What annual recurring revenue actually claims
ARR makes a specific claim: if nothing changes, the subscriptions we hold today will produce this much revenue over the next twelve months. It is a forward-looking run-rate built from signed contracts, which is why it is the number investors, acquirers and lenders anchor on — and why the way you build it matters more than the number itself.
The word doing the work is recurring. A twelve-month subscription at $14,000 recurs. A $9,000 implementation project does not, no matter how reliably you sell one with every deal. Uncommitted usage that happened to spike last quarter does not. A committed minimum spend does. Anything you would not expect to invoice again next year without doing new work belongs outside ARR.
The second word doing work is annual, and it trips people up in both directions. A three-year, $90,000 contract is not $90,000 of ARR — it is $30,000 of ARR and $90,000 of total contract value. A $390-a-month subscription with no term is not $390 of ARR — it is $4,680, even though the customer can leave in thirty days. ARR normalises everything to one year regardless of how long the commitment runs or when the money arrives.
If you already track monthly recurring revenue, ARR is simply that figure times twelve, and the MRR calculator is the faster route. Build ARR from the contract book instead when most of your revenue sits on annual or multi-year terms, because that is where the contract data lives and where the errors hide.
Run-rate ARR, committed ARR and why diligence asks for both
Run-rate ARR counts what is live and billing today. Take every active subscription, divide its total contract value by its term in years, and add the results. It answers "what is the machine producing right now".
Committed ARR answers a different question: "what have we actually contracted". It starts from run-rate ARR, adds the annualised value of executed contracts whose service period has not begun yet, and subtracts the annualised value of accounts that have given notice of non-renewal. A company that closed a large deal on the 28th of the month with a start date six weeks out has real, signed, enforceable revenue that run-rate ARR cannot see.
The two figures diverge most at exactly the moment people care most — during a financing or a sale, when a strong quarter's bookings have not yet converted to billing. That is why an experienced analyst asks for both and reconciles them line by line. It is also why you should never present one without labelling which it is.
Be aware that committed ARR is not a standardised term. Some investors include only contracts with a remaining committed term and therefore exclude your entire month-to-month base. Others count total remaining contract value rather than an annualised figure. Write your definition down before you use the phrase, and expect to be asked to defend it.
The calculator also reports what fraction of run-rate ARR sits on accounts with no committed term. That number is a quality measure, not a size measure: $1M of ARR that is 90% under annual contract behaves nothing like $1M that is 90% month-to-month, even though both annualise identically.
Worked example: a 135-subscription book with mixed terms
Take the calculator's default book. You have 42 one-year contracts averaging $14,000, eight multi-year contracts averaging $90,000 of total value over a three-year term, and 85 month-to-month accounts at $390. You also billed $180,000 of implementation and training work over the last twelve months, hold $95,000 of ARR on contracts that start next quarter, and have received notice on $60,000 of ARR.
- Annualise the one-year contracts. 42 × $14,000 = $588,000. A one-year term needs no conversion.
- Annualise the multi-year contracts. $90,000 ÷ 3 years = $30,000 a year each. 8 × $30,000 = $240,000. Total contract value across those eight deals is $720,000 — a much bigger number, and not ARR.
- Annualise the month-to-month base. $390 × 12 = $4,680 a year each. 85 × $4,680 = $397,800.
- Add the segments. $588,000 + $240,000 + $397,800 = $1,225,800 of run-rate ARR.
- Equivalent MRR. $1,225,800 ÷ 12 = $102,150.
- Average contract value. 42 + 8 + 85 = 135 live subscriptions. $1,225,800 ÷ 135 = $9,080 annualised per subscription.
- Committed ARR. $1,225,800 + $95,000 − $60,000 = $1,260,800.
- Recurring share. $1,225,800 ÷ ($1,225,800 + $180,000) = $1,225,800 ÷ $1,405,800 = 87.2%.
- Uncommitted share. $397,800 ÷ $1,225,800 = 32.5% of ARR has no term beyond the current month.
Two lessons fall out of the arithmetic. First, the average contract value of $9,080 is far below the $14,000 headline on a one-year deal, because 85 of the 135 subscriptions are small month-to-month accounts — ACV computed across a mixed book tells you about the mix, not about your sales motion. Second, the multi-year contracts annualise to $30,000 against $14,000 on a one-year deal, which is a healthy premium per account and the opposite of the discount most people assume multi-year deals carry.
How different contract shapes annualise
| Contract as signed | Term | Total contract value | ARR contribution | MRR contribution |
|---|---|---|---|---|
| $1,200 a year, prepaid | 12 mo | $1,200 | $1,200 | $100.00 |
| $99 a month, no term | 1 mo | $99 | $1,188 | $99.00 |
| $300 a quarter | 3 mo | $300 | $1,200 | $100.00 |
| $60,000 over two years | 24 mo | $60,000 | $30,000 | $2,500.00 |
| $90,000 over three years | 36 mo | $90,000 | $30,000 | $2,500.00 |
| $250,000 over five years | 60 mo | $250,000 | $50,000 | $4,166.67 |
| Ramped: $80k, $120k, $160k | 36 mo | $360,000 | $80,000 in year 1 | $6,666.67 |
| $40,000 subscription + $15,000 setup | 12 mo | $55,000 | $40,000 | $3,333.33 |
The ramped deal is the exception to simple division: report the current contract year's rate as ARR, not the three-year average of $120,000, or you overstate today's run-rate by half. The last row shows the setup fee excluded entirely.
How to read the result: growth, quality and multiples
Read ARR on three axes: size, growth and quality.
Size sets the stage you are at. The common shorthand milestones are $1M of ARR (the product works and someone will pay), $10M (repeatable go-to-market), and $100M (a durable business). They are conventions, not thresholds with any theory behind them.
Growth is what actually gets valued, and it is measured on ARR rather than on GAAP revenue precisely because ARR responds immediately to new bookings. The single most-quoted efficiency test pairs your ARR growth rate with your profit margin — that is the Rule of 40, and it is the reason a company growing 60% at breakeven and one growing 10% at 30% margins can trade on similar multiples.
Quality is where most ARR figures quietly fail. Four questions separate a durable book from a fragile one of identical size: what share is under committed term rather than month-to-month; what share of total revenue is recurring at all; what does net revenue retention look like on the existing base; and how concentrated is the book in a handful of accounts. A book where one customer is 30% of ARR is a different asset from one where the largest is 3%. Work the retention side with the net revenue retention calculator.
On multiples: revenue multiples applied to ARR move with interest rates, growth and market sentiment, and any specific figure quoted here would be stale within months. What is stable is the relationship — higher growth, higher net retention and a higher recurring share all command a higher multiple, and services-heavy revenue is discounted sharply. If you need to translate a multiple into an actual valuation, run it through the enterprise value calculator.
ARR is not revenue, and confusing them is a diligence red flag
ARR is a non-GAAP operating metric. No accounting standard defines it and no auditor opines on it; when a public company presents it, the SEC's 2020 guidance on disclosure of key performance indicators governs what has to be said about how the metric is calculated. Recognised revenue is the opposite: it follows ASC 606 in the United States and IFRS 15 elsewhere, and you recognise as you satisfy the performance obligation. A $120,000 annual contract signed on 1 October produces $30,000 of recognised revenue in that calendar year and $120,000 of ARR from day one. Neither number is wrong; they answer different questions.
Two failure modes recur. Presenting ARR as though it were audited revenue in a fundraising deck invites a painful reconciliation later. And annualising a single strong month — taking December's total billings, multiplying by twelve, and calling it ARR — folds one-time fees and usage spikes into a supposedly recurring figure. Build ARR from contracts, and disclose your definition alongside it.
Mistakes that inflate or distort an ARR figure
- Counting total contract value as ARR. A three-year $300,000 deal is $100,000 of ARR. Reporting $300,000 overstates the run-rate by three times and is the most common error in early-stage decks.
- Including implementation and services revenue. It is real revenue with real margin, but it does not recur and it is valued at a fraction of subscription revenue.
- Annualising a single month's billings. This bakes seasonality, usage spikes and one-time fees into a metric that is supposed to strip them out.
- Averaging a ramped contract. A deal that steps from $80,000 to $160,000 over three years contributes its current-year rate, not the $120,000 average.
- Ignoring accounts under notice. Revenue that has already been cancelled effective at renewal is still in run-rate ARR. Report committed ARR alongside it so the reader can see the difference.
- Counting unsigned pipeline. A verbal yes, a signed order form pending legal, and a countersigned contract are three different things. Only the last belongs anywhere near ARR.
- Using list price instead of net price. ARR is what the customer is obliged to pay after discounts, credits and negotiated concessions.
- Double-counting a renewal. When an annual contract renews, ARR does not increase; only a price rise or added scope is expansion.
What this calculator assumes, and what it leaves out
The model groups your book into three shapes — one-year, multi-year and month-to-month — with an average value for each. That is the right resolution for a board update and for most companies below roughly $20M of ARR. Above that, or where contract values vary by an order of magnitude, compute ARR per contract in your CRM and enter the segment totals here as a single one-year row.
It divides multi-year total contract value evenly across the term, which is correct for flat contracts and wrong for ramped ones. If a meaningful share of your book ramps, enter those contracts separately at their current-year rate rather than their average.
It does not handle foreign-currency contracts (annualise at a consistent rate and disclose which), contractual escalators that bite in later years, ARR held by customers in a payment-failure state, or the cash timing of prepayments — which matters enormously for survival even though it never touches ARR. For that, use the burn rate and runway calculator. To push the book forward with churn and expansion applied month by month, use the subscription revenue forecast calculator.
Key terms
- ARR
- Annual recurring revenue: the annualised value of live recurring subscriptions, excluding one-time and non-committed revenue.
- TCV
- Total contract value: everything a customer is committed to pay across the whole term, including services. Always larger than annual ARR on a multi-year deal.
- ACV
- Annual contract value: the ARR of a single contract. Averaged across a book it is sensitive to mix, so read it by segment.
- Committed ARR
- Run-rate ARR adjusted for signed contracts that have not started and for accounts under notice of non-renewal. Definitions vary between investors — state yours.
- Backlog
- Contracted revenue not yet recognised. A three-year contract carries roughly two years of backlog at the end of its first year, and none of it is visible in ARR.