Most financial models are evaluated wrong. A board member receives a 40-tab spreadsheet, scrolls to the summary, looks at the ARR line on the chart, and forms an opinion. That’s not evaluation — that’s letting the model author tell you what to think.
A useful model evaluation does the opposite. It ignores the headline number and interrogates the engine: how was the number produced, what assumptions drive it, what would have to be true for it to be wrong, and is the structure honest enough that pressure-testing it is even possible?
Here’s the seven-point checklist we use when reviewing models — as board members, as fractional CFOs preparing companies for fundraises, and as advisors helping investors evaluate diligence-stage models.
1. Can you trace any output cell back to an assumption in under 3 clicks?
Open the revenue summary for next year. Click the cell that shows total revenue. Trace it back. How many cells deep does it go before you hit a hardcoded number or a named assumption?
If the answer is 1 or 2, the number is probably hardcoded or the model is too shallow to be trustworthy. If the answer is 6 or 7, you’ve probably found a real model — the kind where revenue is driven by SDR headcount × quota × attainment × win rate × ACV, and each of those drivers is its own defensible assumption.
If the answer is “I can’t trace it because the formulas are too tangled,” that’s its own answer: the model is not built to be defended.
2. Are revenue drivers bottoms-up or top-down — and which is acceptable when?
A bottoms-up revenue model starts with operational inputs (reps, leads, conversion rates, ACVs) and produces revenue as an output. A top-down model starts with a market size and a market share assumption.
For early-stage companies (pre-Seed, Seed), top-down is acceptable because operational data is too thin to support bottoms-up. For Series A+, bottoms-up is required. Series B investors will reject a top-down model as fundamentally unable to support diligence.
Check which one you’re looking at by examining where the revenue formula starts. If the input is “market size × penetration rate,” it’s top-down and not serious. If the input is “monthly leads × conversion to opportunity × close rate × ACV,” it’s bottoms-up and worth your time.
3. Do the unit economics actually math out?
The three numbers to check: CAC, LTV, payback period. And critically, how they’re calculated, not just the numbers shown.
CAC. Is it computed on fully-loaded sales and marketing cost (salaries, benefits, software, marketing spend) or just marketing spend? If just marketing spend, the CAC is understated by a factor of 2–4x. Take the sales and marketing expense line from the P&L, divide by new customers acquired in the same period. That’s the real CAC.
LTV. Is it computed using gross margin or contribution margin? Many models use gross margin, which overstates LTV because it doesn’t account for the ongoing cost of serving the customer (support, infrastructure, success). Contribution margin is the more conservative and more honest input.
Payback. Is it bookings-based or cash-based? Bookings-based payback uses the contracted revenue value to compute payback months. Cash-based uses actual cash collected. For an annual-prepay SaaS company, the numbers differ meaningfully. Cash-based is what investors care about because it determines real burn.
If any of these are wrong by the standard above, the unit economics narrative is wrong, regardless of what the chart shows.
4. Is the hiring plan consistent with the revenue plan?
Open the hiring tab. Find the planned sales rep ramp. Compare to the revenue tab — does the revenue assume each rep is producing X per month starting in month Y after hire, and is that consistent with how revenue is actually growing?
Common failure: the model shows you hiring 4 sales reps in Q1 next year, and revenue inflecting in Q2 by an amount that would require those reps to be at full quota immediately. New SDRs and AEs typically take 3–6 months to ramp to full productivity. A model that doesn’t reflect ramp is overestimating revenue.
Other version of the same problem: the model has a flat headcount plan but accelerating revenue. Unless productivity per rep is realistically increasing (sometimes it is — better tooling, better playbooks), this is fiction.
5. Are there at least three scenarios — base, upside, downside — with explicit assumptions?
Every serious Series B model should include three scenarios. Each scenario should have:
- A clearly labeled set of input differences (faster ramp, slower close rates, higher churn)
- Output that flows through the entire model (revenue, burn, cash position, headcount plan)
- A summary slide or tab showing the comparison
If the model has only a base case, the founder hasn’t pressure-tested their own plan. Investors will, and the founder will be answering questions live in a Zoom call instead of with prepared materials.
If the “downside” scenario isn’t actually worse than the base — same growth rate, same burn — it’s not a real downside. A real downside should show what happens when something the company can’t control goes wrong: win rates drop, deal cycles lengthen, a major customer churns.
6. Is the cash position a live calculation or a hardcode?
Find the cash line. Is it computed from the cash flow statement (operating cash flow + financing + investing), or is it a separate calculation that doesn’t tie?
If it’s a hardcode or a parallel calculation, the model has a structural break — the cash position shown won’t match what would actually happen given the operating assumptions. This is more common than you’d think, because building an integrated three-statement model is hard and many founders’ first models are P&L-only with a separate cash tab.
A real three-statement model has cash as the output of operating cash flow plus financing activities. If you change a revenue assumption, the cash position changes automatically through the entire chain.
7. Does the model integrate to a balance sheet and cash flow statement?
Three-statement models tie. A change in revenue produces a change in receivables (balance sheet), a change in cash collected (cash flow statement), and a change in cash on the balance sheet. The balance sheet balances. The cash flow ties back to the cash line.
If the model has only a P&L and a cash tab, it’s a forecasting tool, not a financial model. That’s fine for a Seed stage company. For Series B, it’s a red flag — diligence will find the gaps the model doesn’t address.
Red flags worth pulling on
Beyond the seven-point checklist, here are the patterns that indicate a model isn’t ready for diligence:
- Hardcoded growth rates. “Revenue grows 8% MoM” sitting in a single cell, with no operational driver behind it. Means the founder hasn’t built the funnel math.
- Hockey-stick revenue reverse-engineered from a target. The growth rate isn’t an output of the operational plan — it’s a curve drawn to hit a target ARR, with assumptions back-solved to match. You can usually spot this because the operational inputs imply impossible per-rep productivity or vanishing CAC.
- Missing burn scenarios. No “what if we miss plan” math. Means the company hasn’t thought about the downside.
- No cohort analysis. Means the company doesn’t know how its customers actually retain. For a recurring-revenue business, this is non-negotiable.
- Lease commitments and other liabilities buried. Off-balance-sheet items (long-term leases, multi-year SaaS commitments, severance obligations) that don’t appear in the cash flow. Diligence will find these; better to surface them yourself.
Reading vs. building
This checklist is written from a board member’s perspective, but it works the other direction too. If you’re a founder building a Series B model, walk through these seven questions on your own model before you send it to investors. If any of them produce an uncomfortable answer, fix the model — diligence will produce the same uncomfortable answer in front of an investor, and it’s much worse when it happens there.
If you’re a board member or investor evaluating a model, the same questions apply. The difference is what you do with the answers: a clean pass on most of these is the difference between a model worth a partner meeting and a model that will fall apart in diligence.
What Pegacorn does
We build investor-grade financial models for Series B and beyond. We also help founders pressure-test models they’ve built internally — a kind of friendly diligence pass before the real diligence — to find the holes before investors do.
If you’re building a model for your next raise or evaluating one as a board member, we can help.