The Real Difference Between ARR and Actual Cash Flow.
ARR describes contracts. Cash describes your bank balance. A public company's filings and a worked example show how the same recurring revenue can leave you with a comfortable runway or an empty account, depending on things ARR never records.
Annual recurring revenue is the number SaaS founders quote first, investors ask for first and dashboards put at the top. It's a useful number. It's also a statement about contracts, not about money. A company can double its ARR in a year and run out of cash in the same year, and nothing about that is a contradiction.
This issue shows how, in two ways. First, a public company's own filings, where ARR, revenue, profit and cash all tell different stories about the same year. Second, a worked example with every assumption on the page, so you can see exactly which levers move cash while ARR stays put. The example is hypothetical and labelled as such. The filings are real.
What ARR Is, and What It Isn't
ARR has no accounting standard behind it. It's an operating metric each company defines for itself, which is why the definitions in public filings differ. Samsara, the fleet and equipment software company, defines it as the annualized value of subscription contracts that have already started recognizing revenue. That wording excludes one-time fees, and it also excludes contracts that are signed but not yet live.
ARR also isn't revenue. Revenue is what the company recognized over the past year. ARR is a snapshot of where the run-rate stands now, so for a growing company it sits above the year's revenue. Samsara's fiscal 2024 shows the gap: ending ARR of $1,102.0 million against full-year revenue of $937.4 million, about 18% higher, because most of the growth happened during the year.
And ARR isn't cash. It says nothing about when customers pay, what it cost to win them, or what you spent before they arrived. Those three things decide whether recurring revenue turns into a bank balance.
Four Numbers, One Company
Samsara's fiscal 2024 earnings release gives a clean view of the difference, because it reports all four measures side by side for two years:
| Measure | FY2023 (Jan 2023) | FY2024 (Feb 2024) |
|---|---|---|
| Ending ARR | 795.1 | 1,102.0 |
| Revenue | 652.5 | 937.4 |
| Net loss | (247.4) | (286.7) |
| Net cash from operations | (103.0) | (11.8) |
| Free cash flow | (136.3) | (22.8) |
ARR grew by about $307 million, or 39%, over the year. The net loss got larger by about $39 million. And free cash flow improved by about $113 million, from a $136 million outflow to a $23 million one. Three measures, three directions. I haven't broken out why from the filing, and I won't guess in detail, but the shape is clear: a loss that grows can sit alongside cash burn that shrinks, because the income statement includes costs that don't leave the bank and the cash flow statement includes timing that revenue doesn't show.
It cuts the other way too. Toast's 2025 results show $2,047 million of ARR, $342 million of net income and $608 million of free cash flow: more cash than profit. Neither company's ARR tells you which way the cash went.
ARR tells you what your contracts say you'll earn. Cash tells you whether you're still around to earn it.
A Worked Example: Same ARR, Three Bank Balances
A public company hides its mechanics in the aggregate. A small hypothetical doesn't. Every input is below so you can change it for your own business. This is an illustration, not a real company, and it's deliberately simple: no churn, and every invoice is paid exactly when the scenario says it is.
The inputs. The company starts the year with 50 customers, each paying $12,000 a year ($1,000 a month), and $300,000 in the bank. It signs 5 new customers every month for 12 months. Serving each customer costs $250 a month. Everything else, payroll, tools and overhead, is $85,000 a month. A sales commission of 10% of the first year's contract ($1,200) is paid at signing. ARR rises from $600,000 to $1,320,000, up 120%, in every scenario. What changes is only how customers pay:
| How customers pay | Cash at year-end | Change in cash | Cash runs out |
|---|---|---|---|
| Monthly billing | −$49,500 | −$349,500 | Month 9 |
| Annual, invoiced at signing, paid 60 days later | $170,500 | −$129,500 | Not within the year |
| Annual, paid upfront at signing | $280,500 | −$19,500 | Not within the year |
Same ARR at the start, same ARR at the end, same customers, same costs. The monthly-billing company is out of cash in month 9. The upfront-annual company ends the year essentially where it started. The $330,000 gap between them is not performance. It's timing: the annual-prepay company is collecting twelve months of revenue at once for the customers it signs, while the monthly company is fronting commissions and service costs and getting paid back slowly.
The middle row is the one most B2B companies live in. Annual contracts sound like prepayment, but if the invoice is net-60 and the customer's accounts-payable team takes its time, you've lent them the money for two months. That alone, in this model, costs about $110,000 of cash against the upfront case.
There's a trade-off I haven't priced in. Annual contracts are often sold at a discount. One billing vendor puts the typical range at 10% to 20% off list price. A discount lowers your ARR per customer, so you're trading some revenue for earlier cash. Whether that's worth it depends on how close to zero you are.
Why a Good Burn Multiple Doesn't Save You
Investors use the burn multiple, net burn divided by net new ARR, to judge how efficiently a company turns cash into growth. Commonly cited bands, attributed to David Sacks, call under 1.0x amazing, 1.0 to 1.5x great, 1.5 to 2.0x good, 2.0 to 3.0x suspect and over 3.0x bad. Those bands are quoted without a published source or sample in the material I found, so treat them as rules of thumb.
Run the example through them. Net new ARR is $720,000. The monthly-billing company burned $349,500, a multiple of about 0.49. The net-60 company burned $129,500, about 0.18. The upfront company burned $19,500, about 0.03. All three score "amazing," and one of them runs out of cash in month nine.
That isn't a flaw in the ratio. It's a reminder of what it measures. Efficiency tells you how much cash a dollar of new ARR costs. It doesn't tell you how much cash you have. A company can be very efficient and still be a few months from zero if it started thin. Samsara's numbers make the same point from the other side: growth and cash moved independently of each other.
What to Track Next to ARR
Runway in months. Cash divided by the current monthly net burn, recalculated every month. It's the only number that answers "how long do we have."
Cash collected against ARR added. For each period, how much cash did customers actually pay compared with the ARR you booked? A widening gap means collections are lagging contracts.
Billing mix and payment terms. What share of customers pay monthly, annually upfront or annually on terms, and how long do invoices really take to be paid? The three rows in the table above are three versions of the same company.
Upfront costs. Commissions, onboarding labor and implementation spend are cash out the door before the first full month of revenue. If they're paid at signing, your fastest-growing months are your cash-hungriest.
A written definition of ARR. What is in it, what isn't, and when a contract starts counting. When you share ARR with investors or buyers (the next issue covers what diligence looks like), the first question will be how you define it.
Where These Numbers Come From
The Samsara and Toast figures come from the companies' own earnings releases. The worked example is my own hypothetical, built so every number can be re-derived. The burn multiple bands and the discount range come from vendor content, which I've marked as rules of thumb. Figures checked November 2026.
- Samsara: fourth-quarter fiscal 2024 earnings release (8-K). Source of ARR, revenue, net loss, cash flow figures and Samsara's ARR definition.
- Toast: fourth-quarter and full-year 2025 results. Source of ARR, net income and free cash flow.
- Ordway: SaaS financial health, GAAP revenue vs. ARR. Vendor blog; source for "no universal accounting standard" and the annual-discount range.
- Runway: burn multiple benchmarks. Vendor blog; the burn multiple formula and bands, attributed without a primary source.
Build a 12-Month Cash Walk With Your Real Billing Terms
Open a spreadsheet and lay out the next twelve months, one row per month. Start with today's cash. Add what customers will actually pay, using your real mix of monthly, annual and invoiced-on-terms customers and your real payment delays, not your contract dates. Subtract payroll, tools, commissions at signing and cost to serve. Find the month the balance would hit zero. Then change one thing, such as moving half your new customers to annual prepay, and see how far the date moves. If the answer surprises you, ARR was hiding something.
