Founders spend weeks polishing the revenue forecast in their Series B model. They obsess over the chart that shows ARR going to $50M. They tweak growth rates until the trajectory looks right.
Then the investor opens the model, ignores the revenue number entirely, and asks: “Walk me through how you got there.”
Investors do not care about your projections. They care about whether you can defend the assumptions underneath them. The number on the cover slide is the output of a thought process; the model is the artifact that shows whether that thought process is rigorous or wishful.
Here’s what they actually look at.
Bottoms-up, not top-down
Top-down models start with “the market is $30 billion and we’ll capture 0.5%.” They are useless. Every investor has seen a thousand of them.
Bottoms-up models start from the unit. For a SaaS company: how many SDRs do you have, what’s their quota attainment, what’s the conversion rate from MQL to opportunity to closed-won, what’s your average contract value, what’s the gross retention. The revenue number falls out of those assumptions. When an investor questions the revenue, you walk back through the funnel and show them where it came from.
If you can’t trace your top-line number to a specific operational input, your model isn’t defensible.
Unit economics, in detail
CAC, LTV, payback period, gross margin per customer. These are table stakes. What separates a good model from a mediocre one is whether they’re computed correctly.
Common mistakes: CAC that excludes loaded sales costs (salaries, benefits, tools). LTV calculated with gross margin instead of contribution margin. Payback computed on bookings instead of cash collected. None of these will survive a Series B diligence call.
Sensitivity, not point estimates
A real model lets you change three or four key assumptions and see the impact ripple through. What if your win rate drops 10%? What if your AE ramp takes 5 months instead of 3? What if churn ticks up by 1 point per quarter?
If your model has a single “base case” and no scenarios, you’ll get asked to build them in the middle of the process. Build them first.
Scenarios that include a downside
Investors will ask: “what happens if you miss plan by 20%?” If your answer is “we’d cut hiring,” that’s not enough. They want to see the model. They want to see when you hit zero cash, what your headcount would look like, and how long the cash you raise lasts under that scenario.
This isn’t pessimism — it’s diligence. Companies that have thought through the downside are the ones that can react fast when reality goes sideways.
The default-alive question
Paul Graham’s framing is now the question every investor will ask in some form: “If you stopped raising tomorrow, would you reach profitability before running out of cash?”
You don’t have to answer “yes.” You have to answer the question precisely, with numbers. “At our current burn, no. But if we paused hiring and reduced marketing spend by 40%, we’d reach default-alive in 18 months at current growth.” That’s a defensible answer. “We just wouldn’t stop raising” is not.
CAC payback by cohort
If you’re a recurring-revenue business, expect investors to ask for CAC payback by cohort — how long does it take a cohort of customers acquired in Q1 2025 to pay back what it cost to acquire them? They’re looking for whether the number is getting better or worse over time. If it’s getting worse, they want to know why.
A model that surfaces this cleanly tells investors you understand your own business. A model that buries it behind a tab of raw data tells them you don’t.
Red flags investors look for
Five things that will tank diligence:
- Hardcoded numbers buried in formulas. If your “model” has hardcodes overriding the math, investors will notice and assume there are more they didn’t catch.
- Hockey-stick revenue with no operator explanation. If revenue inflects in month 14 with no corresponding change in headcount or pipeline assumptions, the model is fiction.
- No downside scenario. Means you haven’t thought about it, or you’re hiding it.
- No cohort analysis. Means you don’t know your retention dynamics.
- Inconsistent definitions. ARR defined three different ways across the same model. Net dollar retention computed wrong. Bookings and revenue used interchangeably.
The model that gets you a Series B isn’t the one with the highest growth rate. It’s the one where every number traces back to an operator’s understanding of the business — and where the founder, opening the model live in a Zoom call, can answer any question without scrolling for two minutes to find a tab.
If you’re 3–6 months from a raise and your model isn’t there yet, we can help.