ARPU Calculator

Setting it out

Average revenue per user is total revenue divided by users in a period, where users includes everyone on the platform rather than only those who paid.

That is the distinction from revenue per customer: ARPU's denominator includes free users, trialists and dormant accounts.

ARPPU, average revenue per paying user, excludes them, and the gap between the two figures describes the free-to-paid conversion.

A concrete case

A business with 40,000 registered users of whom 12,000 purchased in the year, generating $1,027,296.

ARPU is $25.68; ARPPU is $85.61. The ratio between them, 30%, is the share of registered users who bought.

Improving conversion from 30% to 36% would add 2,400 paying customers and $205,000 of revenue with no additional registrations.

Reporting ARPU alone would show the improvement as a rise from $25.68 to $30.81 without revealing that the mechanism was conversion rather than spending.

What the number hides

ARPU falls when a business acquires users faster than it converts them, which looks like deterioration and is frequently growth. A free-tier signup surge cuts ARPU immediately.

The definition of a user also varies enormously: registered, active, monthly active, and comparing figures across businesses using different definitions is meaningless.

Where to go from here

Report ARPU and ARPPU together, since the pair separates spending behaviour from conversion and either alone conflates them.

Then define user precisely and hold the definition constant, because a change in it will move the metric more than any commercial change.

Where the metric came from and where it fits

ARPU originated in telecoms and subscription media, where the user base is well defined and everyone is billed. It transfers awkwardly to businesses with large free populations.

For an ecommerce business with no free tier, ARPU and revenue per customer are the same figure and only one is needed.

Where there is a free tier or a registered base, the useful version is the pair plus the conversion rate between them, since that identifies whether growth should come from more users, better conversion, or more spending from those already paying.

Splitting it by acquisition channel frequently shows a wide spread, and the cheapest channel is rarely the one producing the highest revenue per user.

That comparison is what justifies paying more for a better source rather than optimising every channel toward the same cost target.

Defining the active window explicitly, thirty days, ninety days, and holding it constant matters more than which window is chosen.

Reporting the paying share alongside the figure prevents a rise in registrations being read as a decline in monetisation.

Comparing it against the cost to serve a user identifies whether a free tier is a funnel or a liability.

Cohort ARPU rather than blended ARPU is the version that shows whether monetisation is improving, since the blend moves with the age mix of the user base as well as with behaviour.

A business adding users quickly will see blended ARPU fall while every cohort performs better than the last, which is the opposite of what the headline suggests.