Put SaaS terminology into a corporate three-statement model and you get a model that reconciles and tells you nothing. Revenue grows at a rate nobody can defend, sales and marketing is a percentage, and the metrics an investor asks for are typed in from the board deck. The businesses are different, and a model that respects the difference is built from four objects: annual recurring revenue and the flows that move it, the sales force whose capacity sets one of those flows, the billing terms that put cash ahead of revenue, and the customer as the unit of economics.
ARR is a stock, not a revenue line
Annual recurring revenue is the annualised value of the contracts in force at a moment. It is a balance, like a loan book, and it moves by three flows: new ARR from customers won, expansion from customers who buy more, and churn from customers who leave or buy less. Closing ARR is opening plus new plus expansion less churn, and this quarter’s closing is next quarter’s opening.
Revenue is what is earned from that stock as time passes: roughly the average ARR through the quarter, divided by four. The distinction matters because the two can move differently. A quarter of strong bookings lifts ARR at once and revenue only gradually; a quarter of churn does the reverse. Investors value the stock and audit the flows, which is why every retention metric is an identity on the roll-forward rather than a separate assumption.
- Closing ARR = opening + new + expansion − churn.
- Revenue ≈ average ARR through the quarter ÷ 4.
- Apply churn and expansion to opening ARR, never to closing.
Growth is a sales capacity before it is a rate
New ARR does not come from a growth target. It comes from account executives with quotas. A capacity model counts them: hires join, ramp for two or three quarters at partial productivity, become productive, and a share leave each quarter. The productive base times a quarterly quota times a realistic attainment (80% is a good plan; 100% is a hope), plus what the ramping cohort contributes, is the new ARR the company can close.
Built this way, a growth rate becomes a hiring plan someone can be held to, and the model exposes the lag the board always underestimates: a rep hired in January is not fully productive until the third quarter, so the hiring that produces next year’s growth has to happen this year, and its cost lands first.
Divide new ARR by the average contract value and you have the new customer count, which the unit economics need.
Bookings, billings and deferred revenue
Three words that are used as one and mean three things. Bookings are contracts signed: the value of the commitments. Billings are invoices raised: for the share of contracts billed annually in advance, that is up to a year of service invoiced in one quarter. Revenue is the service delivered, recognised evenly across the contract term.
The difference between what has been billed and what has been earned sits on the balance sheet as deferred revenue, a liability that grows as the business grows. Its build is why a subscription company’s operating cash can be better than its operating profit, and why a slowdown shows up in billings a year before it shows up in revenue. A model that recognises revenue when it bills overstates the quarter a large contract lands and understates the following three.
Renewals bill too. A model that applies the annual-in-advance terms only to new contracts, and forgets that the existing book renews on the same terms, understates billings and deferred revenue by most of their value.
| Line | Q1 | Q2 | Q3 | Q4 |
|---|---|---|---|---|
| Bookings | 120 | 0 | 0 | 0 |
| Billings (annual in advance) | 120 | 0 | 0 | 0 |
| Revenue | 30 | 30 | 30 | 30 |
| Deferred revenue, closing | 90 | 60 | 30 | 0 |
The unit economics, read rather than asserted
Customer acquisition cost is sales and marketing over new customers. Payback is that cost in months of gross margin per customer. Lifetime value is gross margin per customer over its life, and the life is where models lie: margin divided by a 1% monthly churn is an eight-year customer, and by 0.5% a sixteen-year one, so a fixed horizon (five years is common) is the guard. Three times CAC is the conventional floor for the ratio.
Net revenue retention is opening ARR plus expansion less churn over opening ARR, for a cohort a year on; gross revenue retention is the same without expansion, and cannot exceed 100%. Quote them together: a strong net figure over a weak gross one is expansion covering for churn. The magic number is net new ARR over the prior quarter’s sales and marketing, the efficiency of the whole engine rather than of one customer; the burn multiple is cash burned per dollar of net new ARR.
Every one of these should be a formula on the model’s own lines. If a metric on the summary page cannot be traced to the ARR roll and the P&L, it is a claim, not a result.
What the rule of 40 summarises, and what it hides
Growth rate plus operating margin above 40 is the shorthand for a software business that is either growing fast enough to justify losses or profitable enough to justify slow growth. It is a useful screen and a poor model. Two companies at 40 can be a 50% grower burning 10% and a 10% grower earning 30%, with different risks, different investors and different values.
The model behind the number is what an investor actually wants: the sales capacity that produces the growth, the retention that keeps it, the billings that fund it, and the unit economics that say whether each new customer is worth having. Build those and the rule of 40 falls out; assert the rule of 40 and nothing does.
Frequently asked questions
What is the difference between ARR and revenue?
ARR is the annualised value of the subscription contracts in force at a point in time: a stock. Revenue is the service earned from that stock over a period: a flow, roughly average ARR through the quarter divided by four. ARR moves the moment a contract is signed or lost; revenue follows over the contract term.
How do you model SaaS growth properly?
From sales capacity: account executives hired, ramping for two or three quarters, productive at a quota with a realistic attainment, and leaving at an attrition rate. The new ARR that capacity can close, plus expansion, less churn, rolls ARR forward. A growth rate with no capacity behind it is a target, not a forecast.
Why does deferred revenue matter?
Because SaaS customers are usually billed before the service is delivered, so cash arrives ahead of revenue and the difference sits as deferred revenue. Its growth is a source of cash that makes a growing subscription business less cash-hungry than its P&L suggests, and its slowdown is the earliest warning that growth is fading.
What is a good CAC payback?
Under twelve months of gross margin is efficient for most software businesses; twelve to twenty-four is common for enterprise sales; beyond that needs a reason, usually very large customers who stay for many years. Read it with net revenue retention: a long payback on customers who expand every year is a different business from a long payback on customers who churn.
Related guides
Build it from the objects that matter.
The SaaS operating model in the Models library runs three years by quarter from sales capacity to ARR, billings, the P&L and the unit economics, every line a live formula you can inspect on the page.