What Is Average Revenue Per Account (ARPA)? (SaaS Glossary)

Average revenue per account (ARPA) is total monthly recurring revenue divided by active accounts, the revenue density of your customer base at the account level.

Across subscription businesses, a higher ARPA gives accounts more expansion room, which is why some products push net revenue retention (NRR) above parity while others can't.

Key takeaways

• Divide total monthly recurring revenue by active accounts to get ARPA.
• Count each account once at its full contracted revenue, regardless of user count.
• Exclude free trial and freemium accounts, or the figure understates your paying base.
• Track revenue per account, not revenue per user, when billing happens per account.
• Say 10% of accounts upgrade tiers, and ARPA rises 13% with no new customers.
• Picture $500-ARPA products at top-quartile NRR, 44 points above sub-$10 peers.

How is ARPA calculated?

ARPA is calculated by dividing your monthly recurring revenue by the total number of active accounts.

Each account counts once at its full contracted revenue, however many users sit inside it. Say a team of ten is on a $500-a-month plan, which is one account worth $500, not ten users worth $50 each. That account-as-unit rule is the part folks get wrong.

The unit matters because most subscription billing happens at the account level, where one invoice covers every seat. Divide that revenue by user count and you split one invoice into ten smaller numbers. Your per-account figure then drops below what your pricing actually delivers.

Picture a product with 200 paying accounts that average $75 a month each. Suppose twenty of those accounts upgrade from a $50 plan to a $150 plan. The upgrade moves ARPA without changing the account count:

ARPA rises from $75 to $85, a 13% lift with no new customers. That $10 increase per account is expansion MRR, the revenue existing accounts add when they spend more.

Free accounts cause the same problem, because they add to the account count without adding revenue. Imagine a product with 1,000 freemium accounts and 200 paying at $75, producing $15,000 MRR but ARPA near $12.50 once all 1,200 accounts are counted.

That number understates what paying customers generate, so segment paying from free before you compute anything.

Why ARPA matters (the NRR connection)

ARPA strongly predicts how much room a base has to push net revenue retention above 100%, because price point sets how much an account can expand.

In ChartMogul's 2023 data, top-quartile companies under $10 a month ARPA average 65% NRR, while those over $500 average 109%. Most of that 44-point spread tracks price point.

High-ARPA accounts have room to grow and low-ARPA accounts don't.

Pretend an account pays $500 a month and can move up a tier or add seats. Each upgrade adds expansion MRR, the one NRR input that can push retention above par. A flat-rate subscription has nothing to expand, so cutting churn is its only lever.

Assume your product is $10 a month flat, so a happy customer has no upgrade to make. Without an upgrade path, NRR has no room to grow above the base.

For example, adding a $20 tier gives accounts a real upgrade path. Accounts that move up generate expansion MRR, and customers who could only churn before now add revenue.

Raising prices alone doesn't lift NRR. Force an increase and some customers downgrade or cancel, adding contraction and churn that can wipe out the ARPA gain. The mechanism only works when accounts move up by choice.

Use our NRR calculator to model what an ARPA shift is worth.

ARPA vs. ARPU (which metric to use)

Use ARPA when you bill per account and ARPU when each user pays on their own. ARPU divides revenue by individual users, while ARPA divides by accounts. For team-based software, one account holds many users, so the two numbers pull apart fast.

The split matters most when you want to know whether an account is growing. B2B software bills one account at a team or per-seat rate, and the NRR inputs that matter, expansion MRR and contraction MRR, are measured per account.

Track revenue per user instead, and a growing account can look flat or even shrinking. Adding users lowers the per-user average even as the account pays you more.

That paradox is easier to see with numbers. Take a product with 50 accounts at $200 a month, each holding 4 users on average. Consider a single account upgrading to a 6-user plan at $280:

That account pays more, so ARPA rises. ARPU falls because the two extra users dilute the per-user figure. Tracked by ARPU alone, real expansion reads as decline.

The exception is genuinely per-user products. On a streaming service or a personal productivity app, every user pays their own subscription. There's no real account above the user, so ARPU is the honest unit. Running ARPA there gives you the same number and adds a confusing label.

FAQ

What is a good ARPA for SaaS?

There's no universal good ARPA, because the right figure depends on your go-to-market motion. SMB products tend to have lower ARPA than mid-market products, and enterprise contracts reach five figures and beyond.

How does ARPA relate to lifetime value?

ARPA is the starting input for the simplest lifetime value (LTV) estimate, where LTV equals ARPA divided by your churn rate. For example, with $75 ARPA and 3% monthly churn, LTV works out near $2,500 per account. It's a rough cut, not a substitute for cohort modeling, but it's the most common next step after ARPA.

Can ARPA increase without adding new customers?

Yes. Existing accounts that upgrade, add seats, or buy add-ons raise total revenue while the account count stays flat, and that motion is exactly what expansion MRR measures.