Here is the fully revised article, with only the AI-generated-sounding parts rewritten according to all 27 rules. The structure, data, and core arguments are preserved.
---
title: Model Before Revenue: The Only Assumption That Matters meta_description: How to build pre-revenue business model assumptions that investors trust and founders can test, with real data and a repeatable framework. slug: model-before-revenue reading_time_min: 7
You have no customers. You have no revenue. You have a slide deck with TAM/SAM/SOM that you copied from a template.
Here's the uncomfortable truth: every pre-revenue business model assumption you make will be wrong. The question isn't whether you'll be wrong — it's whether you'll be wrong in a direction that lets you learn something.
Most founders treat financial modeling like a costume. They build spreadsheets to impress investors, not to test whether their business works. According to Waveup's analysis of pre-revenue startup fundraising, the gap between "fundable" and "funded" often comes down to how well you can defend your assumptions — not how big your TAM is.
Let me show you what defensible assumptions look like.
The Two Types of Assumptions
Every pre-revenue model contains two categories of assumptions. Most founders confuse them.
Category 1: Structural assumptions — These are things you can calculate from first principles. Unit economics, cost structures, conversion rates from known benchmarks. You can defend these with logic and comparable data.
Category 2: Leap-of-faith assumptions — These are things you cannot know until you ship. Will customers pay $50/month? Will your CAC be under $200? Will churn be 5% or 15%?
The Federal Reserve's pre-provision net revenue models — used to stress-test the largest banks in America — make this distinction explicit. Their documentation, revised in December 2025, adds "additional parameters for the proposed suite of pre-provision net revenue models." Even the Fed, with decades of data on trillion-dollar institutions, knows that models need constant recalibration.
You have zero data. You need even more humility.
Build the Structural First
Start with what you can calculate. Not what you can guess.
Unit economics. If you're selling software, your hosting cost per user is calculable. If you're selling physical goods, your COGS is calculable. If you're selling services, your time per delivery is calculable.
These are not assumptions. They are arithmetic.
The Cambridge Dictionary defines "pre-" as "before (a time or an event)." Pre-revenue means you are before the event of revenue. But you are not before the event of cost. You can know your cost structure before you have a single customer. That's not a guess — that's a constraint.
Your job is to find the cost floor. What's the absolute minimum you must spend to deliver value to one customer? Not your ideal cost. Not your scaled cost. The minimum.
If that number is higher than what the market will pay, you have a cost problem, not a revenue problem. Fix the cost problem first.
The Leap-of-Faith Stack
Now for the hard part. You need to make assumptions about things you cannot know. Here is how to make them defensible.
Tier 1: Comparable benchmarks. Look at public companies in adjacent spaces. If you're building a B2B SaaS for HR teams, look at Workday, Rippling, BambooHR. What were their early conversion rates? What did their CAC look like at $1M ARR?
You won't find these numbers in public filings. But you can reverse-engineer them. If a company spends $50M on sales and marketing and adds $20M in new ARR, their blended CAC payback is 2.5 years. That's a data point.
Tier 2: Proxy experiments. Before you build the product, can you simulate the transaction? Sell a service version. Run a pre-order campaign. Put up a landing page with a "Buy Now" button that goes nowhere and measure click-through.
These experiments give you real data from real humans. They aren't perfect. They're better than nothing.
Tier 3: Sensitivity ranges. Don't give investors one number. Give them three: optimistic, base, pessimistic. And show what each scenario requires to be true.
If your base case requires 3,000 customers at $100/month, and your pessimistic case requires 1,000 customers at $75/month — and you can't get 1,000 people to sign up for a free trial — your pessimistic case is still too optimistic.
The Turn: Your Model Is a Hypothesis, Not a Forecast
Here is where most founders break.
They build a model that goes out 36 months. They project hockey-stick growth in month 18. They call it "conservative." It is not conservative. It is fantasy.
The Merriam-Webster definition of "pre-" is "earlier than : prior to : before." Your pre-revenue model is a pre-hypothesis. It is not a prediction. It is a set of statements that must be falsified before you spend real money.
The goal of your model is to identify which assumption, if wrong, kills the business. That single assumption is your risk axis. Everything else is noise.
For most businesses, it is one of three things:
Find your risk axis. Build your model around testing that specific assumption. Everything else is spreadsheet decoration.
How to Present This to Investors
Investors don't believe your numbers. They believe your thinking.
When you present a pre-revenue model, you're not presenting a forecast. You're presenting a logic structure. You're saying: "Here is how this business would work if these conditions hold. Here is how I will test each condition. Here is what I will do if the test fails."
Waveup's playbook — built from analyzing 600+ pre-revenue funding rounds totaling over $3B — shows that investors reward clarity over optimism. A founder who says "I don't know if CAC will be $50 or $150, but I will know within 90 days and here is my cost structure at both prices" is more fundable than a founder who says "CAC will be $75 because that's what our model says."
The Minimum Viable Model
You don't need a 50-tab Excel file. You need a model that fits on one page.
That's it. Four sections. Three scenarios each. One risk axis.
The Federal Reserve uses complex models because they are stress-testing institutions with 100+ years of data. You have zero data. Simple is better. Simple is testable. Simple is honest.
Resolution
Your pre-revenue business model assumptions will be wrong. That's not a failure. That's the starting point.
The question is: Are you building a model to defend, or a model to test?
If you're defending, you're playing theater. You'll raise money from people who don't ask hard questions, and you'll run out of money when reality contradicts your spreadsheet.
If you're testing, you're building a learning machine. Every wrong assumption teaches you something. Every failed prediction tightens your model. Every iteration brings you closer to a business that actually works.
The prefix "pre-" means before. Use this time before revenue to learn, not to pretend.
---
You have 16 modules of analysis between your idea and a fundable model. Cortex AIF runs them all in the time it takes to read this article. [Button: Test your assumptions before your investors do]