How to Calculate ARR With Mixed Contract Terms

ARR is straightforward when every customer is on a flat annual subscription. Almost no B2B SaaS company is. Once you have multi-year deals, ramped pricing, monthly customers, usage commitments, and pilots in the same book, ARR becomes a normalization problem. This guide covers the rules that make a mixed book add up to one defensible number.

What is the base ARR formula?

ARR is the annualized value of all recurring revenue under contract as of a specific date.

``` ARR = Sum of (Annualized Recurring Value of Every Active Contract) ```

The two words that do the work are annualized and recurring. Annualized means every contract gets converted to a twelve month equivalent regardless of its billing frequency or term length. Recurring means the revenue would repeat if the customer did nothing, which excludes services, one-time fees, and anything the customer has to buy again by choice.

ARR is also a point-in-time snapshot. It is what the business runs at today, not what it billed last year or what it will bill next year.

How do you annualize each contract type?

Convert everything to a twelve month equivalent using the current contracted rate. The ramped contract rule causes the most disagreement. Averaging a ramp across the term reports revenue you are not currently receiving, which breaks the point-in-time meaning of the metric. Using the current year's value keeps ARR honest and moves the increase into expansion when it actually arrives, which is also where it is most useful for retention math.

How should usage-based revenue enter ARR?

Report committed ARR and variable overage as separate lines.

Consumption models break the clean run-rate assumption because the number changes without any customer decision to buy or cancel. Two rules keep it usable:

1. Contracted ARR includes the annual committed minimum at full value. This is the floor the customer has agreed to pay. 2. Variable ARR annualizes trailing usage above commitment, typically over the last three months to smooth seasonal swings, and is reported next to contracted ARR rather than inside it.

A customer with a $150,000 commitment consuming at $220,000 annualized has $150,000 contracted ARR and $70,000 variable ARR. Reporting a single $220,000 figure sets an expectation that a slow quarter will look like contraction, when the customer has done nothing except use less of a variable product.

What breaks ARR calculations most often?

Inconsistent treatment across periods, not the individual rules.

The specific rule you pick for ramps or overages matters less than applying it identically every month. A book normalized one way in Q1 and another way in Q3 produces a growth number that is partly methodology, and no downstream metric built on it can be trusted. Consistency is what makes prediction possible, and a consistent imperfect rule beats an occasionally correct one.

Four common failures:

- Booking multi-year TCV as ARR. A $900,000 three year deal reported as $900,000 ARR overstates the run rate by 200 percent and produces a phantom churn event at renewal. - Including one-time fees. Implementation revenue in ARR guarantees an artificial contraction next period. - Counting signed but not started contracts. A contract with a future start date is backlog, not ARR, until the service period begins. - Leaving cancelled logos in the run rate. Accounts in a notice period are still active ARR, but accounts past their end date are not.

How does ARR normalization affect forecasting?

A clean ARR base is what makes retention and expansion forecasts comparable across periods.

The retention waterfall depends entirely on ARR being calculated the same way at the start and end of each month. ORM's structure runs beginning ARR, churned customer ARR, churned product ARR, product decrease ARR, new customer ARR, new product ARR, increased product ARR, and ending ARR, where beginning ARR always equals the prior month's ending ARR. If ARR is normalized inconsistently, that reconciliation never ties and every retention rate derived from it is an estimate.

The same applies on the new business side. Deal values in pipeline should be annualized on the same basis as booked ARR, otherwise pipeline and bookings are measured in different units. One pattern worth checking: pipeline average deal size frequently runs well above closed-won average deal size, for example $80,000 in pipeline against $40,000 actually closed. That gap distorts coverage and forecast alike.

For the mechanics of turning a clean ARR base into a forward number, see how to forecast revenue and how to create a sales forecast. For how ARR normalization affects retention reporting, see net revenue retention.