- Formula
revenue in the period / users in the period- Unit
- Currency
- Direction
- Higher is better
- Across periods
- Recomputed from totals, never averaged
ARPU is the revenue of a period divided by the number of users in that period. €30,000 of revenue in March from 1,500 users gives an ARPU of €20.
The division is trivial. The hard part is the denominator: “user” can mean everyone who signed up, everyone who was active, or only those who paid. Each choice gives a different number with the same name.
The formula
ARPU = revenue in the period / users in the period
“Users in the period” is either the unique users active in the period or the average of the start and the end of it — say which one. March revenue of €30,000, 1,500 users active in March, 400 of them paid:
ARPU = 30,000 / 1,500 = €20
ARPPU = 30,000 / 400 = €75
Three versions are in common use:
| Denominator | Used for | |
|---|---|---|
| ARPU | all users, free ones included | apps and free plans: the value of an average user |
| ARPPU | paying users only | price level: what one payer brings |
| ARPA | paying accounts (companies) | B2B, where one account has many users; the input of LTV |
The three are linked: ARPU = ARPPU × share of users who pay. The LTV:CAC
calculator takes ARPA, the monthly revenue per paying account.
The period is part of the number. For a subscription, the ARPU of a quarter is about three times the ARPU of a month. Write “per month” or “per year” next to it.
What it does not show
It hides the spread. An average of €20 can be 1,450 users paying nothing and 50 users paying €600. The average describes nobody. Split ARPU by plan, or look at the median next to it.
It does not know your margin. €20 of revenue from a user on a plan with heavy usage costs is not the same as €20 from a light user.
It says nothing about time. ARPU is revenue per user for one period. How long the user keeps paying is a separate question, and only the two together give lifetime value.
It does not tell channels apart. A blended ARPU hides a channel that brings many free users who never pay. ARPU by channel needs both the users and their revenue tied to the campaign that brought them.
How it breaks
ARPU and ARPPU mixed up. More users start paying, many of them on a cheap plan: ARPU goes up and ARPPU goes down. The two numbers move in opposite directions on the same event, and quoting the wrong one tells the wrong story.
Users summed or averaged across periods. The same person active in January, February and March is one user of the quarter and three “monthly users”. The ARPU of a quarter is the quarter’s revenue divided by the quarter’s unique users — not the average of three monthly ARPUs, and not divided by the sum of monthly users. It is a derived metric: recomputed for every period, never added up.
Revenue and users from different places. Revenue comes from billing or the CRM, users from product analytics. They define a person differently and often use different time zones; the ratio inherits both differences.
One-time revenue inside. A setup fee or a yearly payment taken in one month makes ARPU jump for that month and fall after it, with no change in what users pay.
Revenue before refunds. Refunds come later than sales; ARPU built on gross sales is higher than the money you keep.
How to build it in KO
KO does not have ARPU as a ready metric. It counts buyers, not users of your product: a person who signed up and never paid exists in KO only if they came in as a lead.
What you can build from what KO has:
- ARPPU — revenue per paying customer. In the project’s Metrics catalogue, create your own metric with the formula Revenue by purchase date ÷ Buyers. Buyers counts each customer once in the period, however many purchases they made, and Revenue by purchase date has refunds taken out. In a breakdown by campaign this gives revenue per paying customer by channel. Both work when the project has its successful sale stages marked.
- Revenue per buyer. In the Entities section, the Buyer entity shows sales, revenue by purchase date, repeat sales and refunds for each customer: the spread that an average hides.
- Revenue per lead. If every sign-up arrives as a lead, Lead value is the closest thing to ARPU over sign-ups: revenue of leads that had time to reach a sale, divided by their number, per channel. For a channel with few leads the number is pulled towards the project average.
What is missing: the number of active users of your product and the plans they are on. These come from a billing system or product analytics, not from the CRM; we connect Stripe and subscription billing systems on request.
Related
- Customer lifetime value — ARPU times how long a customer pays, minus costs.
- Monthly recurring revenue — MRR divided by paying accounts is the monthly ARPA.
- Average order value — revenue per purchase rather than per person; the two differ by the number of purchases.
- Churn rate — the other half of LTV: ARPU says how much, churn says how long.
- Customer acquisition cost — what you pay for the user whose revenue ARPU averages.