Every founder walking into a fundraise believes their model is solid. Then an associate opens the data room, pulls up the model, and starts asking why pipeline coverage assumes a conversion rate the company has never actually hit. The founder scrambles for an answer they should have had ready three weeks earlier.
This is the moment stress-testing exists to prevent.
What “Stress-Testing” Actually Means
Many founders hear “stress-test the model” and think revenue sensitivity: what happens if growth comes in 20% lower. That’s part of it, but it’s the shallow end. Real stress-testing means going assumption by assumption through the model and asking which ones are load-bearing, which ones are guesses dressed up as inputs, and which ones would break the whole plan if they turned out wrong.
VCs run this exercise on every model they see, whether or not the founder does it first. Associates and principals have a standard playbook: pull the model apart, isolate the assumptions that drive the outcome, and test whether those assumptions hold up against the company’s own historicals and against market comps. They’re not doing this to be adversarial. They’re doing it because the model is the only artifact in the process that claims to predict the future, and their job is to find out how much of that prediction is real.
The founders who come in with their own stress test already done aren’t hiding anything. They’ve simply done the diligence before the diligence, which changes the entire tenor of the conversation.
The Five Areas VCs Stress-Test in Every Model
Revenue assumptions. This is where most scrutiny lands first. Investors will check pipeline coverage ratios against the sales cycle length, question whether the conversion rate assumed in the plan has ever actually been achieved, and probe ACV and churn numbers against what the company has shown historically versus what it’s projecting going forward. A model that assumes conversion rates the company has never hit is a model that will get flagged in the first pass.
Gross margin assumptions. Cost-to-serve, pricing structure, and product mix all roll into this line, and it’s one of the easiest places for optimism to creep in unnoticed. If the model assumes margin expansion, the investor will want to know exactly what’s driving it: better unit economics from scale, a pricing change already in motion, or a shift in mix toward higher-margin products. “Margins improve over time” without a mechanism behind it won’t survive contact.
Headcount plan. Every headcount plan embeds a productivity assumption, whether the founder states it explicitly or not. Investors will back into revenue per rep, or per engineer, or per support agent, and compare it against the company’s own trend and against comparable companies at similar stages. Loaded cost matters here too. A plan that quietly under-counts benefits, taxes, and overhead will produce a burn number that doesn’t match reality once the hires actually happen.
Cash management. DSO, capital efficiency, and burn trajectory get their own line of questioning because they determine runway, and runway determines how much leverage the company has in the round. A model that shows healthy growth but glosses over collections timing or capital intensity will draw questions about how much of that growth is actually cash-generative.
Milestone assumptions. Every fundraise model implicitly depends on hitting certain milestones before the next round, whether that’s a product launch, a key hire, or a specific ARR threshold. Investors will ask what the plan depends on and what happens if those milestones slip. A model that never states its own dependencies looks like it hasn’t been stress-tested at all.
Running a Pre-Process Stress Test
The goal before a raise isn’t to make the model look perfect. It’s to know exactly where it’s fragile before someone else finds out for you.
The pessimistic analyst exercise. Rerun the entire model with every key assumption set 20% worse: conversion rates down, churn up, cost-to-serve up, hiring productivity down. Look at what happens to runway and to the milestones the plan depends on. If the company survives this version with a credible path forward, that’s a strong signal. If it doesn’t, that’s information worth having before an investor surfaces it in front of the whole partnership.
The one-thing-goes-wrong scenario. Rather than degrading every assumption at once, isolate the single biggest risk to the model and run that scenario in isolation. For a usage-based pricing company, that might be a large customer downgrading. For a company with sales-led growth, it might be a lengthening sales cycle. Naming this risk explicitly, and having a plan for it, is far more credible than pretending the risk doesn’t exist.
The growth-costs-money check. Ambitious growth targets require investment to match, and it’s common for models to project the outcome of an aggressive plan without fully funding the plan that would produce it. Check whether the headcount, marketing spend, and infrastructure investment in the model are actually sufficient to hit the growth numbers being projected. A model that shows hockey-stick growth on a flat cost base won’t hold up to questioning.
Fixing the Gaps Before the Process
Not every gap found in a stress test needs to be closed in the model itself. Some gaps are better addressed with narrative, and knowing the difference is part of the judgment a CFO brings to the process.
If the gap is structural, meaning the underlying unit economics don’t support the growth plan, that needs to be fixed in the model before the raise starts. Investors will find it, and no amount of narrative will paper over a model that doesn’t compute.
If the gap is a known unknown, like a milestone the company hasn’t hit yet but has a credible plan to hit, that’s often better handled through the story the founder tells rather than through changing the numbers. Investors don’t expect every assumption to already be proven. They expect the founder to know which assumptions are proven and which are bets, and to be direct about the difference.
The gaps that stall deals are the ones the founder didn’t know were there. A CFO who has run the stress test in advance walks in already knowing which category each weakness falls into, and has an answer ready either way.
The Investor Conversation After Stress-Testing
Once the stress test is done, it changes the entire framing of the model in front of investors. Instead of presenting a single projection and hoping nobody pokes at it, the founder can present a model that has already been pressure-tested against its own weaknesses.
The framing that works: “We’ve modeled what happens if we miss our key assumptions, and here’s our plan for each scenario.” This does two things at once. It shows the investor that the founder understands their own business well enough to know where it’s fragile, and it preempts the exact line of questioning the investor was planning to run anyway.
A model presented this way reads as battle-hardened rather than optimistic. That distinction is often the difference between a partner meeting that builds conviction and one that raises more questions than it answers.