Imagine a dashboard showing a large number labelled “customer lifetime value.” What exactly was counted? Is the number sales revenue, money left after delivering the service, or money left after acquiring the customer too? How many future years were assumed? Did anybody check whether those years are likely? A number without these answers can sound precise while being a weak guide to a decision.
This chapter explains customer lifetime value (CLV) from the beginning. You will calculate a simple contribution-based estimate, compare it with customer acquisition cost (CAC), then improve the estimate using retention and a small discounting example. Every number below is invented solely for arithmetic practice; it is not taken from slides, a company or a customer story. You can reproduce the calculation with a calculator, and all diagrams have complete explanations in words.
At the end, you will have a blank worksheet that names each input, its unit, its source and its uncertainty. The goal is a useful estimate you can challenge, not a number that pretends to know the future.
Figure 1
Figure 1 in words: SOSTAC® asks whether a plan is sustainable and how its assumptions will be reviewed. RACE's Reach stage can involve acquisition cost, while Engage concerns continuing value and service. Control compares forecasts with actual outcomes. CLV connects these questions; it is not a financial value assigned to a person's dignity (Smith, no date; Chaffey, no date).
The simple meaning: an estimate over time
CLV estimates the financial contribution associated with a customer relationship across a defined future period. There are several methods and conventions. One organisation may display lifetime revenue before any costs; another may estimate lifetime contribution after relevant delivery costs; another may subtract acquisition cost to show a net amount. These measures answer different questions. Always read the definition next to the number.
The OpenStax marketing textbook calls CLV a forward-looking estimate, not a guarantee, and shows a simplified calculation using annual profit, years retained and acquisition cost (Gomez Albrecht, Green and Hoffman, 2023). Academic research by Gupta, Lehmann and Stuart (2004) describes customer value as the expected sum of discounted future earnings. Its empirical results concern particular firms and assumptions; they do not supply a universal CLV target for everyone. The American Marketing Association's current calculator guidance also distinguishes revenue lifetime value from contribution lifetime value and allows either an average lifetime or a retention curve (AMA, no date).
“Lifetime” does not require predicting the date when a particular person leaves. It means the period included in a model. State the horizon: three years, for example, is a three-year model. A relationship might continue longer or end sooner. The number is an average or expectation for a defined group, not a prophecy about an individual.
Figure 2
Figure 2 in words: Revenue CLV counts expected sales before delivery and service costs. Contribution CLV subtracts defined variable costs but is before acquisition cost. Net-after-CAC value subtracts acquisition cost once. None of these measures describes a person's human worth. A report must state which of the first three it uses.
A short refresher on CRM and the plan
Chapter 24 explained customer relationship management (CRM) as coordinating helpful service across contacts. A CRM record may help measure repeat purchases and support needs, but a record is not the relationship. CLV uses relevant financial and retention information to estimate a possible outcome; it does not decide how politely to answer a person. Link the previous chapter internally only after its canonical public URL is verified.
SOSTAC® stands for situation, objectives, strategy, tactics, action and control. RACE stands for Reach, Act, Convert and Engage. In this chapter, these are orientation tools. CLV can help test whether an acquisition plan is financially plausible and whether continued service makes sense, while Control asks whether the assumptions were borne out. The frameworks do not dictate a single correct formula (Smith, no date; Chaffey, no date).
1. Begin with the question and the unit
Before calculating, write down the decision. Are you trying to compare two acquisition approaches, plan how much service capacity you need, or understand whether an existing offer covers its variable costs? The answer determines the relevant inputs and horizon. A sales-revenue estimate may be useful for forecasting sales. It is inadequate by itself for deciding what acquisition spend you can afford because delivering the product or service also costs money (AMA, no date; Gomez Albrecht, Green and Hoffman, 2023).
Choose a unit of analysis: one customer, one subscription, one account, or one order is not the same denominator. If several people use one paid account, decide whether the unit is the account. Choose a period: monthly revenue and annual retention cannot be multiplied without conversion. Choose a cohort: a group whose members began in a specified period. Comparing cohorts can expose changes that an overall average conceals. Finally choose a horizon: the number of periods you will include.
Put those choices at the top of every CLV calculation. “£270” alone is nearly meaningless. “Illustrative three-year net contribution after acquisition per newly acquired account, with no churn or discounting” tells a reader what was assumed. Do not call the result observed lifetime profit if it includes future periods you have not seen.
Four words that often get mixed up
Revenue is the money received from sales before the relevant costs are subtracted. Variable cost changes with serving or supplying the unit, within the model's defined scope. Contribution is revenue less those specified variable costs; it is money available to cover acquisition and other costs. Profit usually requires a fuller accounting of costs, including fixed or shared overhead. If your calculation leaves those out, label it contribution instead of claiming complete profit.
The scope of cost matters. Delivery, payment processing, returns and support can vary by business. This article uses only two simple illustrative variable costs to teach the arithmetic. A real model should decide which costs are relevant, whether refunds and taxes are treated consistently, and whether the same period and customer unit apply to every input. A low-quality input will not become trustworthy merely because the formula is elegant.
2. Calculate contribution for one period
Start with a deliberately plain set of hypothetical inputs, all in pounds per customer per year:
Revenue: £200.
Variable delivery cost: £60.
Variable service cost: £20.
Annual contribution before acquisition cost is £200 − £60 − £20 = £120. This is £120 per customer per year, assuming the same costs and sales in each year being modelled. It is not £200 of profit. It also does not include fixed overhead, tax, capital investment or every possible cost. The labels stay attached to the number.
If you prefer to begin with several purchases per year, first calculate annual revenue as average revenue per purchase multiplied by expected purchase frequency, then subtract the relevant annual variable costs. Keep the time unit consistent. An average purchase figure is not automatically representative of every group, especially when a few very large purchases distort it.
Figure 3
Figure 3 in words: The illustrative annual revenue of £200 is reduced by £60 of variable delivery cost and £20 of variable service cost. The result is £120 annual contribution before acquisition cost. These are exercise inputs, not observed business figures.
Where acquisition cost enters
Customer acquisition cost (CAC) is the cost attributed to acquiring new customers under a defined method. A simple calculation is acquisition-related spend divided by the number of new customers acquired over the corresponding period. If the exercise assumes £900 of relevant acquisition spend and 10 acquired customers, CAC is £900 ÷ 10 = £90 per new customer. Both numbers are hypothetical. A real CAC calculation needs a consistent list of included sales and marketing costs, a suitable period and a reliable count of genuinely new customers.
For clarity, the £120 contribution above excludes CAC. We will subtract the £90 once, later. Some sources use the word profit for a period amount and then subtract CAC; if that amount already contains acquisition spending, subtracting again would double-count it. Check the definitions, not just the symbols. The AMA distinguishes contribution LTV from CAC in its calculator and uses their relationship for a payback estimate (AMA, no date).
3. Build the simplest three-year estimate
Suppose the exercise assumes that the same £120 annual contribution occurs for three years with certainty. Add three annual amounts: £120 + £120 + £120 = £360 of contribution before CAC. Subtract the one-off illustrative acquisition cost: £360 − £90 = £270 net contribution after CAC over the three-year horizon.
This is a teaching baseline. It assumes no customer leaves, annual contribution is unchanged, there is no additional acquisition cost and a pound in year three is treated like a pound today. These are strong assumptions. In particular, three years is simply the chosen horizon here; it is not an observed average lifespan. The result does not prove that any real activity is profitable after fixed costs or cash timing.
Figure 4
Figure 4 in words: £200 minus £60 minus £20 gives £120 annual contribution. Multiplying by three assumed certain years gives £360 before acquisition. Subtracting £90 CAC once gives £270 net after CAC. This baseline ignores churn, discounting and fixed costs.
Why a single “average lifespan” can mislead
Sometimes a report multiplies annual contribution by an “average lifespan.” That can be a starting shortcut if the average was calculated from appropriate data and its definition is clear. But a young business may not have observed a full lifespan. People can also join in different periods, buy at different frequencies or have different cost-to-serve patterns. Multiplying uncertain averages can hide the uncertainty rather than remove it.
In some subscription settings, a stable churn assumption can be converted into a rough average lifetime. That shortcut depends on assumptions about the timing and stability of churn, and it is often misapplied to irregular purchasing. The practical beginner approach is to use a short, explicit horizon and show retention period by period. You can expand the model when evidence improves. A professional calculator may offer both simple average-lifetime and cohort-retention modes for precisely this reason (AMA, no date).
4. Add retention without claiming certainty
Retention means continued activity under a stated definition. For an annual subscription it may mean the account remains active at renewal. For irregular purchases it may require a chosen window for what counts as returning. Churn is the opposite event under that definition, such as an account leaving or not renewing. Neither word is meaningful without the period and the rule.
Now replace certain continuation with a hypothetical 70% annual retention probability after the first year. This means the model expects all newly acquired customers to contribute in year one; 70% are expected to remain active for year two; and 70% of that remaining share for year three. The expected active shares are 100%, 70% and 49%. The last share is 0.70 × 0.70 = 0.49. We are calculating an expected amount across a group, not saying that 49% of one person exists.
Apply these shares to £120 annual contribution: £120 expected in year one, £84 in year two and £58.80 in year three. The expected three-year contribution before CAC is £120 + £84 + £58.80 = £262.80. Subtract £90 CAC once to get £172.80 expected net contribution after CAC, before discounting. The change from £270 to £172.80 comes entirely from the retention assumption in this comparison.
Figure 5
Figure 5 in words: With 70% hypothetical retention between years, year-one active share is 100%, year two 70%, and year three 49%. The corresponding expected contributions are £120, £84 and £58.80. Their £262.80 sum less £90 CAC is £172.80. The horizon is three years and discounting is still absent.
Check what the 70% came from
If the 70% is only a guess, label it a scenario. If it came from data, record the cohort, start date, observation window and active-account rule. Count the same original cohort in later periods rather than mixing new and old customers in the numerator. Consider whether people who could not have completed a full second or third year are being accidentally counted as lost. A percentage with a denominator and date is easier to review than an unexplained dashboard label.
Retention need not be constant. The first renewal may differ from the next, an offer may change and service quality may change. You could replace 70% and 70% with two observed or explicitly forecast period rates. A forecast should be updated when real cohort results arrive. Do not silently stretch an early 30-day retention figure across three annual periods.
5. Give future amounts a time value
Discounting expresses future amounts in today's value using a stated rate. A pound expected later is treated as worth less in present-value terms because timing and opportunity to use money matter. The chosen discount rate is an assumption, not a fee charged to a customer. Advanced CLV models use expected, discounted future earnings; Gupta, Lehmann and Stuart (2004) define customer value in these terms. Their paper does not make our illustrative rate of 10% the correct rate for a real business.
For an end-of-year payment, divide the expected contribution for year one by 1 + rate, for year two by (1 + rate)², and for year three by (1 + rate)³. We will assume a 10% annual discount rate solely to see the mechanics. The expected contribution values from the previous section are £120, £84 and £58.80.
Year 1: £120 ÷ 1.10 = £109.09 present value, rounded to pence.
Year 2: £84 ÷ 1.10² = £69.42 present value.
Year 3: £58.80 ÷ 1.10³ = £44.18 present value.
The unrounded amounts sum to approximately £222.69. Subtract the £90 one-off CAC assumed at the start to give £132.69 net present value after CAC. If you round each yearly line to pence and then add them, the sum is also £222.69 in this illustration. Timing matters: if CAC were paid later or contribution received at the start rather than the end of a period, the calculation would differ.
Figure 6
Figure 6 in words: With a hypothetical 10% annual discount rate, the present values of £120, £84 and £58.80 expected at the ends of years one, two and three are £109.09, £69.42 and £44.18. Their rounded sum is £222.69; subtracting £90 CAC gives £132.69. The model retains a three-year horizon and 70% annual retention.
Three figures, three questions
The £270 baseline asks what happens if all three £120 annual contributions arrive with certainty and time is ignored. The £172.80 retention model asks what happens when only expected active shares are counted, without discounting. The £132.69 discounted model additionally expresses those expected future contributions in present-value terms at a hypothetical 10% rate. The figures are not contradictory: each answers a different question under different assumptions. Write the convention beside the result.
Do not use this small model as a substitute for cash-flow planning. Acquisition spend may happen before revenue arrives; a positive estimated lifetime value does not guarantee enough cash to pay near-term bills. A business with variable purchase sizes, returns, renewals or multiple products needs a model that fits those events. Keep the simple model as an explanation of the logic, then increase complexity only when the decision and data justify it.
6. Test how fragile the estimate is
A sensitivity check changes one assumption while holding the others fixed. It tells you which inputs deserve more evidence. Retention is an instructive test because the number of active future periods changes when it changes. We return to the undiscounted, three-year model with £120 contribution in each active year and £90 CAC. Only the annual retention rate varies.
At 50% retention, expected contribution before CAC is £120 × (1 + 0.50 + 0.25) = £210; net after CAC is £120. At 70%, it is £120 × (1 + 0.70 + 0.49) = £262.80; net is £172.80. At 90%, it is £120 × (1 + 0.90 + 0.81) = £325.20; net is £235.20. These are three arithmetic scenarios, not three measured customer segments and not three recommended targets.
Figure 7
Figure 7 in words: Holding the £120 annual contribution, £90 CAC, three-year horizon and no discounting constant, assumed annual retention of 50%, 70% and 90% gives illustrative net values of £120, £172.80 and £235.20 respectively. Higher retention raises this particular model's expected value; a real retention forecast needs evidence.
You should test other inputs too. What if service cost is higher? What if revenue varies? What if acquisition cost includes a sales team that was omitted? What if the observation period is too short to estimate retention? A sensitivity range is often more honest than a single decimal-heavy forecast. When a small change in one unknown reverses the decision, collect better evidence or make a reversible decision rather than hiding the uncertainty.
Gupta, Lehmann and Stuart's (2004) paper finds a particularly strong retention effect in the firms they studied. That is evidence from a stated research setting, not a universal rule that a one percentage point retention change always produces a particular percentage change in business value. Our Figure 7 is original arithmetic with invented inputs; it is not a replication of their empirical estimates.
7. Compare CLV and CAC on the same basis
The CLV ratio compares an estimated value before acquisition cost with the acquisition cost. For the simple certain three-year baseline, contribution CLV before CAC is £360, and CAC is £90. The ratio is £360 ÷ £90 = 4:1. Net after CAC is separately £270. Do not use £270 ÷ £90 and call it the same ratio: that numerator already had the £90 subtracted. Do not compare revenue CLV with a contribution-based CAC decision and imply that delivery costs vanish.
Figure 8
Figure 8 in words: The illustrative three-year contribution CLV before acquisition is £360, CAC is £90, and their ratio is 4:1. After subtracting CAC once, net contribution is £270. The ratio and net amount have different definitions; neither includes all possible overhead or cash-timing effects.
No one ratio proves a healthy business. A higher CLV forecast may depend on several distant years of uncertain contribution. CAC may be incomplete. The money may arrive too late. A low-value account might still be owed the same minimum service and fair treatment. Compare like with like, inspect the assumptions, and look at the timing and service obligations before using the ratio to make a decision. The AMA describes CLV and payback as decision aids built on explicit revenue, margin and retention inputs, rather than as universal guarantees (AMA, no date).
Payback asks how long it takes for contribution to recover acquisition spending. In the simplest level-flow illustration, £90 ÷ £120 per year = 0.75 years, or nine months. That is only a rough arithmetic equivalent: actual billing dates, first-year churn, up-front costs and service timing may change the cash payback. A positive three-year CLV with a very slow payback can still put pressure on near-term cash. Do not confuse a forecasted payback with cash already received.
8. Work with cohorts and actual results
A cohort is a group that started under the same chosen definition in a period, such as accounts acquired in a month. To estimate retention, follow that original group through later periods. An overall active-account count can rise because new accounts joined even while an older cohort is leaving. That is why “we have more accounts now” does not measure retention for a particular group.
Decide whether a cohort includes refunds, trial accounts, repeat buyers or business accounts with several users. Define when the relationship begins. An annual account should not be judged after only two months as though it had completed a renewal opportunity. If data is limited, use a short horizon and label later periods unknown. The AMA's calculator distinguishes a simple average-lifetime input from a cohort retention curve, illustrating two different levels of detail (AMA, no date).
Review forecasts against observed results. For each cohort, compare expected and actual active shares, contribution per active account and acquisition spend under the same definitions. A difference may come from a genuinely changed behaviour, a revised cost allocation, missing data or a bad initial assumption. Update the model with a note of what changed. Do not silently overwrite a previous forecast and then claim the new version had always been accurate.
Where averages need care
An average can hide variation. Some accounts buy once, some return, and costs may differ. A large account can raise the average while most accounts experience a different pattern. You can show a range or compare well-defined groups when you have enough data to do so responsibly. Do not make highly granular predictions about individuals when a broad operational decision only needs a cohort estimate. In small groups, one unusual result can move the average considerably.
If a product is new and there are no completed multi-year cohorts, a long-horizon CLV estimate is mostly a set of assumptions. That is acceptable if stated plainly. You can use scenarios to decide what evidence to collect next: renewal rate, variable cost to serve, purchase frequency or the real acquisition cost. The model is a prompt for learning, not a certificate of profitability.
9. Use the estimate without reducing people to a price
CLV helps an organisation ask whether a marketing approach can sustain its costs and whether long-term service is being considered. It does not tell you whether one person deserves a clear answer, an accessible route, privacy or a fair correction. A predicted low contribution can be wrong, and a person may influence others in ways a simple model does not count. Even when a model is useful for budget planning, the organisation still needs a service standard for everyone.
Think in terms of three decisions. Planning: does the estimate support or challenge a proposed acquisition spend? Improving: which costly point in the experience should be investigated, and would a genuine service improvement help? Reviewing: did the forecast match what happened? Avoid treating “increase CLV” as a command to upsell relentlessly or make leaving difficult. A better outcome might come from a clearer product, fewer avoidable support problems and honest expectations.
Figure 9
Figure 9 in words: Check that data and cost definitions match; check when cash arrives; check whether service promises are realistic; check whether a decision unfairly excludes people; then compare the forecast with actual outcomes. A high estimated CLV is a question for planning, not permission to ignore other customers.
Data used for CLV should be collected and managed for a clear purpose, with relevant access and retention decisions. Chapter 24 explained those boundaries in the CRM context. A lifetime estimate does not require the name or private details of every person on a teaching worksheet. For an actual business, use the authorised data process and a definition suited to the decision.
Try it: one blank estimate and two challenges
Figure 10 is empty so you can practise without copying an invented business story or anyone's personal record. Use hypothetical numbers of your own, or use clearly labelled aggregate information you are authorised to analyse. Do not paste real customer identities into the worksheet.
Define the unit and period. Write “per account per year” or another consistent unit. Name the cohort and horizon if known.
Record revenue and variable costs. Calculate contribution per period. List what costs were excluded, especially shared overhead.
State acquisition cost. Show the spend, the acquired-unit denominator and whether CAC is already present in any other input.
Choose a simple horizon. Calculate contribution before CAC, then net after CAC. Label any years that are assumed rather than observed.
Challenge retention. Try a lower and higher retention rate, or write “not yet measured.” Do not insert a made-up rate and label it actual.
Optionally discount. State the rate, payment timing and reason; leave it out rather than hiding an unexplained choice.
Review the result. Write which real decision it could inform, what it cannot prove and which input you would verify first.
Figure 10
Figure 10 in words: The blank sheet asks for unit, period, cohort, source date and owner. Its rows cover revenue, variable costs, contribution, horizon and retention, optional discounting, CAC definition, and result with limits. It ends with a scenario to test and next review date. Do not put real personal data on it.
A useful answer may conclude “we cannot estimate year three responsibly yet.” That is a result of the exercise, not a failure. You can still estimate year one carefully, record the uncertainty, and define when year-two evidence will be available. A shorter reliable model can serve a decision better than a long confident guess.
Common mistakes to catch
Using revenue as profit. If £200 comes in and £80 of relevant variable cost goes out, the contribution in this illustration is £120, not £200. State which costs remain outside the model.
Subtracting acquisition twice. The calculation subtracts CAC once from contribution that excludes it. Check the labels of imported figures before combining them.
Mixing months and years. Convert revenue, costs, retention and discount rate to compatible time units. A monthly churn percentage is not an annual retention percentage.
Treating a fixed horizon as observed lifetime. Three years in our baseline is an assumption. In a young cohort, there may be no three-year record at all.
Mistaking correlation or an average for a guarantee. Retention, margin and acquisition costs can change together. A ratio from one group does not automatically predict another group's result.
Assuming a high ratio solves cash flow. Timing, acquisition outlay and payback matter alongside lifetime contribution.
Using a financial estimate as a service ranking for people. Give everyone the promised baseline service and a fair way to raise an issue. The model is for resource decisions, not a measure of human value.
Check your understanding
If annual revenue is £200, delivery and service together cost £80, and CAC is £90, what is the first useful step? Calculate annual contribution £120 before CAC. If you assume three certain years without discounting, contribution CLV is £360 before CAC and the net after CAC is £270. The ratio £360:£90 is 4:1. If the £120 already included CAC, you would need to revise the definitions before doing any of those later calculations.
What changes if annual retention after year one is 70%? Over the same three-year horizon, expected contribution before CAC becomes £262.80, and net after CAC becomes £172.80 before discounting. The difference does not prove any person will leave. It shows the effect of an explicit assumption on a cohort expectation. The exercise figures have no relationship to a real enterprise or source slide.
Optional videos
Start with the definition and simple arithmetic:What is CLV? Customer Lifetime Value Explained For Beginners - Formula, Example, Definition, Surfside PPC (26 October 2022). The public description works with average order value, purchase frequency and lifetime, making it a manageable first look. Its arithmetic focuses on revenue CLV, so return to the written contribution-and-cost calculation above before using it to judge acquisition spending. Its examples are not used in this article.
Go deeper with an academic lecture:Lec 23: Customer Lifetime Value, NPTEL IIT Guwahati (5 February 2026). The channel description identifies Prof. Kuldeep Baishya and the Metrics and Analytics of Marketing course; the official NPTEL course page confirms the course and instructor. This is academic teaching, not a peer-reviewed paper. It may go beyond the beginner calculation, so watching it is optional.
The shortlist below the article in the editorial packet records alternatives and point-in-time views. The two selected videos were recommended from public metadata/descriptions and official course provenance checked on 15 November 2026. They were not played or transcript-reviewed; exact runtime, captions, spoken language, UK playback and embedding remain to be verified before publication. All essential teaching is written above.
Gupta, S., Lehmann, D.R. and Stuart, J.A. (2004) 'Valuing customers', Journal of Marketing Research, 41(1), pp. 7–18. Available at: https://doi.org/10.1509/jmkr.41.1.7.25084 (Accessed: 15 November 2026).
Surfside PPC (2022) What is CLV? Customer Lifetime Value Explained For Beginners - Formula, Example, Definition. 26 October. Available at: https://www.youtube.com/watch?v=vA1YX8963ts (Accessed: 15 November 2026).
A beginner-friendly route through 26 short lessons on customers, offers, channels, content and measurement. Start anywhere, with no account or paid tools.
Learn the difference between features and benefits, write a promise customers understand, and check which kinds of online value you can honestly deliver.