A discounted cash-flow model estimates what a business is worth today by projecting the cash it will generate in future years and shrinking each of those future amounts back to present-day money. It is the most direct way to estimate intrinsic value — and the most honest, because it forces you to write down every assumption you are making.
You cannot simply add up the cash a company will earn over the next ten years, because a dollar arriving in ten years is not worth a dollar today. Two reasons. First, a dollar you have now can be invested and grow. Second, a dollar promised later might never arrive. Both push in the same direction: distant cash is worth less, and the further away and the less certain it is, the less it is worth.
The rate you shrink future cash by is called the discount rate, or your required rate of return: the annual compensation a business this risky would have to be priced to offer before you tied up money in it. It is built from the risk-free rate you could get from government bonds, plus a premium for the risk that this company disappoints you. A safe, boring utility gets a low discount rate; a speculative growth company gets a high one, which is precisely why the same projected cash flows are worth less coming from a risky business.
Present value = the sum, over each future year t, of CF_t ÷ (1 + r)^t. CF_t is the cash the business produces in year t — in practice, its free cash flow, the cash left after it pays for the equipment and working capital it needs to keep running. r is the discount rate. The exponent t is what does the work: dividing by (1 + r) once per year compounds the shrinkage, so year 10 is discounted far harder than year 1.
You cannot project cash flows individually forever, so a real DCF splits the future in two. For an explicit forecast period — typically five or ten years — you project each year by hand. Then everything after that is compressed into a single terminal value, most often with the Gordon growth formula: terminal value at the end of year n = CF_(n+1) ÷ (r − g), where g is the modest rate you assume the cash grows at forever. That terminal value is itself a future amount, so it too gets discounted back by dividing by (1 + r)^n.
Note the shape of that terminal term: (r − g) sits in the denominator. As your assumed perpetual growth g creeps toward your discount rate r, the denominator approaches zero and the value explodes toward infinity. This is not a deep insight about business — it is a division artifact, and it is the single easiest way to accidentally justify any price you like. A sane g is bounded by long-run economic growth: no company grows faster than the whole economy forever, because eventually it would be the economy.
Example (illustrative — invented numbers chosen to show the mechanics, not a real company and not a recommendation). A business is expected to produce $100 of free cash flow next year, growing 8% a year for five years. You discount that cash at 10% a year (r = 10%, your required rate for a business this risky), and you assume that after year five the cash grows 2.5% a year forever (g = 2.5%).
The five projected cash flows are $100, $108, $116.64, $125.97 and $136.05. Discounting each by (1.10)^t gives $90.91, $89.26, $87.63, $86.04 and $84.48 — about $438 in total. Notice how flat those discounted values are: the 8% growth and the 10% discount rate nearly cancel, which is a useful intuition about why growth alone is not value.
Now the terminal value. Year-six cash is $136.05 × 1.025 = $139.45, so the terminal value at the end of year five is $139.45 ÷ (0.10 − 0.025) = $1,859. Discounting that back five years gives $1,859 ÷ (1.10)^5 = $1,155. Total estimated value: $438 + $1,155 = about $1,593.
Look at what just happened. The terminal value — the part built entirely on a guess about the indefinite future — is $1,155 of the $1,593, or roughly 72% of the answer. The five years you carefully projected supply barely a quarter of it. That ratio is typical of real DCFs, and it is the most important thing a beginner can learn about this method: most of the number comes from the part you know least about.
Keep every projection above identical and change one input: discount at 12% instead of 10%. The explicit five years now discount to about $416, the terminal value to about $833, and the total collapses to roughly $1,249 — a 22% drop in the estimated value of the business from a two-percentage-point change in a number you essentially chose. (All illustrative arithmetic from the invented inputs above.)
This is the real lesson of a DCF, and it is not a defect of the technique. It is telling you the truth: the value of a business genuinely does depend that heavily on how risky you think it is and how fast you think it will grow, and nobody knows those numbers. A model that returned a stable, confident answer regardless of inputs would be lying to you.
The professional way to use this is backwards. Instead of running a DCF to produce a price, take the actual market price and solve for what the market must be assuming — what growth rate, what margin, what discount rate makes today’s price fair? Then ask whether those implied assumptions are plausible. "This price requires 20% growth for a decade" is a far more useful sentence than "my model says $142.68."
It gives false precision. The output has decimal places; the inputs are guesses. The arithmetic is exact, which fools people into thinking the answer is too. Garbage in, precisely-formatted garbage out.
It is trivially manipulated, usually without dishonest intent. Because you choose g and r, you can always produce the conclusion you already reached — nudge growth up a point, shave the discount rate, and any stock becomes cheap. A DCF built after you decided to buy is not evidence; it is decoration.
It struggles with the companies people most want to value. A business with negative free cash flow today, whose worth is entirely in an uncertain future, has nearly all of its value in the terminal term — where the model has the least to say. And a DCF cannot see a structural break: it assumes the business that generated last year’s cash still exists in year ten. Technology, regulation and competition regularly disagree.
Finally, be clear what a DCF is not: it is not a forecast of the stock price. It estimates what the business is worth; the market can disagree with you for years. Quantustik’s model does the opposite — it forecasts the probable range of the PRICE over a horizon and says nothing about intrinsic value. Neither replaces the other, which is why the discipline of a margin of safety exists.
A way to estimate what a business is worth today by projecting the cash it will generate in future years and shrinking each future amount back to present-day money, because a dollar later is worth less than a dollar now.
Present value = the sum over each year t of CF_t ÷ (1 + r)^t, where CF_t is the free cash flow in year t and r is the discount rate. Cash beyond the explicit forecast period is compressed into a terminal value, usually CF_(n+1) ÷ (r − g), which is then discounted back the same way.
Your required rate of return — the annual compensation a business this risky would have to be priced to offer before you tied up money in it. It is roughly the risk-free rate available from government bonds plus a premium for the risk that this particular company disappoints you. Riskier business, higher discount rate, lower present value for the same projected cash.
Because it stands in for every year after your explicit forecast — the entire indefinite future. In a typical five-year model it can be roughly three quarters of the total value, meaning most of the answer rests on the assumptions you are least able to defend.
It is precise, not accurate. The arithmetic is exact, but the inputs are estimates, and a two-point change in the discount rate can move the result by around a fifth. Use it to find out what assumptions today’s price implies, and pair it with a margin of safety rather than trusting the output as a target price.
Browse all S&P 500 tickers to see this metric applied to individual companies.
Educational research only — not investment advice.