Day 10: Build a Simple Financial Model From Scratch

The Concept

Most early-stage founders avoid building a financial model for one of two reasons: they think they need an accountant or CFO to do it properly, or they believe their business is too early-stage for numbers to be meaningful. Both beliefs are expensive. A financial model is not a prediction. It is a structured way of making your assumptions explicit so you can see which ones matter most and test whether your business can work before you find out the hard way that it cannot.

The three things every investor — and every founder — needs to understand about a business are simple: how the business makes money (the revenue model), whether it makes more from a customer than it spends to acquire and serve them (unit economics), and how long the company can survive at its current rate of spending (cash runway). You do not need a 40-tab spreadsheet to answer these three questions. You need honest inputs and a clear-eyed look at what they imply.

The Difference Between a Model and a Spreadsheet

A spreadsheet contains numbers. A financial model contains relationships between assumptions. The distinction matters because a spreadsheet tells you what happened or what you expect to happen given fixed inputs. A model lets you ask "what if?" — what if my churn doubles? What if my CAC is 30% higher than I estimated? What if revenue growth is half what I projected? A model that shows how your runway changes when key assumptions shift is worth infinitely more than a projection that assumes everything goes according to plan.

Causal is designed exactly for this: it builds interactive models where you can drag a slider on your churn rate and watch your runway recalculate in real time. For founders who prefer flexibility and want full control over the structure, Google Sheets with clearly labelled assumption cells at the top works just as well. The tool matters less than the discipline of separating assumptions from calculations and testing what happens when those assumptions are wrong.

Unit Economics: The Number That Determines Everything

Unit economics answers one question: does your business make money on each customer? The answer lives in the relationship between three numbers — your average revenue per customer over their lifetime (LTV), how much it costs you to acquire that customer (CAC), and what it costs you to serve them (your gross margin). A healthy SaaS business typically runs an LTV:CAC ratio of at least 3:1 and recovers its CAC within 12 months. If your numbers do not currently meet those benchmarks, knowing that is not a crisis — it is information. It tells you whether to focus on increasing revenue per customer, reducing acquisition cost, or reducing churn.

The honest truth about early-stage unit economics is that CAC is often unknown because acquisition channels are not yet established, and churn is often unknown because the product is new. The model still matters even with unknown inputs, because AI can help you reverse-engineer what those numbers need to be for the business to work. If your model only works when CAC is under $50 and your current experiments suggest $200, that is a signal that requires action now, not when the money runs out.

Reading Your Runway With Clear Eyes

Cash runway is the number of months you can operate at your current burn rate before you run out of money. Most founders know this number approximately; few track it precisely. The reason precision matters is that the decisions you need to make — whether to hire, how aggressively to spend on acquisition, when to start fundraising — all change based on whether you have six months of runway or eighteen. Fundraising typically takes three to six months from first outreach to wire transfer. If you wait until you have four months of runway to start, you have already made the problem harder than it needed to be.

Running this model today, even with imperfect inputs, gives you a baseline. The specific numbers matter less than the habit of updating them monthly and making decisions with current data rather than assumptions that were last reviewed at founding.