The capstone of the valuation arc. Module 04 produced the cash flows. Module 05 produced the discount rate. Module 06 produced the multiples cross-check. This module brings them together: a complete DCF valuation with terminal value, sensitivity analysis, and triangulation. By the end, you can value any public company with discipline.
Download the Excel toolkit (complete DCF model with sensitivity)The discounted cash flow method values a business as the present value of all the cash it will generate over its remaining economic life. It is, theoretically, the most rigorous valuation methodology in finance: every other approach — multiples, asset-based, transaction-comparison — either approximates a DCF or extracts a single piece of it. When you do a DCF correctly, you are answering the question "what is this stream of future cash flows worth today, given the time value of money and the risk of those cash flows?"
Three principles make DCF distinctive:
Modules 04, 05, and 06 each produced a piece of the puzzle. Module 04 built the projection — what cash flows the firm will produce. Module 05 produced the discount rate — what return investors require given the risk. Module 06 produced the multiples cross-check — what comparable firms are actually trading for. None of them, alone, gives you a complete valuation. The DCF combines projection and discount rate into a single number; multiples then verify the answer.
Stripped to essentials, the DCF says:
Where FCFFt is the firm's free cash flow in year t, and WACC is the discount rate. Each future cash flow is divided by (1 + WACC)t, which translates it back to present-value terms. Sum across all future years to get the firm's enterprise value.
The challenge: you can't actually project cash flows to infinity. So we split the calculation into two stages — an explicit projection of the next 5 years (Module 04's domain), plus a "terminal value" capturing everything beyond. That two-stage structure is what makes DCF practical. Section 02 covers it in detail.
By the end of this module you'll be able to compute a complete enterprise-DCF valuation for any public company. The mechanics:
The Excel toolkit accompanying this module is a complete DCF model — pulls the WACC build from Module 05, the FCFF projection from Module 04, applies both terminal-value methods, runs sensitivity analysis, and triangulates against multiples. Use it as a template, replace the inputs with your own company, and you have a defensible valuation.
In principle, DCF discounts cash flows from now until the firm's hypothetical end. In practice, you can't project ten or twenty years out with any precision — and beyond that, projecting individual years is essentially fiction. The standard solution: split the valuation into two stages, the explicit period and the terminal value.
Year-by-year projection of free cash flow for a defined window — typically 5 years. Built using Module 04's three-statement model. Each year's FCFF discounted back to today individually.
A single number representing the present value of all cash flows beyond Year 5, computed using either the Gordon Growth formula or an exit multiple, then discounted back.
The split exists because the two stages have very different forecasting characteristics:
The convention of using a 5-year explicit period reflects a balance: long enough that the firm reaches a quasi-steady state by the end, short enough that each year is genuinely projectable. Some industries (pharma with multi-decade patents, utilities with regulated cycles) warrant 10-year explicit periods. Some (early-stage tech) require explicit periods until the firm reaches profitability, even if that's 7-8 years out. The 5-year default is a starting point, not a rule.
Here's a fact that surprises most first-time DCF builders: the terminal value typically accounts for 60-80% of total enterprise value. The explicit period — five years of careful projections — usually contributes only 20-40% of the answer. The "tail" beyond Year 5 dominates.
| Firm type | Typical TV % of EV | Why |
|---|---|---|
| High-growth tech | 80-95% | Most cash flow is far in the future; explicit period FCF is small or negative |
| Mature industrial | 65-75% | Steady cash generation across the explicit period; terminal-value tail is meaningful but not dominant |
| Declining business | 40-60% | Explicit period captures more of the remaining value; terminal value reflects a smaller, stable rump |
| Cyclical commodity | 50-70% | Explicit period must capture cycle dynamics; terminal value reflects mid-cycle steady-state |
Two consequences flow from terminal-value dominance:
First, terminal-value assumptions matter enormously. A 50bp change in terminal growth rate can move the entire valuation by 10-15%. Always run sensitivity on the terminal-value driver (Section 06).
Second, defending a DCF means defending the terminal-value assumption. In any presentation or valuation discussion, expect the question "why did you assume that terminal growth rate?" or "why that exit multiple?" The answer needs to be more than "it seemed reasonable." It should reference industry growth rates, GDP growth expectations, and the firm's specific positioning.
The fact that terminal value dominates the valuation isn't a defect of the DCF method — it reflects an underlying truth about the businesses being valued. Most ongoing businesses generate the bulk of their value in the long run, not the next 5 years. A DCF that produced a 95% explicit-period contribution would be saying "this firm has effectively no value beyond Year 5," which would only be true for a firm in terminal decline. The terminal-value share is a feature of the method matching reality, not a flaw to engineer around.
The mechanics of discounting the explicit period are straightforward: take each year's FCFF, divide by (1 + WACC)t, and sum the results. This converts a stream of future cash flows into a single present-value number.
That's it for the math. Each year's FCFF gets a discount factor of 1 ÷ (1 + WACC)t, which shrinks more aggressively as t grows. Year-1 cash is barely discounted; Year-5 cash gets discounted by a factor of about (1+9.27%)5 = 1.557, so its PV is about 64% of its nominal value at a 9.27% WACC. Year-10 cash would be discounted by about 2.42x — its PV is only 41% of nominal.
One mechanical question: when in each year does the cash flow occur? Two conventions exist:
The toolkit uses end-of-year by default. Switch to mid-year by changing the discount-period column from {1, 2, 3, 4, 5} to {0.5, 1.5, 2.5, 3.5, 4.5}. For most practical purposes, the choice is a matter of convention rather than precision — the variation across analysts who choose different conventions is smaller than the variation from terminal-value or growth-rate assumptions.
Recall Sample Company from Modules 04-06. Module 04's three-statement model produces the FCFF projection for years 1-5; Module 05's WACC of 9.27% is the discount rate. Putting them together using end-of-year convention:
| Year | FCFF ($M) | Discount period | Discount factor | PV ($M) |
|---|---|---|---|---|
| Year 1 | 62 | 1 | 0.9152 | 57 |
| Year 2 | 88 | 2 | 0.8377 | 74 |
| Year 3 | 110 | 3 | 0.7666 | 84 |
| Year 4 | 118 | 4 | 0.7016 | 83 |
| Year 5 | 125 | 5 | 0.6421 | 80 |
| Total: PV of explicit period | 503 | — | — | 377 |
The undiscounted FCFF series sums to $503M. After discounting each year back to today at 9.27%, the PV totals $377M. The discount machinery shrinks the nominal $503M by about 25% — that's the time-value penalty for the average 3-year wait to receive the cash.
This $377M is just the explicit-period contribution. The terminal-value portion — capturing all cash flows beyond Year 5 — is the next step.
The terminal value captures the present value of all cash flows beyond the explicit-period horizon. Two methods dominate practice — both have sound logic, both have weaknesses, and the right approach is usually to compute both and reconcile.
The Gordon Growth method assumes the firm grows at a constant rate forever after the explicit period ends. The formula:
Where FCFFY5 is the final explicit-period cash flow, g is the perpetual growth rate, and WACC is the discount rate. The numerator (FCFF × (1+g)) is the Year-6 cash flow; dividing by (WACC − g) capitalizes that into a perpetuity. The result is the terminal value at the end of Year 5; discount back by (1+WACC)5 to get the PV today.
The terminal growth rate g is the most consequential input in the entire DCF. The standard discipline:
The exit multiple method assumes that, at the end of the explicit period, the firm could be sold at a price reflecting the EV/EBITDA (or other) multiple that prevails for comparable firms. The formula:
Where the exit multiple is the trading multiple (typically EV/EBITDA) of comparable firms, applied to the projected Year-5 EBITDA. The result is the terminal value at end of Year 5; discount back by (1+WACC)5 to get PV today.
The exit multiple is the key input. The standard practice: use the median trading multiple of comparable firms (Module 06's domain), with two adjustments:
The Gordon Growth and Exit Multiple methods make different assumptions and have different weaknesses:
| Aspect | Gordon Growth | Exit Multiple |
|---|---|---|
| Theoretical foundation | Cleanly grounded in DCF logic — perpetuity formula derived from first principles | Empirically grounded — uses observed market prices for comparable firms |
| Key weakness | Tiny changes in g produce large changes in TV. Highly sensitive. | Implicitly assumes future multiples will look like today's. May not hold. |
| Best for | Mature firms reaching genuine steady-state by Y5 | Firms where comparable transactions are abundant and informative |
| Common practitioner approach | Compute both. Use Gordon Growth as primary; Exit Multiple as cross-check. | If they diverge by more than 15-20%, investigate the source. |
Continuing the Sample Company DCF: Year-5 FCFF is $125M, Year-5 EBITDA is $267M. Using a 3% perpetual growth rate (Gordon Growth) or a 9× exit EV/EBITDA multiple:
The two methods produce results within 17% of each other — which is reasonable agreement. The toolkit defaults to using Gordon Growth as the primary method, producing a $1,322M PV of terminal value. (The small difference from $1,319M above reflects exact precision in the toolkit vs. rounded numbers in this worked example.) The Exit Multiple result of $1,541M serves as the cross-check; if the two methods diverged by 30%+ the assumptions would warrant re-examination.
Notice that the PV of the terminal value ($1,322M) is roughly 3.5× the PV of the explicit period ($377M). This is the "TV tail" effect referenced in Section 02 — most of the firm's value comes from beyond Year 5. The terminal-value assumption is the single most consequential lever in the model. Always run sensitivity on it, and always be prepared to defend the terminal-growth or exit-multiple choice you make.
The toolkit includes a fully working DCF model with seven tabs: WACC build (Module 05), 5-year FCFF projection, terminal value (both methods, side-by-side), DCF valuation with bridges, sensitivity table (WACC × terminal growth), and multiples triangulation. Everything traces back to editable inputs. Drop in your own company's numbers and you have a defensible valuation.
Download toolkit (.xlsx)The DCF produces an enterprise value — the value of the entire business operation. To get to a per-share value that an equity investor would care about, three bridges connect EV to equity value to share price:
Each bridge requires a careful judgment call:
Subtract debt because debt is a senior claim that must be paid before equity holders see anything. Add cash because excess cash is technically owned by equity holders — it's not part of the operating business that the DCF valued. The mechanical version is "subtract net debt" (where net debt = debt − cash).
What "debt" to use:
Divide equity value by the diluted share count to get implied per-share value. The diluted count includes:
Most public companies report diluted share counts in their 10-Q/10-K filings. Use the diluted count, not the basic count — basic underestimates effective ownership because it ignores all the contingent equity claims that exist.
Putting the full DCF together:
This is the complete DCF answer: $32 per share (rounded) for Sample Company. If the stock is currently trading at $28, the DCF says it's undervalued. If at $40, the DCF says it's overvalued. Context matters: a 10-15% gap between DCF and market price is normal noise; a 30%+ gap should prompt deeper investigation of either the assumptions or the market's positioning.
Compare to the multiples-implied valuations from the toolkit's Triangulation tab: EV/EBITDA at 10.5× implies $1,904M EV; EV/EBIT at 14× implies $1,813M; P/E at 18× implies $1,748M EV-equivalent. The DCF's $1,700M sits at the low end of this range, with the multiples cluster centered around $1,820M (~$36/share). The 7-15% gap between DCF and multiples is in the comfortable agreement zone — minor differences in growth and margin embedded in different methods, not contradictions.
A DCF point estimate is fragile. Change the WACC by 50bp, the terminal growth rate by 50bp, or the exit multiple by 1×, and the answer can shift 10-20%. The skilled practitioner doesn't present a single number; they present a range, with explicit acknowledgment of which inputs drive that range.
For most DCFs, two inputs dominate the sensitivity:
The standard sensitivity output is a 2D table flexing these two drivers. From the toolkit, varying WACC from 8.0% to 10.5% and terminal growth from 1.5% to 4.5% produces an EV range from roughly $1.2B to $2.9B — a 2.4× spread. This is normal for a DCF; the discipline is presenting the range honestly rather than picking a flattering point estimate.
Beyond input sensitivity, a complete valuation triangulates across methods. The "football field" is a single visual showing the implied EV from each method (DCF, EV/EBITDA, EV/EBIT, P/E, transaction comps, 52-week high/low, etc.), arranged as horizontal bars on a value axis:
For Sample Company (from the toolkit's Multiples Triangulation tab):
| Method | Implied EV | vs. DCF |
|---|---|---|
| DCF (this model) | $1,700M | (baseline) |
| P/E (forward, on Y1 NI) | $1,748M | +3% |
| EV/EBIT (forward) | $1,813M | +7% |
| EV/EBITDA (forward) | $1,904M | +12% |
The range across methods is $1,700M to $1,904M — a 12% spread. The mean is approximately $1,791M. This says: a defensible point estimate is $1,800M, with a reasonable range from $1,650M to $1,950M. Presenting it this way — as a range with central tendency, not a single number — is what separates rigorous valuation from false precision.
When presenting a DCF to a board, investment committee, or client, three rules:
A DCF presentation that hides these caveats — that presents a single confident-looking number — is more likely to mislead than to inform. The discipline of sensitivity analysis is what makes DCF a credible technique rather than a black-box exercise.
A DCF is easy to build mechanically and hard to build correctly. Six classes of error account for most of the trouble:
Setting terminal growth too high (above long-run nominal GDP), or terminal exit multiple too high (above current trading multiples). The TV dominates the valuation, so optimistic TV assumptions drive optimistic results. Fix: cap g at nominal GDP (3-4% in developed markets); use median, not maximum, comparable multiples for exit.
Modeling rapid revenue growth and margin expansion in the explicit period that the firm has never demonstrated historically. Projections "naturally" optimistic toward year 5. Fix: sanity-check projected growth and margins against the firm's historical performance and industry comparables. If you're projecting 25% growth for a firm that's grown 8% historically, you need a specific reason.
Presenting a DCF as "the firm is worth $X" without showing the range or driver sensitivity. Misleadingly precise. Fix: always present a range from sensitivity analysis (Section 06) and identify the most consequential drivers. A single-point DCF is worse than no DCF.
Presenting a DCF that produces a value 30%+ different from the multiples-implied value, without explanation. The disagreement is information, not noise. Fix: always triangulate against multiples (Module 06). Investigate any divergence; either reconcile or document why they shouldn't agree.
Using nominal cash flows with real discount rates, or vice versa. Mixing year-end and mid-year. Using FCFF with a cost-of-equity discount rate. Each is a category error. Fix: match cash-flow basis to discount-rate basis. Most DCF work is FCFF + WACC + nominal + end-of-year. Stick to a consistent convention throughout.
When a DCF answer doesn't match the desired outcome, adjusting inputs to force the result. The model becomes a rationalization rather than a valuation. Fix: document the assumptions before computing the answer; if the answer is uncomfortable, that's information about either the assumptions or the situation.
The discipline of professional DCF work is less about formula precision and more about defensibility. Three habits separate good practitioners from mechanical ones:
DCF mechanics are identical across countries — but the inputs vary substantially, and so do the practical considerations:
Standard 5-year explicit, Gordon Growth terminal with g of 2.5-3% (US long-run nominal GDP), USD cash flows, 10-year Treasury for risk-free rate. Equity research and M&A practice both use enterprise DCF. The US is where the textbook approach was developed and where it works most cleanly. Deep capital markets and stable accounting make the inputs reliable.
A generation of deflation produced terminal growth assumptions of 0-0.5% in many Japanese DCFs. With WACC of 4-6% (low Japanese discount rates), the WACC−g spread becomes very narrow, making valuations extremely sensitive to small changes. Some practitioners switched to multi-stage models (5 years explicit + 5 years fade-down + Gordon) to smooth the transition. Recent inflation has restored more normal terminal-growth assumptions.
Brazilian DCFs face a choice: project cash flows and discount in BRL (capturing local inflation, requiring local-currency WACC including country-risk premium) or in USD (translating cash flows through expected exchange rates, using USD WACC + country risk premium). Both methods are valid; both should produce similar answers in equilibrium. Differences signal mismatched inflation/FX assumptions. The toolkit can handle either by switching the WACC inputs and projection currency.
A decade of negative Swiss bond yields produced WACCs of 4-6% — substantially below US levels. The narrow WACC−g spread amplifies valuation sensitivity. Swiss firms also tend to be high-margin, low-growth multinationals (Nestlé, Roche, Novartis) where terminal growth assumptions matter especially much. Swiss DCFs require careful attention to FX exposure since most cash flows come from outside Switzerland.
Indian firms in growth sectors (financial services, technology, consumer) often project explicit-period growth of 15-20%+ — sustainable for the explicit window given India's economic trajectory, but problematic if extrapolated to terminal growth. The discipline: cap terminal g at India's expected long-run nominal GDP (~7-8%), even if explicit-period growth is much higher. Many DCFs use a longer explicit period (7-10 years) to bridge from high growth to sustainable growth.
Chinese DCFs require multiple judgment-based adjustments: a "China premium" added to WACC (typically 100-300bp) reflecting governance and geopolitical risk; treatment of state-influenced revenue and cost streams (subsidies, regulatory pricing); and currency-control considerations for international investors. Even after adjustments, two careful analysts can produce DCFs differing by 30%+ — a level of disagreement uncommon in developed-market work. Multiples cross-check is especially important.
The pattern across all six: DCF mechanics are universal, but DCF inputs require careful country-specific calibration. A practitioner who treats every market like the US will produce systematically biased valuations. The framework is the same; the assumptions vary.
Set the FCFF growth, WACC, and terminal-growth assumptions. The tool computes implied enterprise value, equity value, and per-share value live, plus the multiples-triangulation cross-check. Defaults reflect Sample Company; modify to model a different firm. For the full model with sensitivity tables and more detailed structure, download the Excel toolkit at the top of the page.
DCF is the synthesis. The questions test whether you understand not just the mechanics, but the discipline that turns a calculation into a defensible valuation.