What investors check in a startup financial model
Most investors do not read a model line by line. They run a short set of tests that tell them within minutes whether the numbers can be trusted. If your model fails these tests, the rest of the pitch loses weight. Here are the seven checks we see most often, and how to prepare for each.
1. Does the balance sheet balance in every period?
This is the first and fastest test. If assets do not equal liabilities plus equity in any month, the model has a linking error somewhere, and every number downstream is suspect.
Add a visible check row on the balance sheet that shows zero in every period, and a master check on the summary page. Investors notice when it is there, and they notice faster when it is missing.
2. Is revenue built from drivers or typed in?
A revenue line that grows by a fixed percentage each month tells an investor you picked a number first and worked backwards. A driver-based revenue build shows how many customers you acquire, at what price, with what retention.
- SaaS and subscriptions: new customers, churn, average revenue per account.
- Marketplaces: active buyers and sellers, order frequency, take rate.
- Services and projects: pipeline, win rate, project size, delivery capacity.
Each driver should be an input on one assumptions sheet that can be traced and challenged.
3. Do cash, burn and runway tell the same story as the plan?
Investors compare your hiring plan, your monthly burn and your stated runway. If you say eighteen months of runway but the payroll growth in the plan shortens it to twelve, the inconsistency is the finding.
Model working capital explicitly. Receivable days, payable days and VAT timing can move cash by months even when profit looks healthy.
4. Are Saudi-specific costs included?
For a company operating in Saudi Arabia, investors expect to see the local cost base. That includes GOSI contributions on salaries, expat-related costs for non-Saudi hires, the effect of Nitaqat on your hiring mix, 15% VAT and its cash timing, and Zakat or income tax depending on ownership.
A model that treats salary as the only people cost, or ignores Zakat entirely, reads as built for another market. Confirm current rates and rules with your accountant or the relevant authority before you finalise assumptions.
5. Are the assumptions defensible?
Every major assumption should have a source: your own historical data, a signed contract, a pilot result or a market benchmark. Where there is no data, label it as a judgement and show the range.
Investors forgive uncertainty. They do not forgive assumptions that look invented.
6. Do the scenarios change the right things?
A base, conservative and upside case should differ in the drivers that actually carry risk: conversion, churn, sales cycle, cost of acquisition. A scenario that only scales every revenue line by the same factor teaches the investor nothing.
Include one downside case in which the raise is delayed, so the investor can see what you would do with six fewer months of cash.
7. Does the use of funds connect to milestones?
The amount you raise should buy something specific: a runway to a defined milestone, such as a revenue level or a product launch. Link the use of funds to the hiring and spend in the model, and show the month in which each milestone is reached.
A quick pre-meeting routine
- Check that the balance sheet balances in every period.
- Change three key assumptions and confirm the outputs move sensibly.
- Compare runway on the summary page with the cash line month by month.
- Make sure every assumption has a source or a stated range.
- Export a one-page investor summary that matches the model exactly.
Common questions
How long does an investor spend on a financial model?
Usually minutes on the first pass. They look at structure, checks, the main drivers and the summary. A detailed review comes only if those pass.
Should the model be in Excel or Google Sheets?
Either works. Excel is the more common expectation in due diligence, but a clean, well-structured sheet in either tool is what matters.
How many years should a startup model cover?
Most investors expect monthly detail for the first one to two years and annual figures out to three to five years. Agree the horizon with your investor if you can.