ARPA (Average Revenue per Account)

What is ARPA?

ARPA (average revenue per account) refers to the recurring revenue that, on average, is attributable to one paying customer account at a reporting date or over a period. In subscription businesses, it is generally calculated as MRR divided by the number of paying customers and is stated as a monthly value or annualised.

The metric comes from the business with corporate customers, where one customer account can comprise many users but represents only one contractual relationship. It is the starting point of unit economics, the profitability calculation per customer. From ARPA, the gross margin (the part of each euro of revenue that remains after the direct costs of delivering the service) and customer retention, the customer value follows. ARPA also shows which costs of acquiring a customer are justifiable. Investors read from its development whether a company can push through prices and grow with existing customers.

Calculating ARPA: formula, numerator and denominator

Formula: ARPA = recurring revenue (MRR or revenue of the period) / number of paying customer accounts

For a reporting date, MRR at month-end is divided by the number of customers on the same day. For a period, the recurring revenue of the period is divided by the average number of customers in the period; this variant smooths out additions and losses within the period. Both variants are in use, but they must not be used side by side without stating the method.

Numerator and denominator

• Numerator: only recurring fees after discounts and excluding VAT. Set-up, consulting and training are left out, and usage-based fees enter only with the contractually committed minimum purchase.
• Denominator: only paying customer accounts. Users of a free tier (freemium) and customers in a trial phase are not included.
• Customer account: a fixed rule on whether group companies with their own contracts are counted individually or as one account. The rule determines the number of customers and thus ARPA.

ARPA and revenue per customer according to the accounts

If the revenue of a financial year is divided by the number of trade debtors, the result is a different figure from ARPA. Revenue includes one-off revenue and reflects customers who were won or cancelled during the year only pro rata. The number of debtor accounts also often differs from the number of customer accounts, for example when a corporate group is billed through several companies.

With annual prepayments, the difference becomes sharper. The outgoing invoice shows the full annual amount; under commercial law it is to be spread over the service period, and the part not yet earned is deferred as a liability at the balance sheet date. An ARPA formed from invoice amounts therefore fluctuates with the invoicing months. Only the derivation from the contract base, reconciled with accounts receivable accounting, that is, the bookkeeping of receivables from customers, is reliable.

What ARPA shows: segments, mix and expansion

A single average value conceals the structure of the customer base. ARPA becomes meaningful only when it is analysed separately by size class, product and year of entry; the analysis by year of entry is called cohort analysis. This shows whether new customers start at higher values than earlier ones and whether existing customers buy more over time.

A rising overall value is not always an improvement. If mainly small customers leave, ARPA rises arithmetically even though revenue falls. Conversely, a falling ARPA can be a good sign if a company deliberately opens up a new segment with smaller contracts that can be won economically through self-service. Changes in ARPA must therefore always be read together with customer numbers and MRR.

Customer structure in valuation and due diligence

A subscription business is valued on annual recurring revenue (revenue multiple). ARPA explains what this revenue consists of. An ARPA that grows within the existing base through price adjustments and additional modules supports the assumption of future growth without proportional sales costs and thus a higher multiple. A high ARPA with few customers, by contrast, means dependence; investors and acquirers then look at the share of the largest customers in recurring revenue.

In due diligence, the number of customers is checked first: who counts as a paying account, how group customers are counted and whether the number matches the debtor master data. ARPA per segment and cohort is then recalculated and past price adjustments are evidenced from the contracts. Anyone who plans price increases as a growth driver must be able to show that the contracts permit them.

Worked example: calculating ARPA and breaking it down by segment

A software provider has MRR of EUR 250,000 and 125 paying customer accounts as at 31 December.

• ARPA per month: 250,000 / 125 = EUR 2,000, annualised EUR 24,000
• Small segment: 100 accounts with EUR 1,000 MRR each = EUR 100,000
• Medium segment: 20 accounts with an average of EUR 4,500 MRR = EUR 90,000
• Large segment: 5 accounts with an average of EUR 12,000 MRR = EUR 60,000
• Check: 100,000 + 90,000 + 60,000 = EUR 250,000 with 100 + 20 + 5 = 125 accounts
• Simplified customer value at 80% gross margin and 10% annual revenue churn: 24,000 × 80% / 10% = EUR 192,000

If 20 small customers cancel in the following month, MRR falls to EUR 230,000 and the number of customers to 105. ARPA rises to around EUR 2,190, that is, by around 9.5%, while MRR falls by 8%. The example shows why ARPA without a segment breakdown leads to wrong conclusions. The five large accounts carry 24% of MRR. The loss of one large account with an average of EUR 12,000 would weigh more heavily than the loss of ten small customers totalling EUR 10,000.

Distinction: ARPA, ARPU, AOV, ACV, ARR and LTV

• ARPU (average revenue per user): revenue per user instead of per customer account. In the business with individual consumer subscriptions the two figures coincide; with corporate customers with many users they do not.
• AOV (average order value): the revenue per order, typical for online retail with individual purchases. ARPA relates to the customer and a period. Three orders of EUR 40 each from one customer in a month give an AOV of EUR 40 but revenue per customer of EUR 120 a month.
• ACV: the value of a contract expressed per year. ARPA is an average across all customer accounts, and one account can comprise several contracts.
• ARR and MRR: aggregate figures of recurring revenue. ARPA spreads them over the number of customers.
• LTV: the customer value, that is, the gross profit to be expected over the likely duration of a customer relationship. ARPA is one of its inputs.

Common mistakes

• Free accounts in the denominator. If users of a free tier are counted, the metric measures the conversion rate rather than revenue per customer.
• One-off revenue in the numerator. A large onboarding project raises a month's ARPA and distorts any comparison with earlier months.
• Only the overall average. Without segments it remains hidden whether an increase stems from expansion or from the departure of small customers.
• Changing way of counting customers. If group companies are counted individually at one time and together at another, ARPA jumps for no economic reason.
• Monthly and annual values mixed. An ARPA without a stated period regularly leads to misunderstandings about the order of magnitude in conversations with investors.

Frequently asked questions

Is ARPA stated monthly or annually?

Both are usual. In ongoing reporting the monthly value dominates, in valuation documents often the annual value. What matters is that the period is stated with every figure.

What is the difference between ARPA and ARPU?

ARPA relates revenue to the customer account, ARPU to the individual user. A corporate customer with 50 users and EUR 2,000 MRR has an ARPA of EUR 2,000 and an ARPU of EUR 40 a month. In business with companies, ARPA is generally the decisive figure, because contracts, cancellations and price negotiations take place at the level of the customer account.

Is a high ARPA always better?

No. A high ARPA with few customers means dependence on individual contractual relationships. Economically, what counts is whether ARPA covers the costs of customer acquisition and support with a sufficient margin.

How does ARPA feed into customer value?

As a starting figure. In simplified terms, customer value results from ARPA times gross margin divided by the churn rate of the same period. Anyone calculating with monthly values must also use the monthly churn rate.

The Roemer Capital view

Roemer Capital assigns ARPA to unit economics, one of the four pillars against which it measures a company's investor readiness. In preparing a financing round, it analyses ARPA by size class and year of entry over the last eight quarters. This makes visible whether the business is growing with its customers or whether a few existing customers are holding up the average.

A flat trend is no ground for exclusion, but it changes the growth story that a company can tell investors. Roemer Capital therefore clarifies early with the management what room exists for price adjustments and additional modules and whether the customer contracts support it.