How to Build a Startup Financial Model: Mise en Place for Your Money

A startup financial model is the spreadsheet where your assumptions about revenue, staff, and costs flow into a forecast of cash, profit, and how many months you have left. But a great kitchen is decided before a single order comes in, in the preparation no one in the dining room can see. So is a startup built to hold through its first bad month.

Look at a real cook before service. The station is set. Ingredients are chopped, sauces are ready, every tool is in its place. Cooks call this mise en place, putting in place. The meal is half cooked before the first order, because the real work happened in the preparation.

Your model is mise en place for your money.

Most founders think they fail at the model, when really they fail at the preparation. A dirty assumptions sheet is a dirty station, and a dirty station can damage the whole service.

Most startups do not fail at the model. They fail at the prep.

I have built these from an empty sheet more times than I can count, and the hard part was never the formulas. It was the discipline to set the station first. So I will walk you through a 3-statement model the way a chef walks a new cook through a line, one station at a time.

What a startup financial model actually is

Set aside the show and a model is one set of inputs flowing into outputs through formulas. The inputs are your assumptions. The outputs are the line items an investor reads, cash above all.

A real startup model is built around three statements. The income statement shows revenue minus costs, with profit or loss at the end. The cash flow statement follows the money in and the money leaving, month by month. The balance sheet holds it all together: what you own and what you must pay. Together they form the backbone. Skip one and the model will break.

You are not building this for accuracy. Your forecast will be wrong about revenue every single month, and that is expected. You are building it for the thinking it forces. A founder who builds the model can sense the business the way a chef can sense a service before the first order: where growth comes from, what the cash going out actually is, and how many months the cash gives you.

Set the station: your assumptions sheet

Every service must begin with mise en place. In a model, that station is your assumptions sheet. This is the one place where you type real numbers. All the rest reads from here.

Keep the inputs simple and separate. Put your price, your growth rate, your wages, your staffing plan, and your fixed costs on one sheet. Color them, so future-you can see these are the controls you pull.

Why does separating the inputs matter so much? Because when a board member asks what if growth is half that, you change one cell and the whole model can adjust. A station with real numbers lost inside formulas is one no one set. It will look ready until the orders land.

Build the revenue station

Now the first station. Revenue is not one number. It is built from parts: items sold times price, or new customers times what each one pays, minus the ones who leave.

Choose the parts that match how your business actually makes money. For a business with members who pay each month, that is new customers, the revenue per customer, and the rate they cancel. For a one-time sale, it is volume and price. Forecast each part forward, month by month, and revenue comes out as the result. Do not type a revenue number. Build it.

This is where the model can earn its keep. When you can see revenue as a chain of parts, you can put the harder questions to it. Is the growth rate realistic, or just a wish? Does the staffing plan support that many new customers? The model forces the hard discussion an easy answer never could.

Cost the menu: staff and the cash going out

A course that costs more than it sells for is one you should not serve. Costs are the other half of the line.

Divide costs in two. The costs that scale with revenue are first, the cost of what you sell. Then the fixed costs that hold steady: wages, software, rent. Staff is usually the biggest line for a startup, so build a small staffing plan and let wages flow from it. Add a cook, and the cost shows up on its own.

Take all of it from revenue and you have profit or loss. Take the cash costs from the cash in and you have the money leaving each month. Line that up against the cash you hold, and the model can tell you the one number every founder must know. How many months you have left.

months left = cash on hand / monthly cash going out

Link the three statements

Here the line finally moves as one. The three statements are not separate sheets. They link.

Profit from the income statement flows into cash. Cash flows into the balance sheet. What you own and what you must pay link back. When the links hold, change one assumption and every statement can adjust at once. That connected, working model is the whole service running off one preparation.

This is also where most models break. If the balance sheet does not balance, a link is wrong in one of them. Find it before you show a single investor. Better still, build the model so it cannot happen, with every line reading from one station, so a wrong number shows up in one place rather than ten. A model that does not add up is a dirty course going to the dining room, and a real investor will find it before you do.

Add the scenarios

One forecast is a single course with no menu around it. Build three: a base case, a best case, and a bad case.

You already separated your assumptions, so this part is simple. Adjust the growth rate, the staffing plan, the price, and the months you have left will grow or get smaller. Scenario planning will not tell you which case happens. What it does tell you, before the orders land, is how much room you have before the cash runs out. That is the steady hand the preparation gives you.

It paid off for me once. A founder I worked with, call her Dana, ran her three cases the week before a board call. The bad case put cash going out two months earlier than she had assumed, because one big customer paid late. She held the room. She had already prepared it. She went into the call with the number, the cause, and what she would do about it. The board funded the next round. The preparation is what gave her that steady hand.

Common questions on a startup financial model

How long should building a startup financial model take?

A first working model is a week, not a long project. Set the station small. A simple 3-statement model with one good assumptions sheet is worth more than a large one you do not understand. You can build out the rest of the menu later.

Do I need to know accounting to build one?

No. You need to understand three statements and how cash, profit, and the balance sheet link to one another. The arithmetic is addition, subtraction, and a few formulas. The hard part is the thinking, not the accounting.

What is the most common error first-time founders make?

Typing numbers straight into the middle of the model instead of onto a separate assumptions sheet. It turns a usable tool into one that will break. Separate your inputs, set the station first, and the rest gets easy.

Spreadsheet or software?

Most startup models live in a spreadsheet, and that is a good place to learn. The problem comes when the sheet grows and no one believes the numbers. That is the gap CX Cash is built to close.

The stand: prep it yourself

A founder who cannot build their own model is cooking without preparation. No one outside the business feels the cash crunch you do at 2 in the morning. So build the model yourself. Set the station, cost the menu, link the statements, run the cases. Own it.

You should know where the money is going. CX Cash gives you the station already set: our 3-statement model and scenario planner, run on your own numbers, with the founders building cash discipline into the company from day one. Build your own. Then share it with a founder still cooking with no preparation at all.