Average Revenue per Account (ARPA) and ARPU

SaaS terminology beginner FOUNDERCFO

Published by 4NLab · Updated · Our methods

What is ARPA?

ARPA is average revenue per account: total MRR divided by the number of active accounts, typically measured monthly. It is a core input to LTV and unit-economics calculations.

ARPA tells you whether recurring revenue moved because you have more accounts or because each account pays more. It works best when the account is the thing that signs and pays for the subscription. A workspace with 50 users can still be one account.

Also known as: Average Revenue Per Account

Formula

ARPA = MRR / Total Number of Accounts
Variable Meaning
MRR Total monthly recurring revenue.
Total Number of Accounts Count of active customer accounts in the period.

How to use it

  1. Define what an account is, and use that definition for both subscription revenue and customer counts. Keep free accounts out of the paying base.
  2. Divide closing MRR by active paying accounts on the same reporting date. If you use annual values, label the result annual ARPA.
  3. Split the average by plan, customer size, or acquisition cohort. When a few large customers dominate, track the median and concentration too.

Worked example

Hypothetical worked scenario

A fictional collaboration tool has $24,000 MRR from 120 paying companies. Together those companies have 1,200 paid seats.

Monthly ARPA = $24,000 ÷ 120 = $200 per account. Revenue per paid seat = $24,000 ÷ 1,200 = $20.

The two denominators answer different questions. Plug the $20 seat figure in as account revenue and, in this example, an account-level LTV estimate comes out ten times too low.

ARPU vs ARPA

ARPU (average revenue per user) and ARPA (average revenue per account) share a numerator and differ in the denominator. In a product where every account is one person they are equal, so consumer apps tend to say ARPU and B2B software ARPA. Once accounts have several users, they answer different questions.

QuestionARPUARPA
Divides recurring revenue byActive users or paid seatsPaying accounts (the customer that signs and pays)
Worked example above$24,000 ÷ 1,200 seats = $20$24,000 ÷ 120 accounts = $200
Use it forSeat pricing and per-user packagingLTV, CAC and other per-customer unit economics
Rises whenPrice per seat goes up, or unpaid users leaveAccounts add seats, upgrade or buy add-ons, or small accounts leave

Mixing them breaks unit economics. Churn and CAC are counted per customer, so pair them with ARPA. If your billing system reports ARPU, convert it by multiplying by average paid seats per account.

What the result tells you

A rising average can mean upgrades worked, the mix shifted toward larger buyers, or small accounts left. Before you credit pricing, compare revenue change within the same accounts against the change in customer mix. The average by itself can’t tell you which one happened.

Assumptions and common mistakes

  • ARPU and ARPA sometimes get used interchangeably. When an account has several users, write down which unit you mean.
  • Don’t divide revenue that includes services by subscription accounts and call the result recurring ARPA.
  • With zero paying accounts the average is undefined. A zero-dollar result would hide the missing denominator.

Frequently asked questions

Should I use average or ending account count?

This guide uses an end-of-period subscription snapshot. An average count can make sense for revenue earned over a period, but that’s a different convention, so label it.

Does higher ARPA always mean better economics?

No. Larger customers can need more service, take longer to close, or come with discounts elsewhere. Read ARPA next to delivery cost, retention, and acquisition cost for the same segment.

Is ARPU the same as ARPA?

Not when an account has several users. ARPU divides revenue by users or seats, ARPA by paying accounts. In a product where each account is one person the two are equal, which is why the terms often get mixed up.

Sources and methodology

These references explain the methods. The worked scenarios use hypothetical figures. They are not company results or current market benchmarks.

Found an error or a definition that differs from your reporting? Send a correction with the page URL and the method you use.

Related Concepts