How to build a SaaS financial model: a founder step-by-step guide

A SaaS financial model is a driver-based forecast that builds your future revenue, costs, and cash from operating assumptions instead of a flat growth percentage. At its core it models revenue as an MRR movement engine, starting MRR plus new, expansion, and reactivation, minus contraction and churn, layered on cohort retention curves. It then subtracts COGS and operating expenses to project profit, and rolls that into a cash and runway view so the model answers real questions about hiring, spend, and fundraising.

Most “how to build a SaaS financial model” guides teach you to type “grow 20% a month” into a cell and let the rows multiply. That produces a chart, not a model. This guide builds revenue the way real analytics tools compute it, then connects the output to the two numbers that actually run your company: CAC payback and runway.

What a SaaS financial model is, and what makes it driver-based

A SaaS financial model projects revenue, costs, and cash, and a good one is driver-based: you forecast the inputs (leads, conversion, ARPU, retention, expansion) and let revenue fall out of the math. You never type a target ARR number. The number is an output, not an assumption.

A flat-growth model says “we’ll grow 8% a month” and compounds that line forever. It has no opinion on why growth happens, so it can’t tell you what breaks when conversion dips or churn rises.

A driver-based model says “we generate 400 leads, convert 4%, at $250 average new MRR, and lose 2% of revenue a month.” Now every output traces to a lever you control, which is what makes the model survive a fundraise. Investors don’t fund the ARR line. They pressure-test the assumptions underneath it.

What a SaaS financial model should include

A complete model has five layers, and a full version links three statements together. The five layers:

Revenue build from MRR movements and cohort retention.
Cost structure split into COGS and operating expenses.
Unit economics: CAC, CAC payback, LTV:CAC, magic number.
Cash and runway, including deferred revenue and burn.
Scenarios: at least a base, upside, and downside case.

A three-statement SaaS model then links the income statement (P&L), balance sheet, and cash flow statement so cash burn projects accurately rather than being eyeballed. The three SaaS-specific balance-sheet items founders most often miss are capitalized software (ASC 350-40), capitalized commissions (ASC 340-40), and deferred revenue. If you bill annually, that last one is the difference between your model and reality.

Here’s the core revenue formula everything hangs on:

Ending MRR = Starting MRR + New + Expansion + Reactivation minus Contraction minus Churn

Multiply ending MRR by 12 for ARR. That single waterfall, repeated month over month, is the engine. The rest of this guide builds it and the layers around it.

Bottom-up vs top-down forecasting

Use bottom-up as your engine and top-down as a sanity check. They should land in the same ballpark; if they don’t, one of them is lying.

Top-down starts from your total addressable market, applies a market-share assumption, and works backward to revenue. It’s fast, and it’s where founders wish-cast, because “1% of a $40B market” sounds inevitable and means nothing operationally.

Bottom-up aggregates real drivers, leads times conversion times ARPU, reps times quota, retention curves, into total revenue. It’s slower and far more defensible. When an investor asks “why 4% conversion,” you have an answer. When they ask “why 1% market share,” you don’t.

	Top-down	Bottom-up
Starts from	TAM / SAM and market share	Leads, conversion, ARPU, retention
Best for	Framing market size	Planning, hiring, fundraising
Risk	Wish-casting	Spreadsheet sprawl
Use it as	A sanity check	The primary engine

Build bottom-up. Glance at the total addressable market to confirm you’re not modeling more revenue than the market can hold.

Step 1: build revenue from the MRR movement engine

Forecast revenue as a monthly MRR waterfall, the same five-driver breakdown Baremetrics, ChartMogul, and Maxio use. Starting MRR plus three additions, minus two subtractions, equals ending MRR.

The five drivers:

New MRR: recurring revenue from brand-new customers this month.
Expansion MRR: upgrades, seat adds, and usage growth from existing customers.
Reactivation MRR: revenue from previously churned customers who return.
Contraction MRR: downgrades and seat reductions from customers who stay.
Churned MRR: revenue lost when customers cancel entirely.

Net new MRR is the sum: New + Expansion + Reactivation minus Contraction minus Churn. Add it to starting MRR and you have next month’s starting point. Repeat.

A worked month. Start at $100,000 MRR. Add $12,000 new, $4,000 expansion, $500 reactivation. Lose $1,500 to contraction and $3,000 to churn.

Net new MRR = 12,000 + 4,000 + 500 minus 1,500 minus 3,000 = $12,000. Ending MRR = $100,000 + $12,000 = $112,000. ARR = $112,000 times 12 = $1.344M.

The reason this beats a flat percent: those five numbers move independently. Churn can spike while new MRR holds, and one blended “growth rate” hides exactly that signal. The SaaS revenue forecasting guide goes deeper, and the MRR forecaster runs the waterfall for you.

Step 2: drive new MRR bottom-up

Build new MRR from the funnel, not a guess. The formula:

New MRR = Leads x Lead-to-customer conversion rate x Average new MRR per customer

Then segment by plan, because a $2,000 enterprise deal and a $40 self-serve signup behave nothing alike and shouldn’t share one conversion rate.

Worked example. Marketing delivers 500 leads a month. Sales converts 5%, so 25 new customers. Average new MRR per customer is $300.

New MRR = 500 x 0.05 x $300 = $7,500 a month.

Now you have real levers. Want to double new MRR? You either double leads, lift conversion from 5% to 10%, or push ARPU from $300 to $600. The model tells you which is plausible and what each costs. A flat “grow 8% a month” tells you nothing.

For self-serve motions, swap “lead conversion” for trial-conversion rate and drive new MRR off signups times trial-to-paid percentage.

Step 3: model expansion and churn with cohort retention curves

Replace a single blended churn rate with cohort retention curves, because that’s the difference between “we’ll grow 30%” and actual math. A cohort is a group of customers who signed up the same month. You track what each cohort does over time, then connect those curves to the revenue bridge.

Two retention numbers anchor the curves:

Net Revenue Retention (NRR) = (Starting MRR + Expansion + Reactivation minus Contraction minus Churn) / Starting MRR

Gross Revenue Retention (GRR) = (Starting MRR minus Contraction minus Churn) / Starting MRR

NRR includes expansion and can exceed 100%. GRR excludes it and caps at 100%, so it shows pure leakage. Model both per cohort.

Here’s why the cohort view changes decisions. Say a cohort loses 5% of revenue to churn but expansion pulls its NRR to 102%. A blended rate buries that. The cohort tells you it’s actually growing in place, so your real problem is acquisition, not retention. Different problem, different spend.

In 2025, median NRR sat at 101% and median GRR at 88% (Benchmarkit, 2,000+ companies); enterprise NRR runs near 118%, SMB closer to 97%. Use those to pressure-test your curves, not replace them. The cohort analysis guide and the net revenue retention calculator help you build and check them.

Step 4: layer in COGS and gross margin

COGS is the direct cost of delivering your service, and for SaaS that’s a specific, short list: hosting and infrastructure, customer support, payment processing fees, and the cost of third-party software baked into your service.

Gross Margin = (Revenue minus COGS) / Revenue

What does not belong in COGS: sales, marketing, R&D, and general overhead. Those are operating expenses (Step 5). Founders who dump everything into COGS understate gross margin and spook investors.

Median SaaS gross margin in 2025 was 77% on total revenue and 81% subscription-only (Benchmarkit). Treat 75% as the floor and 80%-plus as the target; AI-heavy products are pushing early-stage margins down as inference costs bite, so model your hosting line honestly.

Worked: $1.344M ARR at 80% gross margin leaves $1.075M in gross profit to fund everything else. The SaaS gross margin calculator does the split.

Step 5: build the operating expense plan

Operating expenses are everything below gross profit, and in SaaS they’re mostly headcount. Build opex from a hiring plan, not a percentage of revenue. Three buckets:

S&M (sales and marketing): reps, ad spend, marketing tools, commissions.
R&D (research and development): engineering, product, design salaries.
G&A (general and administrative): finance, legal, ops, rent, software.

Model each role as a row with a start month and a fully loaded cost (salary plus roughly 20-30% for taxes, benefits, equipment). When you add a rep in month 4, the cost shows up in month 4, not spread evenly. That’s what makes the burn line real.

A useful check sits here: revenue per employee, median around $130,000 and $175,000-plus considered strong (SaaS Capital 2025). If your plan implies $60,000 per head, it’s too heavy. The ARR-per-employee calculator sizes this fast.

Step 6: pressure-test unit economics

Before you trust the model, run it through four unit-economics checks. CAC payback is the most important, because it answers “how long until an acquired customer pays back what it cost to win them,” and that gates how fast you can spend.

CAC = Total sales and marketing spend in a period / new customers acquired

CAC Payback (months) = CAC / (Average new MRR per customer x Gross margin %)

LTV = (Average MRR per customer x Gross margin %) / Monthly churn rate; LTV:CAC = LTV / CAC

Magic Number = Net new ARR added in a quarter / prior-quarter S&M spend

Worked CAC payback. You spend $30,000 on S&M and win 25 customers, so CAC is $1,200. New MRR per customer is $300, gross margin 80%, so each customer contributes $240 of gross margin a month.

CAC payback = 1,200 / 240 = 5 months. That’s excellent. Anything under 12 months is top quartile; the median has drifted to 15-20 months.

Metric	Minimum	Strong	Top quartile	2025 median
CAC payback	under 18 mo	under 15 mo	under 12 mo	15-20 mo
LTV:CAC	3:1	4:1	4:1 to 6:1	~3.2:1
Magic number	0.75	1.0	above 2.0	~0.9
New CAC ratio	under $2.50	under $1.50	under $1.00	$2.00 per $1 ARR
Gross margin	75%	80%	85%+	77%

Sources: Benchmarkit 2025, KeyBanc 2024-25, Bessemer 2026 State of the Cloud, SaaS Capital 2025.

The median company now spends $2.00 to acquire $1.00 of new ARR, up roughly 14% since 2023. That’s why CAC payback and LTV:CAC are the make-or-break outputs of the model, not afterthoughts. The CAC payback calculator and LTV:CAC ratio calculator let you stress these directly.

Step 7: roll it into cash and runway

The model only earns its keep when it tells you when you run out of money. Two burn numbers, one runway formula:

Net Burn = monthly cash outflows minus cash collections; Gross Burn = total cash outflows

Cash Runway (months) = Current cash balance / Net Burn rate

Runway runs off net burn, not gross. Worked: $1.2M in the bank, $200,000 monthly outflows, $120,000 monthly collections. Net burn = $80,000. Runway = 1,200,000 / 80,000 = 15 months.

Now the SaaS-specific twist most first models get wrong: cash and recognized revenue are not the same thing. If you bill annually upfront, cash arrives in month one but revenue recognizes over 12 months. The gap is deferred revenue, and its month-over-month change is a positive working-capital adjustment in operating cash flow.

That’s why a growing SaaS company can post a net loss on the P&L and still show positive operating cash flow. A model that forecasts burn off recognized revenue alone will overstate it and scare you into raising early. Model collections, not just revenue. The burn rate and runway calculator and the deferred revenue calculator handle both sides.

This is where the three-statement linkage pays off: deferred revenue lives on the balance sheet, flows through cash flow, and reconciles to the P&L. Link them and the burn line stops being a guess.

Step 8: add scenarios and stress test

A single forecast is a prediction; three scenarios are a plan. Build base, upside, and downside by flexing the drivers you already isolated, not by nudging one growth percent.

Downside: conversion drops 30%, churn rises 50%, a key hire slips. Does runway still clear your next raise?
Base: your honest expected case.
Upside: a channel works, NRR climbs. Can you hire fast enough to keep up?

Because the model is driver-based, scenarios are just three columns of assumptions feeding the same engine. That’s the whole payoff of building it this way.

Use the Rule of 40 as a final sanity check, never an input:

Rule of 40 = Revenue growth rate % + profit margin %; target 40 or above

Use EBITDA margin or free-cash-flow margin for the profit half, and state which. Only about 20% of public SaaS clear 40, and the public median score sat near 28% as of Q4 2025 (Aventis / SaaS Capital). If your base case implies a Rule of 40 score of 75, your assumptions are fantasy. The Rule of 40 calculator gives you the number in seconds.

Common mistakes founders make in their first model

Four errors show up in nearly every first attempt, and each one quietly breaks the model:

Flat growth percent. “Grow 8% a month” compounds nicely and predicts nothing. Build the MRR waterfall instead.
Ignoring contraction. Founders model churn but forget downgrades. Contraction is real leakage and it shows up in NRR.
Confusing recognized revenue with cash. Annual billing breaks any model that forecasts burn off recognized revenue. Model collections.
No cohort view. A blended retention rate hides the difference between a leaky product and a slow acquisition engine.

A fifth, subtler one: treating the model as a fundraising artifact you build once. The point is to re-run it monthly against actuals. A model you never reconcile to reality is just a nicer-looking guess.

2025-2026 SaaS benchmark cheat sheet

Plug these medians in to pressure-test your assumptions, then replace them with your own cohort data as it accumulates.

Metric	2025-2026 median	Target / top quartile	Source
ARR growth (private)	26%	50%+ top quartile	Benchmarkit 2025
Net revenue retention	101%	110%+	Benchmarkit 2025
Gross revenue retention	88%	90-92%+	Benchmarkit 2025
Gross margin	77%	80-85%+	Benchmarkit 2025
CAC payback	15-20 mo	under 12 mo	KeyBanc 2024-25
New CAC ratio	$2.00 per $1 ARR	under $1.00	SaaS Capital 2025
LTV:CAC	~3.2:1	4:1 to 6:1	Bessemer 2026
Magic number	~0.9	1.0+	Benchmarkit 2024
Rule of 40 (public)	~28% score	40+	SaaS Capital 2025
Revenue per employee	~$130K	$175K+	SaaS Capital 2025
Runway before raise	6-12 mo	12-18 mo	2024 VC survey

More context on where you sit by stage lives in the state of SaaS metrics 2026 breakdown and on the benchmarks page.

How Mowt feeds your model with real numbers

Your model is only as good as its inputs, and the assumptions that matter most, retention curves, expansion rates, churn, ARPU, CAC, are exactly the ones founders guess at. Mowt computes them from your connected Stripe account, so the model starts from truth instead of a hopeful round number.

We don’t theorize about the MRR waterfall. We compute it live across $1B+ in tracked revenue for 500+ companies at 99.9% uptime, the same five-driver engine this guide describes. Pull your real MRR movements and cohort retention into the model, calibrate every assumption against what your customers actually did, and your base case stops being a wish.

Download: SaaS financial model template

Want the structure without building it from scratch? Use the MRR forecaster to run the revenue waterfall, the burn rate and runway calculator for the cash view, and the SaaS break-even calculator to find the month the model turns cash-positive. Together they give you a working SaaS financial model template you can populate with your own numbers in an afternoon.

FAQ

What is a SaaS financial model?

It’s a driver-based forecast of a SaaS company’s revenue, costs, and cash, built from operating assumptions rather than a flat growth rate. Revenue is built from the MRR movement engine (new, expansion, reactivation, contraction, churn) plus cohort retention; costs split into COGS and opex; and the result rolls into a cash and runway view that drives decisions on hiring, spend, and fundraising.

How do you forecast SaaS revenue in a model?

Build it as a monthly MRR waterfall: Starting MRR plus New plus Expansion plus Reactivation minus Contraction minus Churn equals Ending MRR, then multiply by 12 for ARR. Drive new MRR bottom-up from leads times conversion times average new MRR, and drive expansion and churn from cohort retention curves rather than one blended percentage, so every number traces to an operating assumption.

Should I use bottom-up or top-down forecasting?

Use bottom-up as the primary engine and top-down as a sanity check. Bottom-up aggregates real drivers (leads, conversion, reps, ARPU, retention) and is far more defensible for planning and fundraising; top-down starts from market size and is easy to wish-cast. The two should land in the same ballpark.

How do you calculate runway in a SaaS model?

Runway in months equals current cash balance divided by net burn rate, where net burn is monthly cash outflows minus cash collections. Model cash collections, not just recognized revenue, so annual upfront billing and the deferred-revenue swing are captured. Most VCs advise keeping 6 to 12-plus months of runway before the next raise.

What benchmarks should I plug into a SaaS model in 2025-2026?

Anchor to current medians: gross margin about 77% (target 80%-plus), NRR about 101% (target 110%-plus), GRR about 88-90%, CAC payback 15-20 months (target under 12), LTV:CAC 3:1 minimum, and magic number about 0.9 (target 1.0-plus). Use the Rule of 40 as a sanity check that only about 20% of public SaaS clear, and replace these medians with your own cohort data over time.