What an investment financial model should actually show

Project & Investment2026-10-017 min readAtlas Strategy Group

Companies present financial models as prophecies: here is the future, accept it. Professional readers treat them as something else entirely — a description of the machine, built to be tested. The gap between those two understandings explains most failed raises.

The model shows mechanics, not the future

Nobody believes your forecast — including the reader who approves it. What a serious reader evaluates is the machinery underneath: what converts money into revenue (the drivers, and whether they connect to anything real), what the cash eats on the way (working capital, capex, debt service), when the company needs money and what precisely it will do with it. A model with honest mechanics and a conservative forecast raises money. A model with decorative mechanics and a hockey stick does not — it merely signals that the team has not understood the question.

What the reader actually tests

• Driver logic. Do revenue assumptions trace to capacity, price, channels and conversion — or do they trace to a growth percentage typed into a cell?
• Sensitivity concentration. Every model has two or three assumptions that carry the outcome. The reader looks for whether the builders know which ones they are.
• The stress points. Where the plan breaks: the delay in customer payment, the season of negative cash, the utilization threshold. A model without visible break points is not optimistic — it is unexamined.
• The honesty of the base case. The base case is a claim about judgment. An aggressive-but-defensible base case earns trust; a timid or a fantastic one spends it.

Three futures, not one

A model earns credibility precisely where the forecast is weakest: in the scenarios. What happens to cash if the two central assumptions are wrong — the market grows slower, the ramp costs more? Scenarios are not decoration; they are the model answering the question the reader is actually asking: what is the shape of the downside, and does the company survive it? A single-line forecast invites the reader to invent the downside alone. Inventing it alone is what kills deals.

The deck gets the meeting. The model survives the diligence. Build for the second audience.

The model is a management tool, not just a fundraising artifact

There is a quieter test of a model's quality: does the management team run the company with it after the round closes? If the drivers are real, they keep working as a dashboard of the business — actuals against assumptions, and the conversation about variance instead of hope. If the model dies in the data room, it was never a model of the business. It was a costume for one.

This is the standard we build financial models to: drivers a reader can interrogate, scenarios that pre-answer the hard questions, and a base case the team can defend line by line. The project and investment practice page describes where this work sits in the full engagement — from investment readiness to the deal itself.