Founders tend to assume that an investor opens a financial model and starts at the top line. In our experience supporting fundraises across three continents, that is almost never what happens. An experienced reviewer is looking for reasons to trust the file before they will engage with what it claims.
Here is the order it usually happens in.
1. Does the balance sheet balance?
It sounds trivial. It is the fastest possible integrity test, and a surprising number of models fail it, usually because someone hard-coded a plug to force it closed. A reviewer will flip to the balance sheet, look at the check row, and toggle a single assumption to see whether it stays balanced.
If it does not, everything downstream is now suspect. Not wrong necessarily, but suspect, which is worse.
A model that only balances under the base case is not an integrated model. It is three separate spreadsheets standing next to each other.
2. Are the assumptions separated from the calculations?
The second thing a reviewer does is look for the inputs. If assumptions are scattered through formulas, a growth rate typed directly into a revenue cell, a headcount number buried in an expense row, the model cannot be interrogated. And a model that cannot be interrogated cannot be trusted.
Every input belongs on a dedicated assumptions sheet, colour-coded, labelled, with a source noted against it. Every other cell should be a formula.
3. Where does revenue come from?
This is where most early-stage models lose credibility. Top-down revenue, "the market is worth $4 billion, we will capture 2%", tells a reviewer nothing except that the founder has not thought about how customers are actually acquired.
Bottom-up revenue is built from things you can defend:
- How many leads enter the funnel each month, and from which channel
- What proportion convert, based on your own observed data
- What each converted customer pays, segmented if pricing varies
- How long they stay, based on your actual cohort retention
The number at the end may be smaller. It will be far more persuasive.
4. What happens in the downside case?
The final check is whether the model can survive being wrong. A reviewer will look for a scenario toggle and, finding one, will immediately run the downside.
What they are testing is not the outcome. It is whether you have thought about it. A founder who can say "if conversion drops thirty percent, here is the month we would need to act and here is what we would cut" is materially more fundable than one who has only modelled success.
The uncomfortable summary
None of the four checks above are about your business. They are about whether your model is a serious document. You can have an excellent company and lose a round because the file suggested otherwise. We have watched it happen.
Build the model to be audited, not to be admired.