Enter monthly and annual subscription revenue to get ARR and MRR, then add new business, expansion, contraction and churn to see where the number actually moves.
The first line is trivial and is where most explanations stop. The second is the one that tells you anything, because ARR is not a number you calculate so much as a balance you reconcile. Four forces push it, two up and two down, and the headline hides all four.
ARR only means something if the revenue genuinely repeats without a new decision being made. The test is whether the money arrives again next year if nobody does anything.
In. Subscription fees on monthly or annual contracts, committed usage minimums, and support or maintenance contracts that auto-renew.
Out. One-off implementation and setup fees, professional services, hardware, and usage above a committed floor that varies month to month. All of these are real revenue and none of them are recurring.
The contested middle. Usage-based revenue that is stable in practice but not contractually committed. Including it inflates ARR and makes the number fragile; excluding it understates a business that is genuinely predictable. If you include it, say so, and report the committed figure alongside.
Annual contracts are converted by dividing by twelve to get their monthly contribution, not by counting the whole invoice in the month it was billed. Counting billings as ARR produces a number that lurches with your renewal calendar rather than describing the business.
In the worked example, ARR rose from $1,200,000 to $1,398,000. Reported as 16.5 percent growth, that sounds healthy. The movement underneath tells a different story.
| Movement | Amount | What it means |
|---|---|---|
| New business | +$312,000 | Customers who were not there before |
| Expansion | +$126,000 | Existing customers paying more |
| Contraction | -$84,000 | Existing customers paying less |
| Churn | -$156,000 | Customers who left entirely |
| Net new | +$198,000 | What actually reached the headline |
Two retention figures describe this. Gross revenue retention counts only the losses, here 80 percent. Net revenue retention adds expansion back, here 90.5 percent. Net retention below 100 means the existing base shrinks on its own, so every dollar of growth has to be bought. Above 100 means the base grows without new customers, which is the structural difference between a business that scales and one that runs to stand still.
ARR is a revenue figure, so on its own it says nothing about whether the growth was worth having. The number that connects it to the advertising side is the cost of a dollar of net new ARR.
In the example, $468,000 of acquisition spend produced $198,000 of net new ARR. That is $2.36 spent for every dollar gained, which means payback takes over two years before any cost of delivery is counted.
Measured against gross new business of $438,000 instead, the same spend looks like $1.07 per dollar, which is a completely different story. Both figures are defensible and they support opposite decisions, which is exactly why the one being quoted should always be stated.
The honest version is the net figure, because replacing churned revenue is a cost of staying in business, not an investment in growth. If churn is high, acquisition spend is partly maintenance, and calling it growth spend hides the problem.
Net retention below 100 percent means the base shrinks on its own, so every dollar of growth has to be bought.
The most common inflation is counting non-recurring revenue. Setup fees, services and one-off usage get folded in because they arrived in the same invoice, and the result is a number that cannot renew. If ARR includes anything that requires a new decision to happen again, it is not annual recurring revenue.
The second is annualising a good month. Multiplying the best MRR of the year by twelve produces a figure the business has never actually achieved. Use the MRR in force at the period end, on committed contracts only.
The third is reporting net new without the movement behind it. Two businesses can both report $198,000 of net new ARR: one added $198,000 with no churn, the other added $438,000 and lost $240,000. The first is growing, the second is leaking, and the headline is identical.
The last is treating ARR as cash. Annual contracts billed upfront generate cash long before the revenue is earned, and monthly contracts do the opposite. ARR describes the shape of the business, not the balance in the account. Read it beside the burn rate.
Arithmetically yes, but only if the MRR is committed and recurring. Multiplying a strong month that included setup fees or one-off usage produces a figure the business has never actually run at.
No. The test is whether the money arrives again next year if nobody does anything. Implementation, professional services and hardware all require a new decision, so they are revenue but not recurring revenue.
Gross retention counts only losses from churn and contraction, so it can never exceed 100 percent. Net retention adds expansion back, and above 100 percent means the existing base grows without a single new customer.
Because annual contracts billed upfront bring cash in long before the revenue is earned, and monthly contracts do the reverse. ARR describes the shape of the business, the burn rate describes the balance.