How to Walk Through a DCF Like You've Done It

The tells aren't in the steps. Everyone gets the five steps roughly right: project cash flows, discount them, calculate terminal value, discount that, bridge to equity. The tells are in the small things. Whether you say "discount" with confidence or hedge it. Whether you know why the cash flows are unlevered without being prompted. Whether you treat terminal value like the afterthought every prep guide treats it as, or whether you flag — proactively, without being asked — that it's 60–80% of the answer and most of where the disagreement lives.

This piece walks through the DCF the way someone who's built one would explain it. Not exhaustive. Exhaustive is what gets you tuned out. Structured, with the layer of why underneath each step that separates the candidate who's read a guide from the candidate who's spent a weekend wrestling with circular references in Excel.

The Five Steps, Stripped Down

The skeleton, in ninety seconds:

1. Project unlevered free cash flow for 5–10 years
2. Discount each year's UFCF back to today at WACC
3. Calculate terminal value at end of forecast period
4. Discount terminal value back to today at WACC
5. Sum to get enterprise value, then bridge to equity value

That's the structure. Every prep guide on the internet will tell you this much. The rest of this article is everything that lives inside those five lines: the choices, the why, and the places interviewers push.

#### Step 1 — Unlevered Free Cash Flow

The formula: UFCF = EBIT × (1 − t) + D&A − CapEx − ΔNWC

Each piece earns its place. EBIT × (1−t) is operating profit, taxed as if the business carried no debt. This is the "unlevered" part: you're stripping out the financing decision entirely so the cash flow reflects only the operating business. D&A gets added back because it's a non-cash expense that sat inside EBIT; the business didn't actually write a check for depreciation last year. CapEx gets subtracted because it's the actual cash going out the door to maintain plant and equipment and to fund growth. Change in net working capital gets subtracted because as a business grows, it ties up more cash in receivables and inventory than it gets back from payables.

Now the question that separates candidates: why unlevered?

Because the discount rate you'll apply in step 2 is WACC, which already blends the cost of debt and equity. If you used levered free cash flow, which already nets out interest expense, and then discounted at WACC, you'd be double-counting the debt benefit. Once in the cash flow (lower interest expense reduces taxes, boosting cash flow) and once in the discount rate (after-tax cost of debt is lower than cost of equity, dragging down WACC).

The fix: unlever the cash flows. Strip out the financing decision. Then apply WACC, which captures the financing decision separately. The two pieces — cash flow and discount rate — have to be consistent about whose claim you're valuing, and unlevered means you're valuing the whole enterprise before debt and equity split it up.

Say this sentence out loud in your walkthrough, unprompted, and you've already signaled a tier above most candidates. It is the single highest-leverage piece of why in the entire model.

The forecast period question. Default to 5 years for mature businesses, 10 years for growth businesses that haven't reached steady state. The principle: the explicit forecast period should run as long as it takes for the business to look "normal" — stable margins, stable growth rate, stable reinvestment rate. If you're modeling a software company growing 40% a year, five years is not enough. The business is still in transition at year five, and you can't drop into a terminal value calculation that assumes perpetual stability when the business is anything but stable. Push the forecast to year 7 or 10 and let it normalize before you hand it off to the terminal value formula.

#### Step 2 — Discounting at WACC

The formula: WACC = (E/V) × Re + (D/V) × Rd × (1 − t)

Where V = E + D, and — critically — E and D are market values, not book.

This is a place candidates fumble and a place MDs love to push on. Book value of equity is whatever's left on the balance sheet after liabilities. Market value of equity is shares outstanding times share price. For companies trading at multiples of book, which is almost every public company you'll value, the two numbers are different by an order of magnitude. Using book value of equity in your WACC weights is wrong in a way that suggests you don't understand what WACC is measuring. WACC is the rate at which today's investors discount future cash flows, and today's investors paid today's prices.

Cost of equity comes from CAPM: Re = Rf + β × (Rm − Rf)

The components, with the choices an actual practitioner makes.
Risk-free rate: 10-year Treasury yield. Not the 30-year, which is too long and captures inflation expectations that distort the rate. Not the 3-month T-bill, which is too short and doesn't match the duration of the cash flows you're discounting. The 10-year is the convention because it roughly matches the duration of the forecast plus terminal value.

Equity risk premium: 5–6% is the working range in U.S. practice. Damodaran publishes one every year if pressed for a source, and his number is the closest thing the industry has to a standard reference. Don't say "10%," a number candidates pick when they're guessing, and don't say "2%," a number candidates pick when they're confusing the risk-free rate with the premium.

Beta: for a public company, pull it from Bloomberg or Capital IQ. For a private company, you can't observe it, so you lever and unlever beta from comparable public peers: pick comps, unlever their betas to strip out capital structure differences (Hamada's equation handles this), average the unlevered betas, and relever to the target company's capital structure. You don't need to derive the algebra in an interview, but knowing the procedure by name signals you've seen the move before.

Cost of debt: After-tax, because interest is tax-deductible and you want the post-tax cost the firm actually bears. For investment-grade companies with public debt, use the yield on their existing debt. For others, use the yield on comparable-rated debt from companies with similar credit profiles.

The hidden gotcha most candidates miss: for private companies, you can't observe market value of equity directly, which means you can't calculate the weights, which means you can't calculate WACC, which means you can't value the firm, which means you can't get the market value of equity. The reference is circular. The practical fix in a model is to iterate: assume a target capital structure based on comparables, calculate WACC, value the firm, check whether the implied debt-to-equity ratio matches your assumption, and revise.

In an interview, you handle this by saying "for a private company I'd assume a target capital structure based on comps." Don't pretend the circularity doesn't exist. Flagging it gets you credit. Pretending it doesn't loses you credit when the interviewer pushes.

#### Step 3 — Terminal Value (Where Most of the Answer Lives)

The point interviewers want to hear you internalize, unprompted: terminal value typically represents 60–80% of total enterprise value in a standard DCF. This is not a footnote. This is most of the answer.

Which means the assumptions in your terminal value calculation matter more than the assumptions in your 5-year forecast. Which means most candidates spend most of their prep time on the wrong thing, building elaborate revenue models for years 1 through 5 and then waving their hands at the terminal value.

Two methods, both standard.

• Gordon Growth (perpetuity growth): TV = FCF_(n+1) / (WACC − g)
• Exit multiple: TV = EBITDA_n × Exit Multiple

Which to use, in practice: both. Calculate it both ways and compare. If the two methods give wildly different terminal values, something is off in your assumptions. The standard sanity check is to back-solve: calculate the implied growth rate from your exit multiple, or the implied multiple from your growth rate, and see whether the implied number is reasonable.

The g question. Do not pick a growth rate above long-run nominal GDP growth, which in developed markets is roughly 2–3%. The reason is mechanical, not stylistic. A company cannot grow faster than the economy forever, because if it did, it would eventually become the entire economy. This is a mathematical absurdity, not an estimation choice. Most candidates know this in principle and still pick 4% under pressure because it makes the valuation look better. Don't.

The exit multiple question, which gets asked less often but matters more. Where does the multiple come from? From current trading multiples of comparable public companies. Which means you're assuming that in 5 or 10 years, the comps will still trade at similar multiples. This is a heroic assumption. Multiples compress and expand with interest rate cycles, growth expectations, and sentiment. Flag it. Bankers respect candidates who flag heroic assumptions instead of papering over them, because the alternative — pretending the assumption is precise — is the kind of confidence that gets you torn apart later in the deal cycle.

The discounting step everyone forgets to mention: PV of TV = TV / (1 + WACC)^n

Where n is the last year of your explicit forecast period — year 5 in a 5-year DCF — not n+1. Terminal value is calculated as of the end of year n, representing all cash flows from year n+1 to infinity. So you discount back from year n. Small detail. Interviewers notice when you get it right and they notice when you get it wrong.

#### Step 4 — Sensitivity, Because the Number Is Wrong

The honest framing: every DCF is wrong. The output is a function of a dozen assumptions, each of which could reasonably move 10–20%, and the answer compounds the uncertainty. The point of the model isn't to produce a number. It's to produce a range and understand which assumptions are driving the range.

The two assumptions that move the answer most: WACC, and either terminal growth rate (if you used Gordon Growth) or exit multiple (if you used that method). A 50 basis point shift in WACC typically moves equity value by 8–15%, depending on the duration of the cash flows: longer-duration cash flows are more sensitive because more of the value sits further out. A 0.5 percentage point shift in terminal growth rate moves the answer by similar magnitudes for the same reason.

This is why sensitivity tables exist in every real DCF model. In an interview, you don't need to produce a sensitivity table — you need to say, proactively, "the output is most sensitive to WACC and terminal growth rate, which is why we'd run a sensitivity table on those two variables." That sentence does an enormous amount of work. It signals you understand the model is a range, not a point. It signals you know which assumptions matter. And it preempts the question the interviewer was probably about to ask.

The contrarian point worth making. Candidates often present a DCF result to the interviewer as a single number — "I got $52 per share." This is the wrong instinct. Sophisticated bankers don't believe point estimates from DCFs and they're suspicious of candidates who deliver them with confidence. The DCF in real banking practice is a triangulation tool. You use it alongside trading comps and precedent transactions, and what you care about is the range: does the DCF support the comps, or does it suggest the comps are overheated, or is everything pointing to the same number from different angles. Frame your answer the same way. "The DCF gives me a range of $48 to $58 per share, which is broadly consistent with where the trading comps suggest the company should price." That answer is more sophisticated than $52, and it's also more honest.

#### The Bridge From EV to Equity Value

The full bridge, because candidates routinely skip pieces:

Enterprise Value
− Total Debt
+ Cash and Cash Equivalents
− Preferred Equity
− Minority Interest (Non-Controlling Interest)
+ Investments in Associates
= Equity Value
Equity Value / Diluted Shares Outstanding = Implied Share Price.

Each piece has a why. Debt gets subtracted because equity holders get paid after debtholders in any payout. Cash gets added because it's a non-operating asset already owned by equity holders: the DCF values the operating business, and cash sits separately. Preferred equity gets subtracted for the same reason as debt; it's a senior claim on cash flows that equity sits behind. Minority interest gets subtracted because enterprise value captures 100% of consolidated subsidiary cash flows, but the parent's equity only owns the parent's share — the rest belongs to the minority shareholders of the subsidiary. Investments in associates get added back because their earnings flow in below the EBIT line and weren't captured in the FCF projection.

Diluted shares, not basic. Use the treasury stock method for in-the-money options. If you're not sure what the treasury stock method is, close this article, learn it, then come back. It's a foundational mechanic and not knowing it is a separate gap to close.

What MDs Push On

A handful of concrete pushbacks, with the responses that hold up.

"Your terminal growth rate is too high."

"Why CAPM, given beta is unstable and the equity risk premium is contested?"

"Your WACC looks low."

"What if the company is loss-making?"

"Defend your discount rate to two decimal places."

The meta-point: interviewers asking these questions are not trying to find the "right" number. They're testing whether you understand the model well enough to defend it under pressure without retreating into either false confidence or visible panic. Confident "here are the inputs, here are the assumptions, here is the range" beats "I'm sure it's exactly $52.30" every time.

The Walkthrough, Cleaned Up

A model 90-second version, for pacing. Not a script — a demonstration of how the technical detail above compresses into something that fits the time budget while still hitting the why signals.

A DCF values a company based on the present value of its future cash flows. The five steps are project unlevered free cash flow for the explicit forecast period, usually five to ten years depending on how long the business takes to stabilize, discount each year's cash flow back to today at WACC, calculate terminal value at the end of the forecast period, discount that back to today as well, and sum to enterprise value.

The cash flows are unlevered because we're discounting at WACC, which already captures the cost of debt. Using levered cash flows would double-count the financing benefit.

WACC is the market-value-weighted blend of cost of equity and after-tax cost of debt. Cost of equity from CAPM — 10-year Treasury, plus beta times the equity risk premium of around 5–6%.

Terminal value typically represents 60–80% of total enterprise value, so the assumptions there matter most. I'd calculate it both ways — Gordon Growth using a perpetuity growth rate at or below long-run GDP growth, and exit multiple using current trading comps — and compare for sanity.

Once I have enterprise value, I bridge to equity by subtracting debt, adding cash, and adjusting for preferred, minorities, and investments in associates. The output isn't a point estimate — I'd present it as a range and run sensitivity on WACC and terminal growth.

About 100 seconds spoken at a measured pace. Hits the structure, hits the unlevered why, flags terminal value as most of the answer, names the sensitivity drivers, and ends on the range rather than a number. That's the version that sounds like you've built one.

Where to Drill

The Analyst-level interviewer on HARDO is built around this kind of material — numerical edge cases on the WACC components, follow-ups when the unlevered explanation goes vague, push on the terminal growth rate when you pick something too aggressive. It runs in IB shorthand and gets impatient with the answers that sound rehearsed. If the DCF is the technical floor and you want to stress-test yours before recruiting, that's the seat to take.