Average customer lifespan by industry: compare fairly

Average customer lifespan by industry cannot be compared fairly without matching the relationship definition and the time available to observe it. A canceled contract has an end date; an active account has only a tenure so far. Start with dated customer records and publish those two measures separately before using an outside benchmark.

What the industry label leaves out

A subscription, an occasional equipment purchase, and a continuing supply relationship can all describe a customer. Their ending events differ, so a table of industry averages can hide more than it explains.

These are measurement choices, not published industry averages. Keep the choice visible beside every result. The discussion of dormancy by industry covers how purchase cadence changes an inactivity threshold; use that rule consistently before comparing spans.

An external benchmark is usable only if its source discloses the customer unit, dates covered, end rule, sample, and treatment of active accounts. If any of those are missing, treat the figure as background rather than a performance target. A household, a business account, and a location are different denominators even when they buy the same product.

Build an export that preserves what is unknown

Use one row per customer relationship, with a stable account ID, segment, acquisition date, first eligible purchase or activation date, last eligible purchase date, recorded end date, and status at the cutoff. Add an indicator for accounts whose earlier history is unavailable. A system migration date must not quietly become their acquisition date.

Set one report cutoff and retain customers that left before it. An export of current active accounts alone cannot describe completed customer spans. Keep returns, internal transfers, and duplicate account records from resetting the start date.

For a relationship with a known end:

Completed span in days = Recorded end date − Relationship start date

For an account still active at the cutoff:

Observed tenure in days = Report cutoff − Relationship start date

The second figure is a lower bound on that customer's eventual lifespan. Statisticians call this right censoring: the event has not been observed before observation stops. NIST explains right-censored observations using reliability tests; the same missing-end-date principle applies here. This application does not imply that a customer relationship behaves like a physical product.

A worked cohort: time observed is not full lifespan

Consider an illustrative US contract-services company. These accounts all activated on January 1, 2025. The report cutoff is January 1, 2026, and an explicit service termination defines a completed relationship. Date subtraction counts elapsed calendar days, excluding the start date.

Use the same rows to produce the following clearly labeled summaries. Averages are rounded to one decimal day.

The completed-span mean answers how long the ended relationships lasted. It does not answer how long the original cohort will last. The all-account mean summarizes time observed, with continuing relationships stopped administratively at the cutoff. Neither gives remaining life for accounts C, E, or F.

Check the result by reconciling status counts: three ended plus three active equals six original accounts. The segment day totals, 636 and 1,003, must sum to 1,639. No relationship end date should fall after the cutoff in this report, and no elapsed duration should be negative.

Compare cohorts before declaring an industry effect

The example supports a limited finding: enterprise accounts accumulated more observed time in this acquisition cohort. It does not establish a universal enterprise lifespan or explain why the difference arose. Contract terms, acquisition channels, and account selection could all matter.

For your own comparison, group customers by acquisition period and give each group the same follow-up horizon. Show the original customer count, number with known ends, active count, and missing-history count beside tenure. Compare similar commercial relationships within an industry using the same definition of similar customers.

If decision makers need an estimated retention curve that accounts for censoring, seek a suitable survival analysis with its assumptions documented. NIST's Kaplan–Meier guidance describes a method that accommodates censored observations without assuming a particular lifetime distribution. It still needs correctly recorded events; a curve cannot repair an arbitrary churn definition.

In noncontractual businesses, keep first-to-last-purchase span separate from time since last purchase. An account first buying on January 1, 2025, and last buying on July 1, 2025, has a purchase span of 181 days. At the January 1, 2026, cutoff it has also gone 184 days without another purchase. The latter is an inactivity gap, not additional observed purchasing life.

For ecommerce comparisons, average repeat customer rate answers a different question about buying again within a window. A repeat rate cannot substitute for a relationship-duration measure.

What decisions the report supports

Revenue operations should own account identity and activation dates. Customer success or sales should own recorded status and the evidence behind an end event. Finance should own the reporting cutoff, reconciliations, and the distinction between observed results and assumptions used in customer lifetime value.

Use comparable tenure cohorts to investigate early departures, review renewal timing, and test whether a retention action reaches accounts before they leave. Do not multiply the observed mean by annual margin and call it a remaining-value forecast. Continuing customers have already survived their earlier periods, and their future is a separate estimation question.

Where it goes wrong

• Dropping active accounts. An ended-only average selects the relationships that have already finished and can understate the eventual experience of a recent cohort.
• Mixing observation windows. Older customers have had longer to accumulate tenure. Equal cutoffs alone do not give different acquisition cohorts equal follow-up.
• Calling last purchase a departure. Seasonal and infrequent buyers can return. State the inactivity threshold and retain the last transaction date separately.
• Using incomplete histories. An imported first order may be the first visible order, not the customer's actual start.
• Changing customer units. Merged accounts, branches, and legal entities change the denominator unless the identity rule remains consistent.

Compare observed tenure on your own file in Covirage

Bring a dated customer export with segment, start date, supplied status, and recorded end date or reporting cutoff. Covirage maps the columns, and its tools compute and check observed tenure summaries; the external AI model explains the results with every figure traced to its rows. Create a free Covirage account, then set up your business and bring a sample of your own data.

See renewal watch for the supporting customer-retention context.