DCF Valuation: The Complete Walkthrough (with a Worked SaaS Example)

A discounted cash flow valuation (DCF) is the process of estimating what a business is worth today by projecting the cash it will produce in the future and discounting those future cash flows back to the present at an appropriate rate. The output is a single number: the intrinsic value per share, or the enterprise value of the business.

DCF is the workhorse of corporate finance. Every investment bank uses it. Every private equity fund uses it. Most thoughtful public-market investors use it as one of several methods. Most acquisition decisions go through some version of a DCF.

DCF is also widely misunderstood. The mechanical steps are simple. The judgment calls — what discount rate, what growth rate, what terminal value — are where most of the value is destroyed or created. A DCF model is only as good as its assumptions, and the assumptions are usually where people are loosest.

This guide covers what DCF actually is and when to use it, building a free cash flow forecast, calculating the discount rate (WACC), the terminal value problem (Gordon Growth vs exit multiple), the sensitivity analysis that reveals whether your valuation is robust or fragile, a complete worked example valuing a SaaS company, the most common DCF mistakes, when to use DCF vs other methods, and the FAQ.

What DCF actually is, and when to use it

The intuition: a dollar today is worth more than a dollar a year from now. The dollar today can be invested at some rate of return. To compare cash flows at different points in time, you have to bring them to the same point in time.

The DCF formula says the value of an investment is the sum of all future cash flows, each discounted to present value:

$V={\sum }_{t=1}^{\infty }\frac{C{F}_t}{(1+r{)}^t}$

where $C{F}_t$ is the cash flow in period $t$ and $r$ is the discount rate.

In practice, you cannot forecast cash flows out to infinity. So you break the calculation into two parts: an explicit forecast period (usually 5–10 years where you project cash flows year by year) and a terminal value that captures everything after the explicit period.

$V={\sum }_{t=1}^N\frac{C{F}_t}{(1+r{)}^t}+\frac{TV}{(1+r{)}^N}$

where $TV$ is the terminal value at year $N$.

The two big judgment calls embedded in this formula are how to forecast the cash flows (operational understanding) and what discount rate to use (capital markets understanding). Most DCF errors are errors in one of these two.

When DCF works well. Mature businesses with predictable cash flows: utilities, large consumer-goods companies, telecoms, infrastructure. The forecast can be grounded in years of historical data.

When DCF works poorly. Early-stage companies with no profits and no track record: most startups. Cyclical businesses near a turning point. Businesses undergoing major transformation. In these cases, the cash flow forecast is a guess, and the valuation is fragile.

When DCF works but you should be careful. Growth-stage companies (the SaaS example below) — there is real revenue and real economics, but most of the value sits in the terminal period, which makes the valuation sensitive to long-run assumptions. Use sensitivity analysis aggressively.

Building a free cash flow forecast

The cash flow you forecast in a DCF is free cash flow to the firm (FCFF), sometimes called unlevered free cash flow. This is the cash available to all capital providers — debt holders and equity holders combined — after the business has paid for operations and reinvestment.

The standard formula:

$\text{FCFF}=\text{EBIT}\times (1-t)+\text{D\&A}-\text{CapEx}-\Delta \text{NWC}$

where:

• EBIT = Earnings Before Interest and Taxes (operating profit)
• $(1-t)$ = applied because EBIT is pre-tax; we want after-tax operating profit
• D&A = Depreciation and Amortization (added back because it's a non-cash expense)
• CapEx = Capital Expenditure (subtracted because cash actually goes out)
• $\Delta$NWC = change in Net Working Capital (subtracted because growth ties up cash)

The forecast period is typically 5 to 10 years. For each year, you project each line:

1. Revenue: based on market size, growth rate assumptions, and competitive position.
2. EBIT margin: based on historical margins, expected operating leverage, scale economies.
3. Tax rate: usually the marginal corporate tax rate.
4. D&A: usually as a percentage of revenue or as a function of historical capital base.
5. CapEx: as a percentage of revenue or maintenance + growth components.
6. NWC: usually as a percentage of revenue (receivables, payables, inventory).

The forecast is where operational judgment shows up. A bad analyst plugs in arbitrary numbers; a good analyst grounds each line in business reality.

Detail on FCFF computation: Free Cash Flow: Calculation and Interpretation.

Calculating WACC

The discount rate in a standard DCF is the Weighted Average Cost of Capital (WACC). This represents the blended cost of all the company's financing — debt and equity weighted by their share in the capital structure.

$\text{WACC}=\frac{E}{V}\times r_e+\frac{D}{V}\times r_d\times (1-t)$

where:

• $E$ = market value of equity
• $D$ = market value of debt
• $V=E+D$ = total firm value
• $r_e$ = cost of equity (typically estimated by CAPM)
• $r_d$ = cost of debt (typically the yield on existing debt)
• $t$ = tax rate (multiplied because interest is tax-deductible)

The cost of equity $r_e$ is the hardest piece. The standard formula is the Capital Asset Pricing Model (CAPM):

$r_e=r_f+\beta \times (r_m-r_f)$

where $r_f$ is the risk-free rate (typically the 10-year Treasury yield), $\beta$ is the company's systematic risk relative to the market, and $(r_m-r_f)$ is the equity risk premium (historically about 5–6% in the US).

Detail: WACC Explained: Cost of Capital Step-by-Step.

For our SaaS example below, we will use a WACC of 9.5%, which is reasonable for a mid-stage growth SaaS company in the current rate environment.

The terminal value problem

The terminal value is everything beyond your explicit forecast period. For most DCFs, terminal value accounts for 60–80% of the total valuation. It is the largest single judgment in the model.

Two methods are standard:

Gordon Growth (perpetuity): assumes the business grows at a constant rate forever.

$TV=\frac{FC{F}_{N+1}}{r-g}$

where $FC{F}_{N+1}$ is the first year of cash flow beyond the explicit period, $r$ is WACC, and $g$ is the perpetuity growth rate.

The growth rate $g$ must be lower than long-run economic growth (typically 2–3% in developed markets). If $g\ge r$, the formula breaks down (you get negative or infinite values). The intuition: no business can grow faster than the economy forever.

Exit Multiple: assumes the business will be valued at some multiple of its terminal-year metric (EBITDA, EBIT, revenue).

$TV=\text{Terminal Metric}\times \text{Multiple}$

Common multiples: EV/EBITDA, EV/EBIT, EV/Revenue. The multiple is usually drawn from current market comparables, with adjustments for expected maturity at exit.

Which to use. Best practice: use both, and reconcile. They check each other. If they disagree wildly, you have a problem in one of them.

Sensitivity analysis

A single-number DCF is almost always wrong. The right output is a range. Sensitivity analysis shows how the valuation changes when you change key assumptions.

Standard sensitivities for any DCF:

• WACC: vary ±100 basis points
• Terminal growth rate (or exit multiple): vary ±100 basis points (or ±2 turns of multiple)
• Revenue growth in forecast period: vary ±5 percentage points
• EBIT margin in forecast period: vary ±200 basis points

Build a two-way table for WACC × terminal growth: this is where most of the value sensitivity lives.

A robust DCF has its range — across reasonable sensitivity values — concentrated around a single rough number. A fragile DCF has its range scattered, with the valuation tripling or zeroing out under small changes in assumptions. Always show the range.

A worked example: valuing a SaaS company

Let's value "Atlas Software" — a fictional but realistic mid-stage B2B SaaS company. Year 0 (now) data:

• Revenue: $50 million
• Revenue growth: 35% per year (slowing over time)
• EBIT margin: 5% currently (improving with scale)
• Tax rate: 25%
• D&A: 4% of revenue
• CapEx: 8% of revenue (high because investing in growth)
• NWC: 5% of revenue (typical SaaS, low NWC)
• WACC: 9.5%

Step 1 — Project revenue.

Step 2 — Project EBIT margin (expanding with scale, plateauing around year 7–8 at 25%):

Step 3 — Compute FCFF for each year.

For year 5 (illustrative):

• EBIT × (1 − tax) = $30.0\times 0.75=$22.5M
• D&A = 4% × $158.1=$6.3M
• CapEx = 8% × $158.1=$12.6M
• $\Delta$NWC = 5% × ($158.1-$134.0) = $1.2M
• FCFF = 22.5 + 6.3 − 12.6 − 1.2 = $15.0M

Doing this for each year:

(These FCFF numbers shown above for year 5 are illustrative; the exact pattern depends on the precise sequencing of NWC, CapEx, and margin assumptions. The Excel model gives the line-by-line computation.)

Step 4 — Compute terminal value.

Using Gordon Growth with $g=3\%$ (a reasonable perpetuity rate for a mature SaaS):

$TV=\frac{FCF{F}_{11}}{r-g}=\frac{51.6\times 1.03}{0.095-0.03}=\frac{53.1}{0.065}=\$817M$

Cross-check with exit multiple. Mature public SaaS companies trade around 15× EBIT. Year 10 EBIT is $64.1M.ExitmultipleTV=$64.1 × 15 = $962M.Thetwoarereasonablyclose($817M vs $962M).Forourvaluation,usetheaverage:$890M.

Step 5 — Discount everything to present value.

Sum of explicit-period PV: about $135M.

PV of terminal value: $890M\times 0.404=$360M.

Enterprise value (EV): $135M+$360M = $495M.

Step 6 — Adjust to equity value.

Equity value = EV − Net Debt.

Assuming Atlas has $10Mofdebtand$20M of cash, Net Debt = −$10M.

Equity value = $495M-(-$10M) = $505M.

If Atlas has 50 million shares outstanding, value per share = $505M/50M=\ast \ast$10.10**.

Step 7 — Sensitivity analysis.

Vary WACC and terminal growth rate. The two-way table (EV in $M):

Valuation range: roughly $420Mto$615M. The base case ($495M) sits comfortably in the middle. If you're confident in the assumptions, the value is in this range. If you want to buy the company, you want to pay near the bottom; if you want to sell, you want to be near the top.

Step 8 — Sanity check against comparables.

Are there mid-stage SaaS companies at similar growth and margin profiles trading on public markets or in recent transactions? If the trailing or forward EV/Revenue multiples on those comps imply a value far from our DCF, one of them is wrong — and it might be the DCF.

For Atlas Year 0 revenue of $50M,mid-stageSaaScompsmighttradeat6\times to10\times forwardrevenue,implyingEVof$300M to $500M.OurDCFresultof$495M is at the top of this range. Reasonable.

This is what a DCF actually looks like when done seriously. The Excel model linked at the top of this article contains all of these calculations with every formula visible.

Common DCF mistakes

After fifteen years of seeing DCFs from analysts, the same handful of mistakes recur.

Mistake 1 — Overly optimistic revenue growth.

Forecasting 30%+ growth for ten years compounds to absurd revenue numbers. Most businesses cannot sustain that. Check: at the end of the forecast period, what share of the total market does the business have? If it's implausibly large, growth assumptions are too high.

Mistake 2 — Margin expansion without justification.

Assuming EBIT margins expand from 10% to 35% over five years is a model decision, not an observation. The expansion has to be grounded in something specific — operating leverage at scale, mix shift to higher-margin products, structural cost reduction. Hand-waving "scale" is not enough.

Mistake 3 — Terminal value is most of the value.

If 90% of your valuation is the terminal value, your DCF is barely a forecast — it's mostly a guess about the long run. Either extend the forecast period or accept that the model is fragile.

Mistake 4 — Using book values instead of market values for WACC weights.

The weights $E/V$ and $D/V$ should be at market values. Using book equity (which can be far below market equity) systematically underestimates the cost of equity weighting and biases WACC downward.

Mistake 5 — Inconsistent treatment of inflation.

If your cash flows are nominal (include inflation), your discount rate must be nominal. If real, real. Mixing is a sneaky error.

Mistake 6 — Forgetting working capital and CapEx for growth.

A growing business needs working capital. A growing business needs CapEx. Forecasting EBIT margins to expand while assuming working capital and CapEx don't grow is internally inconsistent.

Mistake 7 — Using sensitivity analysis to confirm a target price.

A common abuse: choose a target price, then back-solve for assumptions that produce it, then label the result a "sensitivity range." Real sensitivity analysis starts with reasonable assumption ranges and shows what the implied valuation range is. The two are not the same.

DCF vs other valuation methods

DCF is one of three main approaches. The others are:

Comparables (trading multiples). Look at how similar public companies are valued (EV/EBITDA, EV/Revenue, P/E). Multiply the target's metric by the comparable multiple.

• Strengths: fast, market-grounded, hard to manipulate.
• Weaknesses: only works if good comparables exist; doesn't tell you what the right multiple should be, only what the market currently uses.

Precedent transactions. Look at multiples paid in recent M&A transactions for similar companies.

• Strengths: includes control premium, real prices paid.
• Weaknesses: transactions are infrequent; market conditions change.

The right practice: use all three. Each has weaknesses. Triangulating across them gives a more defensible valuation than any one alone.

In an investment banking pitch book, you will typically see four to six different valuation methodologies, with the result of each shown as a bar on a "football field" chart. The bars usually overlap; the overlap is the defensible range.

DCF's particular strength is that it forces you to be explicit about the operating assumptions. Comparables give you a number; DCF gives you a theory of why the number is what it is. Both are useful.