Lifetime Value

011 min

Lifetime Value at a glance

• What it is: Expected profit, not revenue, from one customer over their whole time with you.
• Origin: Direct marketers used it for years before Berger and Nasr published general models in 1998.
• Simple formula: Profit per period divided by the churn rate gives the quick estimate.
• Decision it changes: It sets the ceiling on what to spend winning one customer.
• Watch for: One average churn rate can hide segments and understate the value of the customer base.

021 min

Why Lifetime Value matters

A founder of a subscription app has to decide how much to pay for each new signup. If they look only at the first month's revenue, they will stop spending too early on customers who stay for years. If they look only at total revenue, they will keep spending on customers who cost more to serve than they bring in. Lifetime value gives one number that avoids both mistakes.

The second mistake is common. A cross-sectional study of U.S. banks, reported by Sunil Gupta and Donald Lehmann, found that in the early 1990s only 30% of a typical bank's customers were profitable over the long run. The same excerpt points out that this means 70% of customers destroyed value. A team that counts customers or revenue would never see this. A team that estimates profit per customer over time would.

The answer sets the allowable acquisition cost, the most a business can spend to win one customer and still earn that money back. In a small company the founder owns this call. In a larger one a growth product manager and a finance lead usually share it.

032 min

What Lifetime Value actually measures

Lifetime value is a forecast of the profit one customer will bring in, from the first purchase until they stop buying. Four terms carry the idea, so each one is defined here.

• Margin per period is the money left from a customer's payment after the direct costs of serving them, such as payment fees, hosting, and support. It is not revenue. The profit margin on a customer is what counts.
• Retention rate is the share of customers who are still customers in the next period. Churn rate is the share who leave. If 95 of 100 customers stay for another month, retention is 95% and churn is 5%.
• Customer lifetime is how many periods a customer is expected to stay. When the same share leaves every period, the expected lifetime is one divided by the churn rate. At 5% monthly churn that is 20 months.
• Discount rate is the yearly rate used to convert future money into today's money, because profit that arrives later is worth less than profit that arrives now. The time value of money explains why.

Put together, the full method adds up the expected margin from each future period, weights each period by the chance the customer is still there, and discounts it. Berger and Nasr describe the aim as the accumulated and appropriately discounted net contribution margin per customer, once the customer has been acquired. The quick version used in most startups drops the discounting:

Peter Fader of the Wharton School states the idea as a definition: he defines CLV as the present value of the future net cash flows associated with a particular customer.

Lifetime value is not the cost of winning the customer, and it is not average revenue per user. It leaves acquisition cost out on purpose. Berger and Nasr treat the result as the most a manager would be willing to pay to acquire the customer, so the cost is compared against it later, not subtracted inside it.

041 min

Where Lifetime Value comes from

No single person invented lifetime value. It grew out of direct response marketing, where mailers tracked what each name on a list was worth over several years of orders. Berger and Nasr wrote in 1998 that customer lifetime value had been a mainstay concept in direct response marketing for many years, and that it was increasingly being considered in general marketing.

Their paper in the Journal of Interactive Marketing is the usual starting point for the formulas. It gives general models for five types of customer behavior and works a numeric example for each. They also quote Kotler and Armstrong's 1996 definition of a profitable customer: one whose revenues over time exceed, by an acceptable amount, the company's costs of attracting, selling to, and servicing that customer. The excess is the customer lifetime value.

Berger and Nasr cite earlier work by Dwyer (1989), who used a five-year projection because most of the value arrives in the first four or five years. That habit of capping the horizon is still standard. Sunil Gupta and Donald Lehmann later extended the idea into a strategy for managing customers as investments, excerpted by Knowledge at Wharton in 2007.

052 min

A worked example

The numbers below are illustrative, for a subscription app. They are not from a real company. All money is in USD.

Quick version. Expected lifetime is 1 ÷ 0.05 = 20 months. LTV = $9 × 20 = $180. LTV divided by CAC is 3.0. The time to earn back the $60 is $60 ÷ $9, about 6.7 months.

Discounted version. Use a 10% yearly discount rate, which is about 0.8% per month. If margin is collected at the start of each month, then LTV = margin × (1 + d) ÷ (d + c), where d is the monthly discount rate and c is monthly churn. That gives $9 × 1.00797 ÷ 0.05797, about $156.48. Discounting lowered the value by about 13%.

A 3.0 ratio looks comfortable and a 2.6 ratio looks borderline, so the choice of method can change the decision. State which one you used.

Berger and Nasr work a larger case for an insurance company. With a gross contribution margin of $260 per customer per year, $50 of retention spending per year, 75% yearly retention, a 20% discount rate and a 10-year projection, their formula gives $568.78 per customer. Anyone can rerun it from the inputs in the paper.

061 min

How to calculate Lifetime Value

The steps depend on each other, so the order matters.

1. 1

   Choose one group of customers first.

   Calculate for a single segment, such as one plan or one acquisition channel, not the whole base. Every later input is only meaningful for a group that behaves alike.

2. 2

   Find margin per period for that group.

   Take billed revenue from the billing system and subtract direct costs from finance. Leave out fixed costs such as salaries, and leave out the cost of acquiring the customer.

3. 3

   Find retention from cohorts.

   A cohort is every customer who started in the same month. Count how many are still active in each following month. This step needs step 1, because retention differs between groups.

4. 4

   Set the horizon and discount rate.

   Cap the projection, for example at five years, and pick a yearly rate your finance team accepts.

5. 5

   Compute the value.

   Use the quick formula for a first look and the discounted one for a decision.

6. 6

   Compare with CAC.

   Divide LTV by the acquisition cost for the same group and period.

For a benchmark, David Skok, a venture investor, wrote that the best SaaS businesses have an LTV to CAC ratio higher than 3, sometimes as high as 7 or 8. He also stresses that these are only guidelines. They come from one investor's experience, not a measured industry standard.

071 min

How Lifetime Value shows up in tech

Founder. The decision is how much to bid for paid signups. If the group's LTV is $156 and the founder wants the acquisition cost at or below a third of it, the upper limit for a bid is about $52.

Growth product manager. The decision is whether to fund a retention feature or more ads. HubSpot's published unit economics show the trade. According to David Skok's write-up, HubSpot improved its LTV to CAC ratio over five quarters, and the big driver was lowering the MRR churn rate from 3.5% to 1.5%. MRR is monthly recurring revenue. Brad Coffey of HubSpot said the useful lever differed by segment. For small businesses the team improved the product to lower churn and raised the average price. For very small businesses there was less value left to gain, so the team lowered CAC by removing friction from the sales process.

Designer. The decision is which onboarding change to test. A change that raises first-week activation matters only if it later lowers churn, so the design team judges it by its effect on retention, not by clicks.

Engineer. The decision is how to store customer history. Cohort retention needs a start date and an end date for every account and an event record that does not overwrite old values. Without that, nobody can compute the number the others rely on.

082 min

A second case: two companies, two definitions

Real companies define lifetime value differently, and the difference changes what the number means. Two filings with the U.S. Securities and Exchange Commission show it.

EngageSmart, a software company, used a forward-looking formula in its 2021 filing. It calculated LTV as average revenue per customer, multiplied by an adjusted gross margin, less fees paid to channel partners, multiplied by one divided by the annual customer churn rate. It reported that its LTV to CAC ratio exceeded 11 times for 2019 and 2020.

Revolve, an online fashion retailer, used a backward-looking measure in its 2018 filing. It defined LTV as the cumulative contribution profit of a customer cohort, meaning all customers who made their first purchase in one calendar year. Contribution profit is gross profit less fulfillment, selling and distribution costs and the share of marketing spent on retaining that cohort.

The variable that differs is where the lifetime comes from. A subscription company can project it from churn. A retailer has no fixed contract, so it counts what each cohort actually bought. Neither number can be compared with the other company's without reading how it was built.

091 min

Why one average churn rate misleads

Suppose the app from the worked example has two equal-sized groups. Light users churn at 8% a month. Heavy users churn at 2% a month. The blended churn is 5%, the figure used above.

The group-level average is $281.25, which is 56% higher than the $180 from the single rate. The numbers are invented to show the arithmetic. The pattern is real, and Fader describes it. In real data, he writes, there is a steeper drop-off of customers in the first year or two followed by a flattening out. Many companies explain this by saying loyalty grows over time. Fader argues the more likely cause is that the customers with the highest churn leave early, so the people who remain are the ones who were never likely to leave.

Looking only at the customers who remain active repeats this error, known as survivorship bias. The remaining customers look better than a newly acquired one will.

101 min

Common mistakes with Lifetime Value

Teams assume revenue is value. Lifetime value built on revenue overstates every customer who is costly to serve. Use margin after direct costs. The U.S. bank figure above shows how far revenue and profit can differ.

Teams assume one churn rate fits everyone. A single rate treats a new, light user and a three-year heavy user as the same customer. Calculate by segment and by cohort, as in the section above.

Teams assume a long lifetime is better data. At 1% monthly churn the quick formula gives a lifetime of 100 months, more than eight years. No product roadmap or price list is stable that long. Cap the horizon and discount.

Teams assume the number is a fact. LTV is a forecast. It rests on past retention continuing, and a price change, a new competitor or a product change can make the forecast wrong.

Teams assume a high ratio proves the business works. The ratio says nothing about how long it takes to recover the money. A business with a 3.0 ratio that needs 30 months to earn back its CAC can run out of cash first.

111 min

Lifetime Value vs. nearby concepts

Lifetime value is most often mixed up with average revenue per user and with CAC payback. The deciding axis is whether the metric looks at the whole relationship, at one period, or at the time to recover the cost.

Skok writes that months to recover CAC is a good predictor of how well a SaaS business will perform, and that profitability is weak when recovery takes longer than 12 months. Lifetime value tells you whether a customer is worth acquiring at all. Payback tells you how long the money is unavailable. A company needs both.

121 min

Where the evidence on Lifetime Value is contested

The main dispute is about the single retention rate in the quick formula. Fader and Hardie argued in Marketing Science in 2010 that standard textbook CLV uses one retention rate, while at the cohort level retention rates typically rise over time. They showed that ignoring this gives a downward-biased estimate of the value of a customer base, and that the commonly reported retention elasticities also understate the effect of retention improvements.

Their explanation is itself a claim. The paper says it has been suggested that the rising rates come from a sorting effect in a mixed population. Many practitioners still say loyalty grows with tenure, and customers can change. The two accounts predict the same curve, so cohort data alone cannot always separate them.

The formulas have also moved. Skok's page on SaaS metrics notes an update from May 2016 saying his later article on discounted cash flow replaces the formulas in his supplemental document. The author's own update treats the quick formula as superseded.

For a practitioner the dispute leaves a clear rule. Until it is settled, calculate by segment, test how much the result changes when churn changes by a point or two, and report a range, not one figure.

132 min

Frequently asked questions about Lifetime Value

?7 questions

Questions people ask

§7 sources

Sources