Key Takeaways
How investors value: only 22% of 885 surveyed venture capitalists use net present value methods, while 63% rely on cash-on-cash multiples (NBER, 2016).
No forecast at all: 31% of early-stage investors in that survey do not forecast cash flows.
Target margins: Damodaran's January 2026 data puts the pre-tax operating margin at about 33% for software and about 16% for restaurants. That is the destination, not year one.
Failure priced in: in Damodaran's worked example, a 40% failure probability cuts a $177.56 million going-concern value to $106.54 million.
Terminal value: it can reach "90%, 100% or even more than 100%" of a young company's value, so one growth rate can decide the result.
No averaging: the International Valuation Standards (2022) tell valuers to reconcile differing results into one conclusion "without averaging."
Why Does the Valuation Number Come Last?
The number comes last because investors do not examine it. They examine the assumptions underneath it. Many founders build a valuation backwards, starting from the figure they hope to raise at and adjusting growth until the spreadsheet agrees. Investors spot this quickly.
Few investors even run the textbook calculation. Gompers, Gornall, Kaplan and Strebulaev (NBER, 2016) surveyed 885 venture capitalists. Only 22% used net present value methods, while 63% relied on cash-on-cash multiples. Among early-stage investors, 31% did not forecast cash flows at all.
Your model is not a price tag. It is how you argue about price on your own terms.
What Goes Into a Startup Valuation Model?
A defensible startup valuation model has three parts, built in one order: Drivers, Futures, Cross-Check. Drivers are the few inputs value depends on. Futures are a success case and a failure case, each with a probability. The cross-check is a second method that tests the first, with the gap explained.
Drivers = revenue path, target margin, reinvestment and risk.
Futures = a success case and an explicit failure case, each with a probability.
Cross-Check = a second method that tests the first, with the gap explained.
Aswath Damodaran of NYU Stern, whose 2009 paper on young companies underpins this order, puts it bluntly: "less (detail) is often more (precision)." Five sourced drivers beat a hundred-line expense forecast.
Quick test: raise your revenue growth by ten points, then your discount rate by five. Which moves the value more? If you cannot answer, you have not yet found the driver an investor will attack first.
How Do You Build a Business Valuation Model in Three Steps?
Build the model in three steps. First, borrow the target margin and reinvestment rate from mature companies in the same business. Second, value the success case and weight it by the chance of survival. Third, run a second method suited to your stage and write one paragraph explaining why the two numbers differ.
Step 1: Anchor Every Driver Outside the Company (The Mature Twin)
A young company has no history, so borrow one from the businesses it hopes to become. Damodaran sets the target margin from established companies in the same business and plots a "pathway to profitability" toward it. He ties growth to reinvestment through a sales-to-capital ratio. His January 2026 dataset of US listed firms puts the pre-tax operating margin at about 33% for system and application software. For restaurants it is about 16%. That is the destination, not year one.
What to do: take target margin and reinvestment from mature peers, then plot the path.
Why it works: every driver has a source an investor can check.
Best for: businesses with established peers.
Build the revenue path from a bottom-up TAM, SAM and SOM estimate.
Step 2: Price the Failure Case Explicitly (The Two-Branch Value)
Damodaran's paper cites Bureau of Labor Statistics data showing that only 44% of businesses founded in 1998 lasted four years. Instead of burying that risk in the discount rate, value success first, then weight it:
Expected value = Going-concern value x (1 - Probability of failure) + Distress value x Probability of failure
In his worked example, a software company is worth $177.56 million as a going concern. With a 40% failure probability and no distress value, it is worth $106.54 million. That is 177.56 times 0.6.
What to do: assign a failure probability and a realistic salvage value.
Why it works: the investor sees the risk stated, not hidden.
Best for: pre-revenue and early-revenue companies.
Tie that probability to named triggers in your risk section.
Step 3: Run a Second Method and Explain the Gap (The Reconciliation)
Pick the method by stage. Before revenue, investors commonly use the venture capital method: estimate a value at exit, then discount it at a target return. Damodaran lists those targets at 50 to 70% for start-ups and 25 to 35% near an IPO. Once revenue is real, a discounted cash flow does the main work and a forward revenue multiple checks it.
The International Valuation Standards (2022 edition) tell professional valuers to reconcile differing results into one conclusion "without averaging," and to describe how.
Founders are not bound by them, but the discipline transfers.
What to do: write one paragraph explaining why the two numbers differ.
Why it works: the gap shows which assumption carries the value.
Best for: any company negotiating a priced round.
Present the outcome as a range in your pitch deck.
Where Do Investors Take a Valuation Model Apart?
Investors rarely dispute the arithmetic. They go after the two places where models hide optimism: risk counted twice, and a terminal value that quietly carries the whole result. Both are easy to find once you know to look, and both have a plain fix.
The Double-Counted Risk
The founder discounts at a venture-style target return and also applies a failure haircut, charging risk twice. Or the reverse: an ordinary discount rate on a forecast where everything works. Damodaran shows that venture targets already price in failure.
Early-stage venture investors in his data to 2007 earned 21.4% a year over twenty years, far below the 50 to 70% targets. The NBER survey adds that early-stage funds "may calculate IRRs from 'if successful' scenarios."
The fix: put failure risk in one place only. With a failure branch, discount the success branch at a rate for a company that survives, and say so in the notes.
The Terminal Takeover
Most of the value sits in the years after the forecast ends. Damodaran notes that terminal value can reach "90%, 100% or even more than 100%" of a young company's value. One generous growth rate then decides the result.
The fix: show the terminal year's share of value. Keep stable growth below the risk-free rate, as his worked example does, and let margins reach the mature twin's level first. Then run a sensitivity analysis on the two drivers that move value most.
Stage | Main method | Cross-check | What investors challenge |
Idea, no revenue | Venture capital method | Cost to recreate what exists | Exit value and target return |
Early revenue | Two-branch discounted cash flow | Forward revenue multiple | Failure probability, margin path |
Growing, near profit | Discounted cash flow | Peer multiples | Terminal growth, reinvestment |
Conclusion: Build the Model, Then Value It
A business valuation model earns its keep by making the argument visible. Drivers come from mature peers, the futures include failure, and a second method shows where the value sits. Built in that order, the number stops being a hope and becomes a range you can defend.
Drivers first, failure priced, a second method reconciled. A valuation an investor can take apart line by line is the one they can agree to.
Build the financial model first and value it second. Then test each scenario in PrometAI's valuation tools, which let you adjust assumptions across cases. You will know which one the investor will push on, and why it holds.
)
)
)
)
)
)
)
)
)
)
)