Valuation Methods for Startups

Professional valuators organize their work into three master approaches, codified in the International Valuation Standards (IVS, effective 31 January 2025) and the American Society of Appraisers Business Valuation Standards. The income approach values the business as the present value of expected future cash flows, with DCF as the primary tool. It is central to venture analysis because a venture's value lies almost entirely in future cash flows that do not yet exist. The market approach values the business by reference to what similar businesses have sold for, in public trading markets or completed transactions. It is observational: it tells you what buyers have paid, not what the asset is intrinsically worth. The cost approach values the business at the cost to reproduce or replace its underlying assets, relevant for asset-heavy or distressed businesses and rarely primary for operating companies with intangible value.

For entrepreneurial ventures, the income approach is primary because it captures future growth. The market approach provides a reality check on what investors are actually paying. The cost approach sets a floor, what the business is worth if broken up. Professionals use all three and triangulate.

FCFF measures the cash available to all capital providers, debt and equity, after the business funds operations and reinvestment. It is the cash the entire enterprise generates before any financing decisions.

Formula. FCFF = NOPAT + Depreciation & Amortization - Capital Expenditures - Increase in Net Working Capital, where NOPAT = EBIT x (1 - Tax Rate)

NOPAT strips out interest expense because FCFF is pre-financing; the cost of debt is captured in the discount rate, WACC, not in the cash flow. Discounting FCFF at WACC produces enterprise value, the value of the whole firm. To reach equity value, subtract net debt (total debt minus cash).

FCFE measures the cash available to equity holders specifically, after the business pays its debt obligations. It accounts for actual interest payments and net borrowing.

Formula. FCFE = Net Income + D&A - CapEx - Increase in NWC + Net Borrowing

Discount FCFE at the cost of equity, not WACC, and it yields equity value directly, with no need to subtract debt. Use FCFF when the capital structure will change significantly, when comparing companies with different leverage, or in standard investment-banking and M&A work. Use FCFE when the capital structure is stable, when valuing a single company with predictable debt service, or for banks and financial institutions.

Worked Example: An FCFF Calculation

A SaaS company reports EBIT of $5M, a 25% tax rate, D&A of $0.8M, CapEx of $1.2M, and a $0.3M year-over-year increase in net working capital. NOPAT is $5M x (1 - 0.25) = $3.75M. FCFF is $3.75M + $0.8M - $1.2M - $0.3M = $3.05M. If WACC is 10%, this single year's FCFF contributes $3.05M / 1.10 = $2.77M of present value to enterprise value.

Fama and French (1993) showed that CAPM's single market factor leaves substantial return variation unexplained. Adding a size premium (SMB, Small Minus Big) and a value premium (HML, High Minus Low) materially improves explanatory power across most equity samples tested. The magnitude varies by geography and period, but the three-factor model is the standard academic benchmark for expected returns.

Formula. E(R) = Risk-Free Rate + b1 x Market Premium + b2 x SMB + b3 x HML

For entrepreneurial finance, the size premium matters most. Small-capitalization firms have historically earned higher returns, compensating for higher risk, lower liquidity, and greater information asymmetry. When valuing a startup or small private company, adding a size premium of roughly 2 to 6% on top of CAPM produces a more realistic cost of equity. Treat that band as a practitioner range, not a precise constant.

Formula. WACC = (E/V) x Cost of Equity + (D/V) x Cost of Debt x (1 - Tax Rate), where V = E + D

The after-tax cost of debt reflects the interest tax shield: because interest is tax-deductible, the true cost of debt is lower than the stated rate. Typical WACC ranges, as practitioner rules of thumb, run 7 to 10% for large-cap stable companies, 10 to 14% for mid-cap growth, 15 to 25% for small private companies, and 30 to 70% for early-stage ventures, if DCF is used at all.

Worked example

Take the CAPM cost of equity from above, 10%, for a company financed 70% equity and 30% debt, with a pre-tax cost of debt of 8% and a 21% tax rate. The after-tax cost of debt is 8% × (1 − 0.21) = 6.3%. WACC is then 0.70 × 10% + 0.30 × 6.3% = 7.0% + 1.9% = 8.9%. Notice the debt contributes only 1.9 points, both because it is the smaller weight and because the tax shield lowers its effective cost from 8% to 6.3%.

WACC requires the market value of equity to compute E/V, but the market value of equity is exactly what the DCF is trying to solve for. This is a circular reference. Three solutions exist.

• First, iterative calculation: assume an initial capital structure, run the DCF, check whether the resulting equity value reproduces the same weights, and iterate to convergence, which modern spreadsheets automate.
• Second, a target capital structure: use management's target or the industry average, treating capital structure as a policy choice rather than a market observation.
• Third, Adjusted Present Value: Myers (1974) proposed valuing the firm as if all-equity financed, discounting FCFF at the unlevered cost of equity, then separately adding the present value of the interest tax shield. APV eliminates the circularity entirely because it does not use WACC.

Gordon (1962) treats post-forecast cash flows as a perpetuity growing at a constant rate forever.

Formula. Terminal Value = FCF_n x (1 + g) / (WACC - g), which requires g < WACC

FCF_n is the final year's free cash flow, g is the perpetuity growth rate, and WACC is the discount rate. The formula requires g below WACC; otherwise it produces a negative or infinite value, signaling internally inconsistent assumptions. This produces the terminal value as of Year n, so to incorporate it you discount it back to the present.

Formula. PV of Terminal Value = Terminal Value / (1 + WACC)^n, and Enterprise Value = Sum of PV(explicit FCFs) + PV of Terminal Value

Keep the two components visible; the split reveals how much of the valuation rests on near-term projections versus the perpetuity assumption. The perpetuity growth rate should approximate long-run nominal GDP growth, typically 2 to 3% for developed economies. A company rarely outgrows the broader economy for long, so using g above 4% is generally indefensible. Sensitivity matters enormously: with WACC of 10% and FCF of $10M, changing g from 2% to 3% raises terminal value from $127.5M to $147.1M, a 15% swing from one percentage point. As a rule, present terminal value in a sensitivity table across WACC and g.

Instead of assuming perpetual growth, apply a valuation multiple to the final year's metric, typically EBITDA or revenue.

Formula. Terminal Value = EBITDA_n x Exit Multiple

The exit multiple is usually derived from current trading multiples of comparable mature companies, on the logic that the firm will have matured by the end of the forecast. The advantage is that it is intuitive and anchored to observable data. The disadvantage is that it embeds current market conditions into a future exit: if today's multiples are inflated by a bubble, the terminal value inflates too. Best practice is to compute terminal value both ways and compare. If the Gordon result exceeds the exit-multiple result, the perpetuity growth may be too aggressive. If the exit multiple exceeds Gordon, the market may be pricing in optionality that a steady-state model misses, or WACC is too high.

Using IRR to rank compound returns implicitly assumes that all interim cash flows can be reinvested at the IRR itself. If a project's IRR is 35%, comparing on that basis assumes every dollar of interim cash earns 35% from receipt to the end of the horizon. The formula does not literally reinvest anything, but the moment you use IRR to rank projects or to equate an IRR with a realized compound return, you have made this assumption. It is rarely realistic. A distribution from an early exit does not automatically earn 35%; it sits in a money market at 4 to 5% or gets recycled into a new deal with its own risk. The higher the IRR, the more the assumption inflates the reported return relative to actual wealth accumulation.

When cash flows switch signs more than once, the IRR equation can produce several valid solutions. Consider an investment with cash flows of minus $100, plus $320, and minus $220. Solving the NPV equation yields two valid IRRs, 0% and 120%. Neither is wrong mathematically, but neither is economically meaningful. The project changes direction twice, which is the situation where a single IRR breaks down.

MIRR addresses both problems by requiring two explicit rates: a finance rate for funding negative cash flows (typically the cost of capital) and a reinvestment rate for positive interim cash flows (typically WACC or a conservative market return). MIRR compounds all inflows forward to the terminal date at the reinvestment rate, discounts all outflows back to time zero at the finance rate, then solves for the single rate connecting them. This yields a unique solution and reflects realistic reinvestment.

Formula. MIRR = (Terminal Value of Inflows / PV of Outflows)^(1/n) - 1

Worked Example: IRR Versus MIRR

A venture has cash flows over five years of minus $500K, then plus $200K, plus $200K, plus $200K, plus $200K, and plus $1,000K. Traditional IRR is 45.0%. Now assume interim distributions reinvest at 12%, not 45%. The terminal value of inflows at 12% is $200K x 1.574 + $200K x 1.405 + $200K x 1.254 + $200K x 1.120 + $1,000K = $2,070.6K. The present value of outflows is $500K at time zero. MIRR is ($2,070.6K / $500K) to the 1/5 power, minus 1, which is (4.141) to the 0.2 power, minus 1, or 32.9%. The MIRR is 12 points below the IRR because interim cash does not earn 45% after distribution. MIRR gives a truer picture of actual wealth accumulation.

A standard DCF applies a single WACC to every projected year, assuming the business carries the same risk in Year 1 as in Year 10. For mature, stable companies this is a reasonable simplification. For startups and high-growth ventures it is demonstrably false. A pre-revenue startup faces binary risk: will the product work, will customers pay, will the team execute? Each milestone (product-market fit, first revenue, positive unit economics, a repeatable sales process) tends to remove a category of risk for good. A company past $5M ARR with healthy retention is categorically less risky than the same company at $100K ARR, even at the same growth rate. The discount rate should reflect this.

Empirical evidence from Bhagat (2014) shows venture capitalists implicitly use declining discount rates through their target return multiples by stage.

Instead of discounting every cash flow at one WACC, assign a different rate to each year based on the expected risk at that point. The present value of each year's cash flow becomes the cash flow divided by the product of one plus each year's rate.

Formula. PV(CF_n) = CF_n / [(1 + r1)(1 + r2)(1 + r3) ... (1 + rn)]

For a startup expected to be pre-revenue in Year 1 (50%), reach product-market fit in Year 2 (40%), hit $1M ARR in Year 3 (30%), reach $5M ARR in Year 4 (20%), and be acquisition-ready in Year 5 (15%), the cumulative discount factors are 0.667, 0.476, 0.366, 0.305, and 0.265. Compare a flat 30% WACC: the Year 5 factor would be 1 / (1.30)^5 = 0.269, close to the stage-adjusted 0.265, but the Year 1 factors differ sharply (0.667 versus 0.769), reflecting the concentrated risk in the earliest period. Stage adjustment matters most for ventures with high early binary risk, such as biotech and deep tech, and for long forecast horizons, where applying a seed-stage rate to Year 10 makes those cash flows nearly worthless and understates value for a company that will clearly have de-risked. It also gives founders negotiating leverage: if the company has de-risked since the last round, the appropriate rate for the next round should be materially lower.

A discounted-cash-flow value rests on a hidden assumption: that management commits to a single plan today and follows it regardless of what happens next. Real ventures do not behave that way. Founders and investors can expand a promising line, delay a bet until the picture clears, or abandon a plan that is failing. That flexibility has value, and a static DCF omits it, which is one reason a mechanical DCF tends to understate the worth of an early-stage venture.

The intuition splits cleanly in two directions:

• The option to abandon caps the downside. If a project sours, management can stop funding it and walk away, so the loss is bounded rather than open-ended.
• The option to expand captures the upside. If early results are strong, management can pour in more capital and scale, capturing gains a fixed plan would leave on the table.

Formal valuation prices this flexibility with option-pricing methods, generally a binomial lattice or a Black-Scholes framework, treating the decision to continue as a call option on the venture's future value. The mechanics matter less than the core insight: flexibility is worth more when uncertainty is higher. When outcomes are nearly certain, the right to change course is close to worthless because there is little to react to. When the range of outcomes is wide, the ability to expand into good news and abandon bad news is generally where much of the value sits.

Venture investors rarely fund an entire plan in one check. Instead they stage capital across rounds, seed, then Series A, then Series B and beyond, with each round tied to milestones the company is expected to hit. This structure is itself a real option. Each round buys the right, not the obligation, to continue. A disappointing milestone lets the investor decline the next round and cap losses, while a strong result justifies a larger follow-on at a higher valuation.

A brief numeric intuition makes the point. Suppose a plan needs $2,000,000 in total.

• Fund it all up front. Committing the full $2,000,000 at once puts the entire amount at risk if the company fails early.
• Stage it. Investing $500,000 now with the remaining $1,500,000 contingent on hitting a defined milestone limits the early loss to $500,000 if the venture stalls, while preserving the upside of continuing if the milestone is met.

The staged path bounds the downside to the money already deployed and keeps the option to scale into good news, which mirrors the abandon-and-expand logic of real options. Staging is not free of tradeoffs. It also imposes discipline on the founders, who must deliver against milestones to unlock further capital, and it shifts the timing of dilution, since later rounds are priced on results the company has since produced rather than on the original plan.

Precedent transactions analyze multiples paid in completed M&A deals involving similar companies. They differ from trading comps in one critical respect: they include a control premium, typically 20 to 40% above the trading price, that buyers pay for the right to control strategy, capital allocation, and operations. Use precedent transactions when valuing a company for an acquisition, since you are buying control, and trading comps when valuing a minority stake, since you are not.

The quality of a comparable analysis depends heavily on the quality of the comparable set. Selection should match:

• industry and business model,
• revenue scale,
• growth rate,
• margin profile, and
• geography and end-market.

A $10M revenue company is not comparable to a $500M one, and a 50% grower should not be comped to a 10% grower without adjustment. The critical limitation is that comps tell you what the market pays, not what something is worth. During the 2021 low-rate bubble, SaaS companies traded at 30 to 50x revenue, multiples that reflected euphoria, not intrinsic value. An analyst who used 2021 comps to justify a valuation was anchoring to a distorted market. As a rule of thumb, cross-check comps against a DCF to test whether the implied growth rate is achievable.

The basic method assumes the investor's entry ownership equals their exit ownership. In reality, later rounds, option pool expansions, and anti-dilution provisions reduce ownership before exit. Sophisticated investors adjust with a retention ratio, the fraction of ownership they expect to retain through exit.

Formula. Required Current Ownership = Required Final Ownership / Retention Ratio

If a VC needs 10% at exit and expects 30% cumulative dilution (a 70% retention ratio), they need 10% / 0.70 = 14.3% today, not 10%.

VCs typically require an option pool refresh created before their investment, out of the pre-money, diluting existing shareholders but not the new investor. A 15% pre-money pool means founders give away 15% before the VC's dollars create further dilution. It is a negotiation lever disguised as standard practice: the larger the pre-money pool, the lower the effective pre-money from the founders' perspective. Anti-dilution protection gives current investors extra shares in a down round, full-ratchet or weighted-average, diluting founders further. Participation rights give participating preferred holders their liquidation preference plus a pro-rata share of remaining proceeds, which does not dilute ownership but dilutes the economic value per common share. Pro-rata rights let an investor maintain ownership by investing in future rounds, improving the retention ratio at the cost of more capital.

VCs do not use WACC for three reasons. Portfolio math demands it: if 6 of 10 investments fail completely, the 4 winners need to return the entire fund plus target profit, so a 3x gross fund return requires winners to average 7 to 8x. CAPM inputs generally do not exist for pre-revenue firms: there is rarely an observable beta, a stable capital structure, or a meaningful cost of debt. And the discount rate embeds failure probability implicitly: a 50% required return does not mean the VC expects 50% on each deal, but that across the portfolio, including zeros, the blended return needs to reach the fund's target.

The cost approach values a business at the cost to reproduce or replace its net assets. It is the least common primary method in entrepreneurial finance but serves three roles. As a floor, it sets the minimum value in any negotiation: a company with $10M in net assets is generally not worth less than $10M unless it is burning cash faster than it can be liquidated. For asset-heavy businesses, such as real estate holding companies, equipment leasing, and natural resources, value is in the assets, not operating cash flows. In distressed situations, where cash flows are negative and the going-concern assumption is questionable, the relevant question becomes what the parts are worth separately.

Three methods sit within the cost approach. The adjusted net asset method restates all assets and liabilities to fair market value, not book value, and computes equity as adjusted assets minus adjusted liabilities. Replacement cost asks what it would cost to build the business from scratch today, including physical assets, assembled workforce, customer relationships, and technology, which is useful for IP-heavy companies. Liquidation value asks what a quick, distressed sale would yield, typically 50 to 70% of fair market value for tangible assets and near zero for most intangibles. The approach fails for most startups: a SaaS company with $500K in laptops and servers, $2M in capitalized development, and $50M in annual recurring revenue is worth far more than its asset base suggests. The cost approach captures little of the value in customer relationships, growth trajectory, network effects, or brand. For knowledge businesses it is a necessary floor and rarely the primary valuation.

When future outcomes are uncertain, not merely volatile around a trend, a single-point DCF produces false precision. Probability-weighted scenario analysis models multiple futures, assigns probabilities, and computes an expected value. It is more honest and avoids optimizing around a single forecast that may never materialize. A standard three-scenario structure has a downside (20 to 30% probability, growth stalls or execution falters, value a fraction of the base case), a base case (40 to 50%, the operating plan is achieved), and an upside (20 to 30%, category tailwinds or successful adjacent-market entry, value 1.5 to 3x the base). For early-stage ventures, add a failure scenario (30 to 50% at seed, 15 to 25% at Series A) where value is zero or recovery of remaining cash.

Worked Example: A Series A Expected Value

The expected value is (0.20 x $0) + (0.25 x $15M) + (0.35 x $50M) + (0.20 x $120M) = $0 + $3.75M + $17.5M + $24M = $45.25M. This can be compared to a VC-method valuation. If the VC method (at a 40% target return) produces a $40M post-money, the probability-weighted approach suggests the deal is slightly favorable to the investor, paying $40M for a $45.25M expected value. If the VC method produces $60M, the deal may be overpriced relative to the outcome distribution.

Use probability-weighted valuation for binary outcomes, such as biotech drug approval or platform network effects; for wide confidence intervals, when reasonable bull and bear projections differ by 3x or more; as a negotiation tool, since presenting explicit scenarios demonstrates rigor and grounds the discussion in assumptions that can be debated individually; and for board-level capital allocation, where comparing expected values across scenarios beats comparing single-point NPVs.

Professional valuators rarely rely on a single method. They compute value several ways and present a range, and the degree of convergence is itself informative. Tight convergence, all methods within 15 to 20%, signals high confidence that the asset is priced consistently regardless of the analytical lens. Wide divergence, methods differing by 50% or more, signals that one or more sets of assumptions is likely wrong, or that the company is in a transitional state where different methods capture different aspects of value.

Investment banks present valuations as horizontal bar charts, football fields, showing the range from each method: a DCF range across WACC and terminal-growth assumptions, a trading-comps range from the 25th to 75th percentile, a precedent-transactions range, a VC-method or target-return analysis, and a 52-week trading range if public. The negotiated price almost always falls in the overlap zone where multiple methods agree. If a buyer proposes a price below every method's range, the seller has strong analytical ammunition to push back. If the seller demands a price above every range, they need a compelling strategic rationale to justify the premium.

A method is only as good as its inputs. A DCF with fabricated growth produces a fabricated value; comps with poor comparables produce misleading multiples; the VC method with an unrealistic exit produces a fantasy. The analytical skill is not running the formulas but selecting defensible inputs and pressure-testing them against reality. Before accepting any valuation, ask three questions. What growth rate is implied, and has this company or any comparable ever sustained it? What discount rate was used, and does it reflect the actual risk of this specific business at this specific stage? What exit assumption drives the terminal value, and is there observable evidence, recent transactions or public comps, supporting it? If all three answers are grounded in evidence, the valuation is defensible. If any requires trust me or the market will get there, the valuation is speculation dressed in financial modeling.

Each method converts an uncertain future into a present price, and each tends to bury some risks in a single number. The output is only as good as the inputs and the judgment behind them.

The discount rate is the most consequential and least observable input. For private and pre-revenue firms, beta, the equity risk premium, and the market value of equity are estimates, and small errors move value by tens of percent.

Terminal value drives most of a DCF, yet it rests on the least knowable assumptions about the distant future. Present it as a sensitivity range over WACC and g.

The benchmarks and ranges used here, including WACC bands, terminal value of 60 to 80%, 2 to 3% perpetuity growth, a 20 to 40% control premium, a 2 to 6% size premium, and VC target returns by stage, are practitioner heuristics and observed ranges, not laws, and they shift with the market and the cycle.

Time-sensitive figures, including a risk-free rate near 4.3% and Damodaran's implied equity risk premium of 4.23% as of January 2026, will change. Re-source them at the time of any real valuation.

The final discipline is not running the formulas but selecting defensible inputs and pressure-testing them against observable reality. A valuation that requires trust rather than evidence is speculation dressed in financial modeling.