Key Takeaways

• ACV divides recurring contract value by contract years. A $360,000 three-year contract has an ACV of $120,000 before adjustments.

• One-time fees don't belong in ACV. Stripping a $30,000 implementation fee from that same contract moves ACV to $110,000, so including it overstates the figure by 9.1%.

• ACV measures one contract; ARR measures the whole book at a point in time. Averaging ACV across customers gives you average deal size, not ARR.

• Contracts shorter than a year annualize, which is why pilots inflate an ACV average. A four-month $20,000 pilot reports as $60,000 ACV.

• ACV is only useful next to acquisition cost. The number that matters is how many months of ACV it takes to repay what you spent winning the account.

How is annual contract value calculated?

Divide the contract's total recurring value by the number of years it runs, after removing anything non-recurring.

Worked on a three-year enterprise deal:

Include the implementation fee and you report $120,000 instead, a 9.1% overstatement that repeats on every deal with a setup charge. Across a portfolio where most enterprise contracts carry one, the aggregate error is systematic rather than random, which is what makes it worth catching.

Short contracts annualize the same way. A four-month pilot worth $20,000 has an ACV of $60,000, since $20,000 divided by 0.333 years is $60,000.

What counts as a good annual contract value?

There's no absolute threshold. A good ACV is one that repays acquisition cost fast enough for your funding position, which makes the ratio the metric and the raw number a detail.

The tests that actually decide it:

• ACV against CAC. If it costs $80,000 in sales and marketing to land a $40,000 ACV account, you wait two years for the first dollar of contribution. The same $40,000 against a $10,000 CAC is a strong business.

• ACV against sales motion. A field sales team with a six-month cycle can't be supported by a four-figure ACV. Self-serve can.

• ACV against gross margin. $100,000 ACV at 30% margin contributes less than $50,000 at 80%, which matters most for AI products where inference cost is real.

• ACV trend by cohort. Rising ACV on new cohorts signals successful upmarket movement; a rising average driven by churning small accounts signals the opposite.

Comparing your ACV to a published industry average tells you almost nothing, because the average blends self-serve and enterprise motions that have nothing in common.

What breaks ACV calculations?

Usage-based revenue breaks it first, because a contract with metered overage has no fixed annual value to divide.

The cases that produce a wrong number:

• Metered contracts. A $50,000 committed floor with uncapped overage might bill $50,000 or $300,000. Reporting the floor as ACV understates the account, and reporting the trailing actual makes ACV move every month.

• Ramped commitments. A contract that steps from $60,000 to $100,000 to $140,000 across three years averages to $100,000, which describes no year the customer actually pays.

• Mid-term expansions. An upsell in month seven either restates ACV for the whole term or creates a second contract record, and teams do both inconsistently.

• Multi-currency contracts. Locking the exchange rate at signature and reporting at that rate diverges from what the invoices actually collect.

For metered contracts I'd report the committed floor as ACV and track realized revenue separately rather than blending them, so the forecast stays stable and the variance stays visible.

Related reading

Further reading on contract value and revenue reporting:

• how to measure billing impact on revenue covers the reporting ACV feeds into.

• how billing system affects saas revenue connects billing mechanics to the revenue line.

• saas billing at every mrr stage maps what to track at each revenue stage.

• how to prevent revenue leakage in usage based pricing covers the gap between contracted and collected.

• enterprise billing controls covers the contract terms that decide the number.

FAQ

What is the difference between ACV and ARR?

ACV describes one contract's average annual value; ARR describes the annualized recurring revenue of the entire customer base right now. Sum the ACV of every active contract and you approach ARR, but they answer different questions. ACV tells you how big your deals are, and ARR tells you how big your business is.

Does ACV include one-time fees?

No. Implementation fees, migration charges, training, and professional services are non-recurring, so they belong in total contract value and stay out of ACV. Including them inflates the figure and breaks the comparison with any other company's ACV, since the whole point of the metric is measuring what recurs.

How do you calculate ACV for a month-to-month customer?

Multiply the current monthly rate by 12 and label it clearly as annualized rather than contracted. A month-to-month customer has no annual commitment, so the number is a run rate rather than a contract value. Mixing annualized month-to-month accounts into an ACV average alongside real multi-year contracts makes the average describe neither group.

Should usage-based revenue count in ACV?

Count the committed minimum and report overage separately. The commitment is contractual and belongs in ACV; overage is variable and reporting it as contract value turns a forecast into a trailing measurement. Teams that blend the two end up with an ACV figure that changes every month for reasons that have nothing to do with new business.