DCF versus comparables

How comparables work in practice

You identify transactions involving similar businesses, extract the multiple each implied, and apply an appropriate multiple to your own earnings. The method's strength is that it reflects what buyers have actually paid rather than what a model says they should.

Its weakness is comparability. Genuinely similar transactions are rare, private deal terms are frequently undisclosed, and the multiple reported often omits the earn-out, the debt assumed or the working capital adjustment that materially changed what was really paid.

In thinner markets the problem compounds. There may be very few relevant private transactions in a given sector and geography, so advisers reach for listed comparables and apply discounts for illiquidity and size — each of which is itself a judgement, stacked on top of another judgement.

How a DCF works

You forecast the free cash flow the business will produce over a defined period, estimate a terminal value representing everything beyond it, and discount both back to today at a rate reflecting the risk of not receiving them.

The logic is unimpeachable: a business is worth the cash it will generate, adjusted for the fact that future money is worth less than money now, and uncertain money is worth less than certain money.

The difficulty is that it requires forecasting the future. Every input is an assumption, and the model presents the output to the nearest rand, which lends a false precision that persuades people far more than it should.

The two inputs that do most of the work

The discount rate. Small changes produce large swings. A business valued at a 15% discount rate can be worth roughly a third less at 18% — and both rates are defensible for a private company in a volatile market. For African businesses the rate carries country risk, currency exposure and illiquidity on top of the ordinary cost of capital, which is why headline rates here are materially higher than in developed markets and why the resulting valuations look conservative to anyone comparing internationally.

Terminal value. In most DCFs this represents well over half the total value and sometimes far more, meaning the majority of the answer rests on an assumption about what happens after the forecast period ends — the least knowable part of the exercise. Whenever you see a DCF, ask what proportion of value sits in the terminal figure. Above about seventy percent, the model is largely an opinion about perpetuity wearing arithmetic.

Where each method breaks down

Comparables fail when the business is genuinely unusual, when the market is illiquid enough that few transactions exist, when reported multiples hide their real terms, and when sector conditions have shifted since the comparable deals closed.

DCF fails when the forecast is unreliable — which is most acute for young or volatile businesses, precisely where owners most want a rigorous number. It also fails quietly, because a model with poor assumptions still produces a confident-looking output. And it is easy to steer: adjust growth by two points and the discount rate by one, and you can reach almost any figure you were hoping for.

Reading a DCF critically

Five questions expose most of what matters. What growth rate is assumed, and how does it compare with what the business has actually achieved? What discount rate, and what is it built from? What share of total value is terminal? What margin trajectory is assumed — does it improve, and why would it? And what happens to the answer if growth is two points lower and the discount rate two points higher?

That last question is the whole exercise. A valuation presented as a single number is a claim. A valuation presented as a range with the sensitivities shown is an analysis, and it is what a serious counterparty expects to see.

What a forecast has to survive

Because a DCF rests on its forecast, the forecast is what a serious counterparty attacks first. Three tests recur.

Does the growth rate have a mechanism? A plan that shows revenue rising thirty percent a year needs to say where the customers come from, who serves them, and what it costs to acquire them. Growth asserted without a mechanism is treated as decoration.

Does margin improvement have a cause? Forecasts routinely show margins expanding as revenue grows, on an implied assumption of operating leverage. Sometimes correct. Frequently it ignores the additional management, systems and premises that scale actually requires.

Is capital expenditure realistic? A forecast showing strong growth and flat capital expenditure is usually wrong. Growth consumes working capital as well — more stock, more debtors — and a plan that shows cash rising in lockstep with revenue has typically forgotten it.

A useful discipline before presenting any forecast: compare the first forecast year against the last three actual years. If the step change is dramatic and unexplained, the whole model will be discounted rather than the single assumption corrected.

Using both together

The methods answer different questions and are most useful when they disagree. Where a DCF materially exceeds a comparables valuation, the model is assuming something the market currently does not believe — which may be a genuine insight or wishful growth assumptions. Where comparables exceed the DCF, the market may be pricing in strategic value the cash flows alone do not capture, or the forecast may be too conservative.

Neither divergence is a problem to be resolved by picking a favourite. It is the most informative output of the whole exercise, and it tells a seller precisely which part of their story they need evidence for.