In a typical first-round case that asks for a project NPV or a target's enterprise value, the terminal value is about three-quarters of the answer. Candidates who polish five years of neat cash-flow rows and then shrug at the perpetuity lose the number before they finish writing. The thesis here is simple: a case DCF is a logic check, not a banking modeling test, and the interviewer is listening for whether you can name which assumption moves the NPV the most.
After this piece you should be able to build a five-year unlevered free-cash-flow forecast, pick a WACC you can defend with published market data, add a Gordon growth terminal value, discount everything once, and run a two-way sensitivity that shows whether WACC or growth is doing the damage.
What a case DCF is actually asking for
Banking interviews want a clean walk-through of the full model: mid-year convention debates, diluted share count, circular WACC, and a bridge from enterprise value to equity value per share. Consulting cases usually want something narrower. The prompt is "should we invest $180 million in this plant" or "what is Midland Spark worth as a standalone business." You need the present value of future cash available to all capital providers, compared to the cash out today.
That means unlevered free cash flow (UFCF), discounted at WACC, plus a terminal value for year six onward. You almost never need a share-price bridge in the room. If the interviewer wants equity value, they will say so and hand you cash and debt.
I stop listening when a candidate recites six steps from a prep sheet and then cannot decide whether CapEx belongs in the cash flow. The sheet rewards a short, verbalizable build: cash flows, discount rate, terminal value, sum, then one sensitivity.
Anchor cash-flow lines in something real
Before inventing Midland Spark, look at how a real beverage company reports the two lines that matter most for free cash flow. PepsiCo's fiscal 2024 results, filed with the SEC, show net cash provided by operating activities of $12.507 billion and capital spending of $5.318 billion. Their disclosed free cash flow for the year is $7.531 billion: operating cash flow minus CapEx. That is the same skeleton you use in a case, scaled down.
In the room you will not have a full cash-flow statement. You will build UFCF from operating profit:
UFCF = EBIT × (1 − tax rate) + D&A − Δ net working capital − CapEx
EBIT is operating profit before interest. Tax it as if the firm had no debt, so financing does not pollute the cash flow. Add back depreciation and amortization because they are non-cash. Subtract the increase in net working capital when growth ties up cash in inventory and receivables. Subtract CapEx because cash actually leaves. That is the cash that belongs to debt and equity together, which is why you discount it at WACC rather than at the cost of equity alone.
Pick a discount rate you can defend
Guessing "ten percent because it is round" is how candidates get challenged. As of January 2026, Aswath Damodaran's U.S. industry cost-of-capital table puts Beverage (Soft) at a 6.33 percent WACC across 27 firms, with a beta of 0.64 and an after-tax cost of debt near 4 percent. That is a listed-industry average, not your private regional target.
For CAPM inputs on the same date, Damodaran's implied equity risk premium update starts from an 8.41 percent expected return on the S&P 500 and a 4.18 percent U.S. Treasury bond rate, which nets an equity risk premium of 4.23 percent. A mid-market soft-drink target is riskier than PepsiCo. In a case I would take rf = 4.2 percent, ERP = 4.2 percent, beta = 1.0 for a smaller brand, cost of equity ≈ 8.4 percent, 20 percent debt at a 5 percent pretax rate, 25 percent tax, and land on a WACC of about 8 percent.
Say the 8 percent out loud and cite the industry table as the floor. Interviewers like candidates who know where the number came from even when they round it for the whiteboard.
Worked example: Midland Spark, five years plus terminal value
Invented company, original numbers. Your client is considering acquiring Midland Spark, a regional sparkling-water brand. The ask is enterprise value on a standalone basis. You have built the following unlevered free cash flows (all figures in $ millions):
| Year | UFCF | Notes |
|---|---|---|
| 1 | 22 | Post-synergy run-rate not included; standalone only |
| 2 | 25 | Volume growth and mild price mix |
| 3 | 28 | CapEx still above D&A |
| 4 | 30 | Working-capital intensity stabilizes |
| 5 | 32 | Last explicit year; feeds the terminal value |
Assumptions you state before you write the table: WACC = 8 percent; perpetual growth g = 2.5 percent after year 5. Keep g well below the risk-free rate you used in CAPM. Damodaran's January 2026 update puts the U.S. Treasury bond rate at 4.18 percent; a 2.5 percent perpetual growth rate for one brand sits comfortably under that ceiling and does not imply Midland Spark eventually outgrows the economy.
Gordon growth terminal value at the end of year 5:
TV₅ = UFCF₅ × (1 + g) / (WACC − g) = 32 × 1.025 / (0.08 − 0.025) = 32.8 / 0.055 ≈ 596
Discount each cash flow and the terminal value at 8 percent. Present-value factors: year 1 = 0.926, year 2 = 0.857, year 3 = 0.794, year 4 = 0.735, year 5 = 0.681.
| Item | Nominal ($M) | Discount factor | Present value ($M) |
|---|---|---|---|
| Year 1 UFCF | 22 | 0.926 | 20.4 |
| Year 2 UFCF | 25 | 0.857 | 21.4 |
| Year 3 UFCF | 28 | 0.794 | 22.2 |
| Year 4 UFCF | 30 | 0.735 | 22.1 |
| Year 5 UFCF | 32 | 0.681 | 21.8 |
| Terminal value | 596 | 0.681 | 405.9 |
| Enterprise value | ≈ 514 |
Read the table the way an interviewer will. The five explicit years sum to about $108 million of present value. The terminal value alone is about $406 million, or roughly 79 percent of enterprise value. If you spent twelve minutes debating year-3 working capital and thirty seconds on g, you optimized the wrong row.
If the ask is project NPV rather than an acquisition price, subtract the upfront investment. A $180 million plant against a $514 million present value of cash flows is a clear yes on these assumptions; a $600 million plant is a clear no. Same math, different framing.
The sensitivity that actually matters
A single-point DCF is a hypothesis. Flex the two inputs that dominate the terminal value: WACC and g. Holding everything else fixed, Midland Spark's enterprise value ($ millions) looks like this:
| g = 2.0% | g = 2.5% | g = 3.0% | |
|---|---|---|---|
| WACC 7% | 575 | 631 | 701 |
| WACC 8% | 478 | 514 | 557 |
| WACC 9% | 408 | 433 | 462 |
Moving WACC by one percentage point at the base growth rate swings value from about $433 million to about $631 million, a range of roughly $200 million. Moving g by half a point at the base WACC swings value from about $478 million to about $557 million, a range of about $80 million. A 10 percent miss on year-5 cash flow moves the answer by roughly $50 million because most of the value is the capitalized terminal cash flow.
WACC is the assumption the Midland Spark answer is most sensitive to. Say that before the interviewer asks. Then explain why: the denominator (WACC − g) is small, so a one-point change in the discount rate rewrites the perpetuity.
Ned's rule. In a case DCF, spend one minute on the explicit forecast and two minutes on WACC, g, and a two-way sensitivity. If terminal value is more than two-thirds of enterprise value and you have not flexed the discount rate, you are not finished.
What the interviewer scores, and what they ignore
They score: correct identity of unlevered cash flow, a WACC that is not plucked from air, a terminal growth rate below economy-wide growth, clean discounting, and a sensitivity that names the driver. They do not score: mid-year conventions, circular references, or a three-statement model. Those are banking hygiene. Using them in a consulting case burns time without earning points.
Common failure modes I see in the room:
- Discounting UFCF at the cost of equity (overstates value).
- Setting g at 4 or 5 percent "because the brand is growing," which quietly assumes Midland Spark eventually outgrows the economy.
- Treating terminal value as a plug instead of a formula the interviewer can audit.
- Giving one number with false precision ("$513.74 million") and no range.
The independent check is verbal: "On Damodaran's soft-drink WACC as a floor, I used 8 percent for a smaller brand, 2.5 percent perpetual growth, and the value is about $510 million. It is most sensitive to WACC; at 9 percent we are closer to $430 million."
CoachNed is not affiliated with McKinsey, BCG, Bain, Deloitte, or any other firm named in interview prep.
Practice this on a live case
Take one M&A or investment case from the library and force a full DCF on paper in twelve minutes: five UFCF rows, one TV, one WACC, one sensitivity table. Then say the answer out loud, including which assumption moves NPV the most. Everything on CoachNed is open for seven days with no card; after that it is $120 for a recruiting season or $49 a month.
For a timed math warm-up before the full case, use /drills/quick-math. For a voice run where you have to defend the discount rate under pressure, try /interview.
Frequently asked questions
How many years should a DCF forecast in a case interview?
Five years is the default. It is long enough to show a transition from growth to a steadier state and short enough to finish on a whiteboard. Use three years only if the interviewer gives a short project life; use longer only if they hand you year-by-year drivers.
Do I use levered or unlevered free cash flow?
Use unlevered free cash flow and discount at WACC when the ask is enterprise value or project NPV for the firm as a whole. Use levered cash flow and the cost of equity only when the interviewer explicitly wants equity value from cash flows to equity holders.
What perpetual growth rate is safe to use?
Stay at or below a conservative view of long-run nominal GDP growth for a single firm or brand, typically 2 to 2.5 percent in U.S. dollar cases today. If your g approaches your WACC, the model explodes and the interviewer will stop you.
Should I use Gordon growth or an exit multiple for terminal value?
Gordon growth is cleaner in cases because every input is explicit. An exit multiple is fine as a cross-check if the interviewer gives you a comps multiple; then state the growth rate that multiple implies so the two methods do not contradict each other.
How do I get from enterprise value to equity value?
Add cash, subtract debt, preferred stock, and minority interest, then divide by diluted shares if they want a price per share. Most consulting cases stop at enterprise value or project NPV unless they ask for the bridge.
Sources
- PepsiCo Q4 2024 earnings exhibit (SEC filing) — operating cash flow, capital spending, and free cash flow line items for fiscal 2024. Checked 2026-09-24.
- Damodaran Cost of Capital by Industry (US) — Beverage (Soft) industry WACC, beta, and capital-structure weights as of January 2026. Checked 2026-09-24.
- Damodaran Data Update 2 for 2026 — implied equity risk premium (4.23%) and Treasury bond rate (4.18%) used as the stable-growth ceiling for g. Checked 2026-09-24.
- Damodaran Historical Implied Equity Risk Premiums — time series of T-bond rates and implied ERP through January 2026. Checked 2026-09-24.