Your model says you need $1.5 million. Your deck says $1.2 million. Your investor update implies 18 months of runway, while the cash flow tab shows 13. This is how good founders lose time in a raise: not because the business is weak, but because the numbers tell different stories.
Learning how to audit a financial model means checking whether the story survives contact with its own assumptions. You are not trying to make the model look prettier. You are trying to find the wrong number while it is still in the draft.
Start with the question the model must answer
A financial model is not a spreadsheet full of formulas. For an early-stage startup, it is an operating argument: this is how customers arrive, this is what they pay, this is what it costs to serve them, this is when the company needs cash and this is what new capital changes.
Before checking a single formula, write down the decisions the model is meant to support. Usually, that includes how much to raise, when to start hiring, what growth is required to reach the next fundable milestone and how long the company can operate if sales take longer than planned.
If the model cannot answer those questions plainly, adding more tabs will not fix it. Strip it back. A pre-seed model with a clear revenue build, hiring plan, operating expenses and monthly cash balance is more useful than a complicated workbook nobody can explain.
Audit the assumptions before the formulas
Most model failures begin with an assumption that sounded reasonable in a conversation and became dangerous when copied across 36 months.
Open the assumptions tab and ask where every material input came from. A price may come from signed contracts, customer interviews, competitor research or a founder's best judgment. Those are not equally reliable. Label them accordingly.
Look hardest at the inputs that drive the outcome: price, sales cycle, conversion rate, churn, headcount timing, compensation, gross margin and payment timing. A small change in any one of these can move the cash-out date by months.
For example, a founder may assume a $20,000 annual contract value and a 60-day sales cycle. That can be a fair starting point. But if the model also assumes customers pay annually upfront, never churn and require no implementation work, the forecast is carrying several favorable bets at once. The issue is not optimism. It is unmarked optimism.
Use three cases for the assumptions that matter most: base, downside and upside. The base case should be the plan you intend to run, not a compromise between hope and fear. The downside case should reflect a plausible delay or miss, such as a longer sales cycle, slower hiring or a customer segment that converts at half the expected rate.
How to audit a financial model line by line
Once the assumptions have a source and a rationale, test whether the spreadsheet applies them correctly. Work from the top of the operating logic down to the cash balance. Do not jump first to the valuation output or the chart on the summary page.
Revenue must match the selling motion
Start with the unit that creates revenue. For a self-serve software company, that may be website traffic, trial conversion, paid accounts and monthly retention. For enterprise software, it may be qualified opportunities, win rate, contract value, sales cycle and the timing of closed deals.
Check that new customers become revenue in the correct month. Check whether annual contracts are recognized over time or counted as full revenue when cash arrives. Check whether churn removes customers from the revenue base. If the company has usage-based pricing, make sure expansion assumptions are distinct from new customer acquisition.
Then compare the revenue build with the narrative in the deck. If the deck says the company will close 30 customers next year, the model should show where those customers come from. “Growing demand” is not a revenue driver. A defined sales motion is.
Costs must arrive when the work arrives
Headcount is often the largest cost line and the easiest place to accidentally create fiction. Confirm each hire has a start month, salary, payroll tax or benefits load and a reason for being hired at that point.
A sales hire should not appear three months before the company has enough pipeline to support one. An engineering team should not remain flat while the product roadmap adds major work. There is no universal right answer, but the sequence has to make operational sense.
Review non-payroll costs with the same discipline. Hosting should move with usage or customers where appropriate. Marketing spend should connect to a channel and an expected outcome. Legal, accounting, insurance and recruiting costs may be lumpy rather than evenly spread each month.
Cash is not the same as revenue
Investors will look at revenue, but founders run out of cash. Audit the cash flow separately from the profit and loss statement.
Ask when customers pay, when vendors are paid and whether there are deposits, annual prepayments, implementation costs or commissions that occur before revenue is recognized. A company can look healthy on an annual income statement and still hit a cash problem in a specific month.
Trace the ending cash balance month by month. Beginning cash plus cash received, minus cash paid, should equal ending cash. If it does not, stop there. Do not move on until the bridge works.
Test the spreadsheet mechanics
A model can have sensible business assumptions and still fail because one row stops pulling through after month 14. This is where a methodical review pays for itself.
Check the first, middle and last months of each major schedule. Look for hardcoded numbers inside formula ranges, broken references, formulas that change unexpectedly across columns, circular references and rows that no longer feed the statements they are supposed to support.
Use a few simple tests:
- Change one key assumption, such as price or hiring date and confirm every related output moves as expected.
- Set revenue to zero and verify that variable costs, cash and runway respond logically.
- Delay the fundraise by three months and see whether the model identifies the cash gap.
- Compare monthly totals with annual totals and make sure rounding does not hide a material difference.
- Search for spreadsheet errors, blank cells in active ranges and numbers typed over formulas.
These tests are not glamorous. They catch the mistakes that turn a reasonable model into an awkward investor meeting.
Reconcile the model with every other company document
The model does not live alone. It must agree with the deck, cap table, data room, investor update, hiring plan and founder's spoken pitch.
Pull the core figures from each document: current cash, monthly burn, runway, revenue to date, forecasted revenue, customer count, headcount, raise amount, use of funds and next milestone. Put them side by side. Any mismatch needs one owner and one correction.
Some differences are valid. The deck may present annual recurring revenue while the model tracks recognized monthly revenue. The key is that you can explain the difference in one sentence and reconcile it on paper. If you cannot, an investor will assume the company has not decided which number is true.
This is the practical value of one shared company context. At Firmgrove, the standard is simple: every number agrees everywhere it appears and the wrong number dies in the draft.
Read the model like an investor will
An investor is not only looking for a forecast. They are looking for judgment. They will notice a model that assumes rapid growth without showing acquisition costs, a team plan that ignores compensation or a raise amount that ends exactly at zero cash.
Give yourself room for the questions they will ask. What happens if the next round takes longer? What evidence supports the conversion rate? Why does gross margin improve? Which milestones does this capital buy and what would make the company ready for the next raise?
You do not need to pretend the future is certain. A clean answer is stronger: “This is our base case. If enterprise procurement adds 90 days, we have this downside plan and here is the hiring decision we would delay.” That is a founder who knows what the business needs.
Run this audit before every fundraise, board discussion and major hiring decision. The goal is not a perfect forecast. It is a model honest enough to help you make the next decision before cash, time or credibility gets expensive.