A customer lifetime value statistic is useful only when its population, contribution definition, and horizon are clear. Read external evidence against a worked internal customer file, and distinguish observed contribution from expected lifetime value.
Customer lifetime value statistics are comparable only when they describe the same kind of customer, value measure, and time horizon. Check those definitions before treating a published figure as a target. A large number can reflect a longer forecast, a revenue basis, or a few exceptional accounts rather than stronger customer economics.
A statistic labeled customer lifetime value might mean historical sales, expected future contribution, or a combination of realized and projected amounts. Those measures answer different questions. The lifetime value glossary defines the core concept; the evidence review starts by establishing exactly what a source calculated.
For every external figure, record the publisher, publication date, customer population, currency, cohort dates, observation cutoff, and calculation method. Identify whether the denominator includes everyone acquired or only customers still active. Then find the treatment of returns, service costs, acquisition costs, and discounting.
Timing alone can produce different results. Fader and Hardie's research note on CLV formulas explains how including the first payment and changing the timing of cash flows changes the appropriate formula. A shared metric name does not guarantee a shared calculation.
Start with the customer file you can reconcile. Keep realized revenue and attributable costs separate from assumptions about future purchases, retention, contribution, and discount rates. A completed observation window is evidence about that window; it does not establish the full duration of the relationship.
For future value, finance should approve the cost basis and discount convention. The commercial owner should explain the business conditions supporting the retention and purchase assumptions. Retain the valuation date so an estimate can be reviewed against what subsequently happened.
Use the existing LTV formula and worked examples for the calculation method. The purpose here is to decide whether statistics built from that method belong beside another source's figures.
An illustrative US business-services company reviews customers acquired in one cohort. Each customer has completed the same first-year observation window. The export includes net revenue and agreed attributable delivery and service costs; it excludes acquisition spending, corporate overhead, and future activity.
| Customer | First-year net revenue (USD) | Attributable costs (USD) | Observed contribution (USD) |
|---|---|---|---|
| A | 600 | 400 | 200 |
| B | 900 | 600 | 300 |
| C | 1,200 | 800 | 400 |
| D | 1,500 | 1,000 | 500 |
| E | 4,800 | 3,200 | 1,600 |
| Total | 9,000 | 6,000 | 3,000 |
Mean observed contribution = Total observed contribution / Customers in the cohort
Median observed contribution = Middle customer contribution after sorting the values
The arithmetic is explicit:
| Step | Calculation | Result |
|---|---|---|
| Reconcile contribution | 9,000 − 6,000 | 3,000 USD |
| Count customers | A through E | 5 |
| Mean contribution | 3,000 / 5 | 600 USD |
| Median contribution | Middle of 200, 300, 400, 500, 1,600 | 400 USD |
| Mean net revenue | 9,000 / 5 | 1,800 USD |
| E's share of contribution | 1,600 / 3,000 | 53.3% |
The mean contribution reconciles back to the total: $600 × 5 = $3,000. E's share is rounded to one decimal place; dollar results are exact. The table shows why a revenue-based average of $1,800 cannot be compared directly with a contribution-based average of $600.
These are first-year observed statistics, not lifetime estimates. They contain no future purchases, survival assumptions, or discounting. The ledger-based customer value guide covers what historical records can establish and where estimates begin.
The mean is $600 while the middle customer contributes $400. Customer E accounts for 53.3% of the total contribution in this small file. The NIST guide to measures of location explains how extreme values affect the mean and why the median can describe a different aspect of the distribution.
As a sensitivity check, contribution excluding E is 200 + 300 + 400 + 500 = $1,400. Dividing by the remaining four customers gives $350. That check reveals concentration; it is not a reason to remove E from the reported cohort average.
Report the full base and the sensitivity result together. Investigate whether the large account represents a repeatable customer type, an unusual contract, or a data error. Do not generalize from these five illustrative records to an acquisition budget for a larger market.
Put an external statistic beside the internal result only after completing a comparison table. Missing information is itself a result: label the figure as not comparable and explain the missing definition.
| Dimension | Internal example | What an external source must establish |
|---|---|---|
| Customer unit | One business account | Account, subscriber, buyer, or relationship |
| Population | All customers in the acquisition cohort | Inclusion and exclusion rules |
| Value basis | Net revenue less listed attributable costs | Revenue, gross profit, or contribution |
| Horizon | Completed first year | Observed period and any projected period |
| Timing | Historical amounts, undiscounted | Valuation date and discount convention |
| Summary | Mean and median with count | Statistic type and customer count |
Do not repair an unknown horizon with an assumed multiplier. Do not substitute a current exchange rate for a missing currency basis and call the comparison settled. Use customer lifetime value by industry for sector differences, and customer profitability versus lifetime value when the disagreement is historical versus future value.
Finance owns the comparison's measurement basis. Sales and marketing own the proposed decision, such as which acquired cohort to investigate. An external statistic can prompt a question, but the approved business case should use reconciled internal evidence and explicit assumptions.
Bring customer revenue and attributable cost exports, with cohort dates and any separately stated assumptions; Covirage maps the columns, and its tools compute and check the figures. The external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.
See Covirage for customer profitability for the supporting financial workflow.
There is no useful universal dollar threshold. The answer depends on contribution, acquisition spending, service costs, retention, and the time value of money. Compare customers and cohorts using the same definitions before considering whether an external industry figure offers a meaningful reference.
Report both when possible. The mean helps connect a customer-level figure to the total value of a defined base. The median describes the middle customer and is less affected by unusually large accounts. Neither removes differences in cost scope or observation horizon.
Historical revenue records sales already observed. It excludes costs and usually does not cover the customer's complete relationship. Label the measure by its actual scope, such as first-year net revenue, and keep any future contribution estimate and its assumptions separately identified.