Most SaaS revenue forecasts I review fail the same way. Someone takes this year's ARR, applies a growth rate that felt right in the board meeting, and extends the line to the right. It looks tidy. It also cannot answer the only questions that matter: what has to be true about pipeline, hiring, and retention for that line to happen, and what breaks first if it does not.
A forecast investors trust, and one you can actually run the company on, is built from drivers. This guide covers the three methods that do that, how to reconcile them, how to turn the result into revenue and cash, and the benchmarks worth checking your assumptions against.
Forecast ARR first, then revenue
Start by separating three numbers that get blurred together.
ARR is the annualized value of the recurring contracts in force today. It is the operating metric: it moves the moment a contract is signed, expanded, or cancelled.
Revenue is what you recognize under ASC 606 as you deliver the service, usually ratably across the contract term. It trails ARR.
Bookings and cash are what customers commit to and pay. An annual prepaid contract brings in twelve months of cash on day one and recognizes it over the year.
The order matters. Forecast ARR from its drivers, derive revenue from the ARR forecast, and derive cash from billing terms. If you forecast revenue directly, you lose the ability to see why it moved. For the full distinction between the four, see bookings vs. billings vs. revenue vs. ARR.
Method 1: The ARR bridge
The ARR bridge is the backbone of every good SaaS forecast. For any period:
Ending ARR = Starting ARR + New ARR + Expansion ARR - Contraction ARR - Churned ARR
Each line has a different driver and, usually, a different owner. New ARR belongs to sales and marketing. Expansion belongs to customer success and pricing. Contraction and churn belong to product and support. When the total misses, the bridge tells you which of them missed and by how much.
Here is one quarter for an illustrative Series A company starting at $3.0M ARR:
| Line | Q1 | Driver |
|---|---|---|
| Starting ARR | $3,000,000 | Closing ARR last quarter |
| + New ARR | $450,000 | Pipeline and rep capacity (Method 3) |
| + Expansion | $120,000 | Seat growth and upsell on the existing base |
| - Contraction | ($45,000) | Downgrades |
| - Churn | ($90,000) | Logo cancellations |
| Ending ARR | $3,435,000 |
Net new ARR for the quarter is $435,000, which is also the denominator of your burn multiple. Build the bridge monthly for the next 18 to 24 months and roll it up to quarters for the board.
The discipline is simple and rarely followed: never type a number into the ending ARR row. Every figure comes from the lines above it, and every line above it comes from an assumption you can defend in a sentence.
Method 2: Cohort retention for the existing base
The existing base is the most predictable part of the forecast, and the part founders most often get wrong, because they assume every customer behaves like the average customer.
Instead, group customers by the quarter they started and track what share of their starting ARR remains in each later period. Two ratios matter:
- Gross revenue retention (GRR): the share of starting ARR still there after churn and contraction, ignoring expansion. It can never exceed 100 percent.
- Net revenue retention (NRR): the same measure with expansion added back. It can exceed 100 percent.
Apply each cohort's observed curve to its remaining ARR and you have a defensible forecast for the base without touching new sales at all. Young cohorts usually churn faster than mature ones, so a single blended rate overstates retention for recent customers and understates it for older ones. The more your customer mix has shifted, toward smaller accounts, a new segment, or a new pricing plan, the more that blend misleads you.
Benchmarks are a sanity check, not an input. Benchmarkit's 2025 research put median NRR for private B2B SaaS at 101 percent and gross retention around 88 percent. SaaS Capital's 2025 retention data found a median NRR of 102 percent for companies with $25,000 to $50,000 contract values, with the top quartile at 111 percent and the bottom quartile at 97 percent.
Read together, the implication is uncomfortable. A typical private SaaS company loses roughly 12 percent of its base each year to churn and contraction, and expansion only just fills the hole. If your forecast assumes 120 percent NRR, you need cohorts that show it, not a slide that says best-in-class companies achieve it.
Method 3: Pipeline and capacity for new ARR
New ARR is the least predictable line, so forecast it two ways and trust the lower one until you have evidence for the higher.
The pipeline view. New ARR in a period equals qualified pipeline created in earlier periods, multiplied by your historical win rate, shifted forward by your sales cycle. If you create $1.5M of qualified pipeline in January, win 25 percent of it, and your cycle runs about 90 days, most of that $375,000 lands in April, not January.
The capacity view. New ARR equals the number of quota-carrying reps, multiplied by the share of quota they actually attain, adjusted for ramp. A rep hired in March with a $600,000 annual quota who takes six months to ramp contributes very little before September. Forecasts that add reps and their full productivity in the same month are the most common error I see in Series A models.
Then check one against the other. If the capacity view says $500,000 of new ARR next quarter but your pipeline coverage is well under three times that number, the capacity view is optimistic. The gap between the two views is not noise to average away. It is the conversation to have with your head of sales before the board has it with you.
For product-led companies, replace reps with the funnel: signups, activation, conversion to paid, and the average starting plan, each measured from your own data rather than a benchmark from a company with a different price point.
Which method carries the weight at your stage
The methods do not change by stage. What changes is how much of the forecast each one can support.
- Pre-seed and seed, under roughly $1M ARR. Cohorts are too young to trust and pipeline history is thin. Forecast from the funnel or a short list of named deals, keep the monthly horizon to 12 to 18 months, and show your assumptions plainly. Investors at this stage are underwriting your understanding of the drivers more than the number itself.
- Series A, roughly $1M to $5M ARR. You have enough history for a real bridge. Cohort retention should replace blended churn, and the capacity view starts to matter because the plan almost always includes hiring sales.
- Series B and beyond. Pipeline and capacity carry most of new ARR, cohort curves are stable enough to drive the base, and the board will expect the forecast to reconcile to recognized revenue and cash every month.
Reconcile top-down against bottom-up
A bottom-up forecast built from the three methods is only half the job. Check it against the top-down question: what growth rate does it imply, and is that plausible for a company like yours?
SaaS Capital's 2026 survey of more than 1,000 private B2B SaaS companies put median ARR growth at 22 percent, with equity-backed companies at 25 percent and bootstrapped companies at 20 percent. Venture-backed companies heading into their next round are expected to grow well above that median, and at Series A the bar is set by the ARR benchmarks for that stage. A bottom-up model that produces 200 percent growth is not wrong by definition. It does mean every assumption behind it needs evidence: pipeline already created, reps already hired, retention already observed.
When the two views disagree, fix the drivers, not the output. Never scale a bottom-up forecast to hit a top-down target. That is how a model loses the one property that made it useful, which is that every number traces to a cause.
From ARR to revenue and cash
Once the ARR forecast holds, convert it.
Revenue. Under ASC 606, subscription revenue is typically recognized ratably across the service period. A $120,000 annual contract that starts July 1 adds $120,000 of ARR on July 1 but only $60,000 of revenue in that calendar year. Implementation fees, usage charges, and contracts with several performance obligations each need their own treatment, and this is where a forecast built on ARR alone quietly overstates recognized revenue.
Cash. Billing terms drive cash, not ARR. Annual upfront billing pulls cash forward and builds deferred revenue; monthly billing does the opposite. Model collections separately, including how long customers actually take to pay, and feed the result into a 13-week cash forecast for the near term.
The three numbers should reconcile every month: the change in ARR, recognized revenue, and the change in deferred revenue alongside collections. If they do not, one of the models has a bug, and it is better to find it before an investor does.
Usage-based and hybrid pricing
Usage pricing breaks the clean ARR definition. Committed minimums behave like subscription ARR. Consumption above the commitment does not, because it can fall as quickly as it rose.
Forecast the two separately. Treat contracted minimums as ARR in the bridge, and forecast usage revenue from active customers, usage per customer, and price per unit, with its own seasonality and a wider downside case. A customer with a $50,000 annual commitment that has been consuming about $6,000 a month above it should carry $50,000 in the bridge and roughly $72,000 a year of usage in a separate line, and that usage line is the first thing to cut in the downside case if consumption has been volatile. The mechanics are covered in more depth in forecasting revenue and ARR when pricing is usage-based.
Build scenarios, not a single line
A single forecast invites a single argument. Build three, and make each one a change in drivers rather than a percentage haircut on the output:
- Base: current win rates, observed cohort retention, and the hiring plan you have actually approved.
- Downside: win rates a few points lower, sales cycles 30 days longer, two planned sales hires delayed by a quarter, and churn at the level of the worst cohort you have seen.
- Upside: pipeline you are confident in closing, plus expansion from a pricing change you have already tested with real customers.
Then tie each scenario to burn and runway. The point of a SaaS revenue forecast is not the revenue line. It is knowing, before the board meeting, how many months of runway the downside leaves you and which decision you would make if the numbers started trending toward it.
Run the forecast monthly, against a locked plan
A forecast is only useful if it changes when reality does. The cadence that works:
- Lock the annual plan. Once the board approves it, the plan does not move. It is the yardstick everything else is measured against.
- Roll the forecast monthly. Replace each month's forecast with actuals as it closes, and update the remaining months from current drivers: pipeline actually created, reps actually hired, retention as observed in the latest cohorts.
- Explain variance by bridge line. Missed new ARR, missed expansion, and higher churn are three different problems with three different owners, and the bridge shows which one you have in a single table.
- Re-plan on a trigger, not a calendar. If the rolling forecast falls far enough below plan to put runway or the next raise at risk, rebuild the plan rather than waiting for the next budget cycle. Agree the trigger with your board in advance, for example ending ARR more than 15 percent below plan or runway falling under 12 months.
Two habits make the monthly roll trustworthy. First, keep every assumption in one place rather than scattered across tabs, so a change in win rate flows through new ARR, revenue, cash, and runway at once. Second, record why each assumption changed and when. Six months later, that log of what you learned about your own business is worth more than any single version of the forecast.
The mistakes that make investors discount a forecast
- A growth rate on top-line ARR. No drivers, so no way to diagnose a miss.
- New reps productive on day one. Ramp and sales cycle both push new ARR out by quarters, not weeks.
- Blended retention. Recent cohorts almost always churn faster than your historical average.
- NRR assumed, not observed. Above roughly 110 percent without cohort evidence, expect a haircut.
- Revenue equals ARR. Ratable recognition and implementation fees break that assumption.
- Cash equals revenue. Billing terms and collections decide runway.
- One scenario. It tells investors you have not thought about what happens when you miss.
What a forecast investors trust looks like
When I prepare a SaaS revenue forecast for a board or a data room, it has the same shape every time:
- A monthly ARR bridge for 24 months, with every line tied to a named driver.
- Cohort retention tables that the retention assumptions come from.
- A pipeline build and a capacity build for new ARR, reconciled to each other.
- A revenue schedule under ASC 606 and a cash schedule built from billing terms.
- Base, downside, and upside scenarios, each connected to burn and runway.
- A one-page summary of the five assumptions that move the result most.
That last page is the one investors read first, and often the only one they read closely.
If you want a quick first pass before building the full model, the free SaaS revenue forecast tool projects MRR and ARR from growth and churn in a couple of minutes. Treat it as a starting point, not a substitute for the bridge. For the acquisition side of the model, the unit economics calculator and our CAC payback benchmarks by stage show whether the growth you are forecasting is affordable. When you want the full driver-based model built and maintained alongside your books, that is part of the fractional CFO work on StartupCFO's Growth plan.
Sources
- Benchmarkit, 2025 SaaS Performance Metrics: median net revenue retention and gross revenue retention for private B2B SaaS
- SaaS Capital, 2026 Private B2B SaaS Company Growth Rate Benchmarks: median growth across more than 1,000 companies, bootstrapped and equity-backed
- SaaS Capital, What Is a Good Retention Rate for a Private SaaS Company: net revenue retention by contract value
- FASB ASC 606, Revenue from Contracts with Customers
- Worked examples in this guide are illustrative, not client data.
