, and who actually needs one

Fader and Hardie's answer to that sorting problem is the shifted-beta-geometric (sBG) model. It assumes each customer has their own fixed chance of cancelling at each renewal, that those chances vary across customers, and it fits the spread from a few years of cohort retention figures. The paper shows it working in a plain Excel spreadsheet, and it's built to project the tail of a cohort like Zwolle's 2021 group, which one blended churn rate can't do.

For businesses without contracts, the standard tools come from the same researchers. The BG/NBD model (Fader, Hardie and Lee, 2005) estimates how many more purchases each customer will make, and the chance they're still active at all, from just three facts: how many times they've bought, when they last bought, and how long they've been a customer. The authors present it as a far easier alternative to the older Pareto/NBD model that gives very similar results. A companion gamma-gamma model of monetary value then predicts likely spend per transaction for each customer, and the two multiply into a CLV per person.

Do you need any of this? A webshop with tens of thousands of repeat buyers does, because open-source libraries make the models cheap to run and the per-customer output decides who gets a win-back offer. For a B2B company with a few hundred accounts, we'd usually stop at a cohort table and the discounted formula, with retention read year by year from the cohort table rather than one blended average. A model fitted to four cohorts of 40 accounts mostly measures noise.

Subscription CLV is not e-commerce CLV

Fader and Hardie draw the line that matters here. In a contractual setting the customer's departure is observed: they cancel, or they don't renew. In a noncontractual one, the business never sees the customer leave. They simply stop ordering, and you can't tell a lost customer from one in a long gap between purchases.

• The unit. B2B subscription CLV is measured per account, a company with several users and often several contracts. E-commerce CLV is per person, or per email address in practice.

• Where lifetime comes from. Subscriptions have a churn rate you can count. A webshop has to estimate the probability that each customer is still active, which is exactly the job BG/NBD was built for.

• Growth inside the customer. B2B accounts expand through seats, tiers and second products, so a good CLV includes expansion, as Zwolle's cohort did. Repeat purchase value grows through frequency and basket size instead, which the spend model handles separately.

• Margin. In software it's mostly hosting, support and onboarding. In e-commerce, margin per order has to come after product cost, shipping, payment fees, discounts and returns, and in categories where many orders come back, returns can erase most of the margin.

• Skew. A handful of very large accounts makes a B2B average misleading, so report CLV by segment or contract size. A webshop's skew sits in a long tail of one-time buyers.

The scale differs too. A Breda webshop selling coffee beans might keep €13.50 from a €45 order after costs and returns, so its CLV depends almost entirely on how many repeat orders a first-time buyer goes on to place.

CLV sets the ceiling on acquisition spend

A customer's lifetime value is the most you could rationally spend to win them, and you should spend far less, which is why CLV is read next to customer acquisition cost, and why the LTV to CAC ratio exists. That entry covers the arithmetic, the common rule of thumb that lifetime value should be at least 3 times acquisition cost, and why the ratio needs reading next to payback.

What the ratio inherits from this entry is the method. Feed it the €48,000 revenue figure and Zwolle seems able to afford more than twice what the €22,176 discounted figure supports, so pick one definition and keep it.

CLV also sets how much effort a deal deserves. At €8,000 of lifetime value, a sales cycle with five meetings and a tailored proposal doesn't pay. At €80,000, it does, and that logic decides how your revenue engine splits work between self-serve, inside sales and field sales.

Raising lifetime value starts with who you sign

Gupta, Lehmann and Stuart measured how sensitive customer value is to each input across five listed companies (Amazon, Ameritrade, Capital One, eBay and E*Trade). A 1% improvement in retention raised customer value by 3 to 7%, while a 1% improvement in margin raised it by about 1% and a 1% cut in acquisition cost by only 0.02 to 0.3%. The division in the formula explains why. Small changes in a denominator make large changes in the result.

The cheapest retention is decided at the sale. Split CLV by segment and some customer types stay for years and expand while others leave at the first budget squeeze. Feed that back into your ideal customer profile, so the companies with long lifetimes are the ones you go looking for, and use automated prospecting that screens accounts for fit before a rep spends time on them so the short-lifetime segments never fill the pipeline in the first place.

After the sale, Zwolle's cohort points to the levers. Most revenue is lost in the first year, so that's where onboarding effort goes. Watch accounts with a customer health score that flags falling usage or a departed champion before the renewal, and treat downgrades as losses rather than as saves. Then make room for expansion in the accounts that are getting value, since the customers who stay and grow are also the ones who refer the next customer, which is what turns a GTM flywheel.

Recap

Customer lifetime value is a forecast with a definition attached, and the definition does most of the work. Revenue or margin, discounted or not, one blended churn rate or a churn curve from real cohorts: each choice moved Zwolle's figure by thousands of euros, from €48,000 down to €22,176, before a single customer did anything different.

The sensible way to work with it is to let the methods check each other: the cohort table for what customers actually paid, the formula for the years the cohorts haven't reached, and a probability model once you have the volume to fit one. Whichever version you settle on belongs next to acquisition cost, split by segment, with the method stated.

The figure is also a verdict on targeting. Customers who fit stay longer and grow, so lifetime value is decided as much by who sales and marketing go after as by anything customer success does later, which is why it belongs at the centre of how a GTM setup connects targeting, selling and retention.

If your CLV currently lives in a spreadsheet nobody can rebuild, we can set up the signing cohorts, revenue history and segment fields in your CRM so the cohort table refreshes on its own, and leave your team with a setup they can run and adjust without us.

Frequently asked questions