A DCF valuation, short for discounted cash flow, estimates what a company is worth today by forecasting the free cash flow it will produce and discounting that back to present value. It is at once the most rigorous valuation method in common use and the easiest to bend, because the answer can only ever be as good as the assumptions fed into it.

This module walks a DCF from end to end, covering the forecast, the discount rate and the terminal value. It then shows why the method produces a range rather than a single figure, and which input choices quietly inflate a target price.

What a DCF actually does

The whole method rests on one idea, and it is an idea you already use without naming it. A pound of cash next year is worth less to you than a pound today. You could invest today’s pound, you are certain of having it, and next year’s pound might not arrive at all. So future cash has to be marked down, or discounted, before it can be compared with money in your hand.

A DCF does exactly that, systematically. It forecasts a company’s free cash flow for several years ahead, discounts each of those years back to what it is worth today, and adds the results together. What comes out is an intrinsic value: an estimate of what the business is worth based on the cash it will actually generate, arrived at without reference to what anyone is currently willing to pay for the shares.

That last clause is why the method is worth the trouble. Multiples tell you what the market thinks. A DCF is an attempt to work out the answer from first principles, so that you have something to compare the market’s opinion against.

Building the forecast: revenue to free cash flow

The engine of any DCF is a forecast running several years out that turns a revenue number into a free cash flow number. Every step in that chain is an assumption, and every assumption is something the analyst has to be able to defend.

The chain runs like this. Start with a revenue growth path. Apply an operating margin, which may widen over the forecast if the business gets more efficient as it grows. Apply a cash tax rate to the operating profit. Add back depreciation, because it reduced profit on paper but no cash left the building. Then subtract capital expenditure, because that cash genuinely did leave.

The step that most often goes wrong is the margin path, and it goes wrong through a story called operating leverage. The argument is that fixed costs stay fixed while revenue climbs, so margins should widen. That is frequently true and entirely legitimate. The trouble is that it is also easy to assume, hard to disprove within a five year window, and enormously powerful in the model. A forecast that quietly widens the margin every year will produce a large valuation without ever making a claim you can obviously challenge.

When you read someone else’s forecast, compare the assumed margin in the final year against what the company has actually achieved in its best year, and against what its strongest competitor achieves today. If the model has the business becoming more profitable than any peer has managed, that is a claim, and it should be argued for somewhere in the note.

The discount rate: small change, big swing

The discount rate, usually the weighted average cost of capital (WACC), represents how risky the cash flows are and what return investors require to take that risk on.

It is the most sensitive input in the model, and it is worth understanding why rather than just being told. The rate does not apply once. It compounds, appearing in the denominator of every single year and raised to a higher power each time. So a rate that is one percentage point too low is one percentage point too low in year one, and compounded nine times over in year ten, and compounded indefinitely inside the terminal value. A change that looks trivially small at the input end arrives at the output end magnified.

This is why a suspiciously low WACC is among the fastest ways to inflate a target price, and among the hardest to spot, because a single number buried in an appendix does not look like an argument. It is one.

Terminal value: the biggest and most fragile number

A forecast cannot run forever, but a company does not stop generating cash at the end of year five. So a DCF estimates a terminal value to stand for everything beyond the explicit forecast window.

Here is the fact that surprises people meeting this for the first time. That single estimate frequently accounts for the majority of the total valuation, sometimes well over half. Most of the answer therefore comes from the part of the model with the least detailed work behind it, which is exactly the opposite of where you would want the weight to sit.

There are two ways to calculate it. A perpetuity growth rate assumes the cash flows carry on growing forever at a fixed rate. An exit multiple applies a plausible sale multiple to the final forecast year, as though the business were sold at that point.

The rule is not just tidy arithmetic. It is a real constraint, and small breaches of it are startlingly expensive. In a perpetuity calculation the terminal value depends on the gap between the discount rate and the growth rate, so as growth creeps towards the discount rate the gap narrows and the value climbs steeply. Nudging terminal growth from 2.5% to 4% can lift the total valuation by a third or more without a single word changing in the written thesis. This is the most powerful lever in the model and the least visible.

Enterprise value to share price

The DCF produces an enterprise value, which is the value of the whole business: everything the operations are worth, regardless of who has a claim on it.

But you are not buying the whole business. You are buying a share, and the lenders are paid before you are. So to reach a share price you subtract what is owed. Equity value = enterprise value − net debt, and then you divide by the number of shares in issue.

This step explains something that otherwise looks strange. Two companies can run identical operations, produce identical cash flows and carry identical enterprise values, yet their shares can be worth very different amounts. The debt sits in between the business and the shareholder, and it is subtracted before you get anything. It also explains why leverage cuts both ways for an equity holder: when enterprise value rises, the fixed debt is subtracted from a bigger number and the equity gains disproportionately, and when it falls, the equity absorbs the loss the same way.

A DCF is a range, not a point

Having seen how far the output moves when the inputs shift, you should now be suspicious of any DCF quoted as a single figure. A precise number implies a precision the method does not have.

Serious analysts run a sensitivity table, varying the WACC and the terminal growth rate across the range they consider plausible and reporting the spread that results.

To illustrate with figures of the kind you would see: varying WACC from 7.5% to 10.5% and terminal growth from 2% to 4% might produce intrinsic values anywhere from £44 to £121 a share, with a central estimate around £54.

Look closely at those three numbers, because they teach the method’s real character. The central estimate of £54 sits far nearer the bottom of the range than the top. That is not a mistake. Value depends on the gap between the discount rate and the growth rate, and as that gap narrows the value does not rise steadily, it accelerates. So the optimistic corner of the table runs away to £121 while the pessimistic corner only falls to £44. Whenever you see a wide DCF range, expect the top end to be the least reliable number in it.

A cross-check against comparable companies helps triangulate, since the two methods fail in different ways. Say an EV/EBITDA comparison gives £52 to £80 with a central figure of £67. If the share trades at £38, below both ranges, that points towards potential undervaluation. But notice what has actually been established. Two models built on assumptions both say the share looks cheap. What matters is whether those assumptions hold, not that two ranges happened to agree.

Common DCF mistakes: how to spot a flattering model

Most models that have been steered towards a desired answer use the same small set of levers. Once you know them you can check a valuation in a few minutes.

• Optimistic growth. Revenue or margin assumptions well above the company’s own history and above what its peers achieve. Always check the final forecast year against the best year the business has actually had.
• A terminal growth rate above GDP. Impossible to sustain, and it sits on the largest single component of the answer.
• A WACC set too low. Understates the risk, and compounds through every year of the model.
• Capital expenditure ignored or minimised. Forecasting the cash a business produces without the spending required to produce it.

Any one of these can turn a fair valuation into a flattering one, and they are more effective in combination because each looks defensible on its own. This is why the right question to ask about a DCF is never “what is the number?” It is always “what assumptions produced it?”