What a Founder-Grade Pro Forma Actually Looks Like

Most founder pro formas don't survive a real investor read.

They look right at a glance. Revenue tab, cost tab, summary tab. Big numbers grow over time. The Excel-fluent founder feels confident sending it to investors. Then the first real check-writer opens it, finds three formula errors in the first ten cells, and the round dies before it started.

That's not a math problem. It's a credibility problem. And it's avoidable.

A founder-grade pro forma does three things that most founder-built models don't. Each one is the difference between getting a second meeting and getting a polite pass.

What a real model is built around

A real model isn't a forecast. It's a defense of a strategy.

The headline numbers exist to support a position. "We'll do $4M in year two" isn't the point. The point is the underlying logic. How many customers. At what acquisition cost. With what retention. At what unit economics. The summary numbers fall out of those assumptions, not the other way around.

When investors read a model, they're not trying to verify the top-line forecast. They're trying to figure out whether the founder understands the business. The model is the artifact that proves you do.

The founder-built version usually starts with the headline. "We want to be at $4M." Then works backwards to fit the assumptions to the goal. Investors can spot this in 90 seconds. The assumptions don't hang together. The retention rates are too high. The acquisition costs are too low. The growth curve is too smooth.

The founder-grade version starts with the assumptions. Customer acquisition cost we can defend. Retention we have data on. Average order value tied to actual cohort behavior. Then the model rolls up. The headline is the output, not the input. When an investor pushes on a number, you can show the assumption it came from and defend it.

That single shift, from output-driven to input-driven, is what separates a real model from a pitch artifact.

The three layers every model needs

Any founder-grade pro forma has three layers, and most founder-built ones only have one.

Layer one is the assumptions tab. This is where every input lives. Customer acquisition cost. Conversion rate. Average order value. Retention by cohort. Headcount ramp. Salary by role. Office and overhead. Tooling. Each cell on this tab has a source. A real number from a real spreadsheet, a comparable company benchmark, a stated assumption you're willing to defend. The investor reads this tab to understand whether you've thought about the business.

Layer two is the model tab. This is where the assumptions roll into a P&L, balance sheet, and cash flow. Every cell on this tab is a formula. No hardcoded numbers. The reason matters. If the investor wants to stress-test the model, they need to be able to change the assumption tab and watch the model recalculate. A model with hardcoded numbers can't be stress-tested, which means it can't be trusted.

Layer three is the scenario tab. Three scenarios at minimum. Base case, downside, upside. Each one with a one-sentence description of what changes. "Downside: customer acquisition cost is 40% higher and retention is 15% lower." "Upside: we land one enterprise customer in year two." This tab is what an investor uses to do their own diligence. They'll change inputs to see what breaks the business. If the downside doesn't break anything, your model isn't honest. If the upside requires three miracles, your model isn't credible.

A founder-built model usually has just the model tab. The assumptions are baked into the formulas. There's no scenario work. The investor can't see how the founder thinks. So they pass.

What investors actually look for

Three specific things that signal a founder understands their business.

First, unit economics that hold up under pressure. Not just CAC and LTV in isolation, but the ratio over time. If the LTV is calculated on a 36-month cohort but you're 14 months old, that's a problem. A real model uses what you actually know and projects forward conservatively from there.

Second, headcount that maps to milestones. Adding two engineers in month six should be tied to a specific outcome. "We hire two engineers in month six to ship the second product line by month ten." Not "we hire engineers to grow the team." The investor wants to see that hiring is purposeful, not aspirational.

Third, cash runway that doesn't lie. Most founders model cash like it's revenue minus expenses. It isn't. Cash is collections minus payments, and the timing matters. A model that shows positive cash flow because it's confused about timing fails the basic credibility test.

If the model gets these three things right, the investor reads it as a sophisticated founder. If it gets them wrong, the investor reads it as someone who hasn't really thought about the business yet.

Why this is hard to do yourself

The reason most founders don't build founder-grade pro formas isn't laziness. It's that the work requires three specific skills that rarely live in the same person.

Spreadsheet engineering. Real models have hundreds of formulas, structured carefully so they don't break when an input changes. The skill of building a model that's both flexible and stable is its own discipline. Most founders are smart enough to write formulas, not skilled enough to architect a model.

Financial accounting fluency. The income statement, balance sheet, and cash flow have to tie. Net income flows to retained earnings. Working capital changes flow to cash. If the three statements don't reconcile, the model is broken. Most founders don't know what these reconciliations should look like.

Industry pattern recognition. The benchmarks for SaaS look different from DTC look different from marketplace. A founder-grade model has assumption ranges that match the industry. If a DTC pro forma shows 90% gross margin, it's wrong. If a SaaS pro forma shows 35% net retention, something's broken. Knowing the patterns is decades of practice, not an afternoon of research.

Most founders try to do this themselves and the result is a model that fails one or all three tests. Which is why we offer Pro Forma + Pitch Deck as a packaged engagement. We've built dozens of these. We know the patterns. We know what investors look for. We know what kills a round.

When to build the founder-grade version

Don't build a real pro forma if you're not raising. A simple operating budget is fine for internal use.

Don't build it if you're at the friends-and-family round. A simple cap table and a one-page financial summary is what those investors expect. A 14-tab model is overkill.

Build the real version when you're raising from professional investors. Seed funds. Family offices that write real checks. Strategic partners. These reads will be careful and pattern-matched. A weak model gets you a polite no. A strong model gets you the second meeting.

The cost of a real pro forma is small relative to the cost of a failed round. Most failed rounds aren't because the business isn't good. They're because the founder couldn't defend it credibly in numbers.

About Shelter 84. We're a brand and product studio in Los Angeles helping founders and small business owners build brands end-to-end. Financial models, pitch decks, brand identity, and the systems behind them.

Building a model for a real raise? Schedule a call at calendly.com/shelter84/30min.