A venture-grade financial model has four parts: a revenue build with unit economics, a cost build including headcount and COGS, a three-statement output (P&L, cash flow, balance sheet), and scenario toggles. Most founder-built models fail in due diligence because of weak assumptions, circular references, and missing scenario logic. This guide explains how to build each part in order.
A financial model is a spreadsheet that projects your company's revenue, costs, and cash position over time. For a venture-backed startup, your model is the document an investor reads to decide whether to write the cheque. This pillar guide covers everything we have written about financial modelling, from the core three-statement structure through scenario analysis to cap table integration.
A financial model is a forward-looking spreadsheet that translates your business assumptions into projected financials. For a startup raising venture capital, the model has three jobs: prove that the business can reach a valuation that justifies the round, show the investor how you will spend their money over 18 to 24 months, and give the board a tool to monitor actuals against plan after the round closes.
Every venture-grade model is built around three interlinked financial statements: the profit and loss (P&L), the cash flow statement, and the balance sheet. The P&L shows revenue and expense by period. The cash flow statement shows operating, investing, and financing cash movements. The balance sheet shows the snapshot of assets, liabilities, and equity at the end of each period. The three must balance every period and the balance sheet must foot to zero.
The revenue build is the most scrutinised part of any model. For SaaS, the core formula is ARPA × customers, with churn and expansion layered on top. For marketplaces, it is GMV × take rate. For e-commerce, it is traffic × conversion × AOV. Every revenue build should start from drivers that the founder can actually influence, not from top-down market-share assumptions.
Costs break into three buckets: cost of goods sold (COGS), operating expenses (opex), and capital expenditure (capex). For most early-stage software companies, headcount is 70 to 80 percent of opex. Build headcount as a plan with names, titles, start dates, locations, and fully-loaded costs including salary, employer taxes, benefits, equipment, and software. Overlay non-headcount opex as a separate schedule.
A model without scenarios is not a model, it is a wish. Investors expect to see at least three cases: base, upside, and downside. Downside should show what happens if your lead metric is 30 percent worse than plan. Upside should show the path to 50 percent outperformance. Sensitivity tables should isolate the two or three variables your business is most exposed to.
Your model connects to your cap table through the financing tab. The financing tab schedules each round: amount, pre-money, post-money, option pool top-up, and dilution to founders and existing investors. Waterfall analysis shows what each shareholder receives in a sale at various exit values, critical for investor conversations and for founder peace of mind.
Five failure modes we see repeatedly: circular references that blow up when you toggle scenarios, hardcoded numbers inside formula cells, mismatched units (months versus years), missing balance sheet roll-forward that breaks the cash flow statement, and revenue assumptions that do not reconcile to the cost build (you cannot grow revenue 300 percent without growing headcount). Every venture-grade model audits against these five.
Common questions founders ask about this topic.
Use Google Sheets or Excel for the first two years. Spreadsheets are faster, cheaper, and easier to share with investors than dedicated software. Only move to paid tools once your model has more than 20 tabs or three scenarios that need version control.
Model 60 months (five years) at monthly granularity. Investors expect monthly detail for the first 18 months, then quarterly or annual views for the remainder. Longer horizons become guesswork and erode credibility.
The five most common errors are: hard-coded revenue instead of a driver build, missing cost-of-revenue scaling with volume, payroll that skips taxes and benefits, a cash balance that does not reconcile to the P&L, and overly aggressive growth assumptions with no supporting evidence.
Three: base, bear, and bull. The base case is what you believe. The bear case shows the minimum funding you need to survive. The bull case shows what happens if your best assumptions land. More than three dilutes the story; fewer looks naive.
Yes. Integrate a cap table tab that links to the funding rounds on the P&L so every raise automatically updates ownership, dilution, and preference stacks. Standalone cap tables drift out of sync within weeks.
A financial model is a belief engine. It turns your view of the business into numbers that other people can check, stress, and argue with. Founders who build their own models raise faster because they can answer diligence questions in real time. The path from blank spreadsheet to investor-ready model runs through five disciplined steps, each taking about two to three days of focused work.
Every SaaS model starts with a single assumption about how revenue scales. The most common drivers are customers acquired per month, leads per channel, website visitors with a funnel, or seats per customer multiplied by price per seat. Pick the driver that matches how your team actually sells and document the conversion rates you used. Hard-coded revenue curves hide reality and collapse under the first investor question.
Unit economics answer the question of whether each customer makes money. The three numbers that matter are customer acquisition cost, average revenue per customer, and gross margin. Investors divide lifetime value by acquisition cost to get an LTV to CAC ratio. A ratio of three or higher shows you are building a durable business; a ratio below two shows you are buying growth. Payback period in months is the companion metric and should land under eighteen months for venture-backed software.
Costs scale with headcount, cloud infrastructure, software tools, sales commissions, and marketing spend. Build each cost category as a function of revenue or headcount, not as a flat dollar amount. When revenue doubles, headcount should lag by six months and hosting should scale with a known cost curve. Models that use flat cost lines are easy to build but impossible to defend.
A three-statement model has a profit and loss statement, a cash flow statement, and a balance sheet. The P&L shows whether the business is profitable. The cash flow shows how much cash the business consumes or produces. The balance sheet shows what the business owns and owes at a point in time. Each statement pulls from the others and must reconcile. When they do not reconcile, there is a formula error somewhere upstream.
A single-case model is a marketing document, not a financial model. Build a base case, a bear case, and a bull case with different growth rates, retention assumptions, and funding timelines. Investors will ask what happens if the bear case lands, and you should be able to click a toggle and show them. Scenarios that differ by more than forty per cent on any single input signal overconfidence in your forecasts.
During diligence investors rarely ask about the output numbers; they ask about the assumptions that feed the outputs. Expect questions about conversion rates, churn assumptions, pricing power, competitive response, and headcount plans. Keep an assumption log beside the model with the source of every input, the date it was last updated, and the confidence level. When diligence questions land, the log makes the answer obvious.
Raise Ready offers a free financial model builder that ships with the revenue driver, unit economics, cost stack, and scenario toggles pre-wired. The Excel file and Google Sheets version are identical. Start with the template, replace the sample inputs with your own assumptions, and you will have an investor-ready model in two weeks of part-time work. Paid tools exist but rarely pay for themselves until your business passes five million in annual recurring revenue.
A financial model is only as reliable as the language used to describe its inputs. Founders and investors often use the same words for different concepts, and a glossary inside the model itself removes ninety per cent of the diligence friction. The terms below are the ones most frequently miscategorised in early-stage models, and every founder should know how to compute each one from first principles without opening a browser.
Recurring revenue is the subscription or contract income that a customer pays on a predictable schedule, usually monthly or annually. Non-recurring revenue covers setup fees, professional services, hardware sales, and anything that requires the customer to sign a new contract each time. Investors value recurring revenue at three to ten times the multiple of non-recurring revenue because it compounds across the life of the customer. Blended revenue mixes the two and obscures the underlying engine, which is why any model that fails to split them invites a diligence deep-dive.
Cost of revenue is the cost that scales directly with each sale: hosting, payment processing, direct support, and customer-specific infrastructure. Operating expense is everything else: salaries, marketing, rent, software, and general overhead. The difference matters because gross margin (revenue minus cost of revenue) is the single most-used lens for benchmarking a software business against peers. Misclassifying operating expense as cost of revenue inflates gross margin and triggers immediate distrust in the model. Misclassifying cost of revenue as operating expense inflates contribution margin and hides the scaling problem.
Salary is not the full cost of an employee. The fully loaded cost includes taxes, benefits, equipment, software, rent allocated per seat, training, and a small buffer for attrition replacement. The rule of thumb for a UK or US startup is to multiply base salary by 1.3 to 1.4 to reach the true monthly cost, and by 1.5 for executives who need travel and large software budgets. Models that use the base salary directly understate burn by fifteen to twenty per cent and lead to running out of cash a quarter earlier than planned.
Runway is the number of months of operation the business can fund before cash runs out at current burn. Net burn is the monthly cash consumed after revenue collection. Default-alive is the state where current growth and current burn, with no new funding, would produce profitability before the cash runs out. The default-alive calculation is the most important single number in a model for founders raising in a tight market, because it tells you exactly how much runway extension you need to negotiate from the next round without cutting headcount.
Pre-money valuation is the company value before new investment lands. Post-money valuation is pre-money plus the new investment amount. The dilution per round equals the new investment divided by the post-money valuation. Founders frequently understate dilution by forgetting option pool top-ups that preferred shareholders require before closing, and by ignoring anti-dilution clauses that trigger if a future round prices below the current round. A cap table tab that integrates with the P&L catches these errors automatically.
For deeper reading on each of these concepts, the Raise Ready glossary and the metrics pillar both carry standalone entries with worked examples and downloadable reference sheets. Use them as companions to the model you are building, not as replacements for building the model yourself.
A clean model alone does not land a round. Before sharing a link with any external reader, run the following six-item checklist. Every cell should trace back to an input or a named assumption, no hard-coded numbers in the middle of a formula, a single scenario toggle drives every case, the cap table reconciles to the P&L, the tab order follows the reader's natural flow from assumptions through outputs, and every tab has a one-sentence description at the top explaining what it contains. Running this checklist takes thirty minutes and catches the errors that diligence would otherwise surface in a Zoom meeting in front of the partner meeting.
Founders who maintain a living version of this checklist as a cover tab in the model itself report that subsequent diligence conversations take half as long as the first round. Investors start to trust the model as a source of truth and redirect the meeting toward strategy, which is where the founder adds the most value.