Why FDD deliverables increasingly present EBITDA as a sensitivity range rather than a single number, and how to build one that genuinely helps a buyer price the deal.
Write "normalised EBITDA: €10.2m" on a page and you have made a promise you cannot keep. That figure is not a measurement; it is the sum of a dozen judgement calls - is this cost really non-recurring, is that customer sticky, does this add-back survive scrutiny - each of which a reasonable, well-informed analyst could have played slightly differently. The €0.2m of apparent precision is false. Increasingly, the better FDD deliverables refuse to make that false promise: they present a sensitivity range instead, and understanding why matters as much as knowing how to build one.
A single-point EBITDA number invites the buyer to treat a matter of judgement as a matter of fact, and then to lever, price and negotiate off that spurious precision. A sensitivity range does the opposite - it tells the buyer exactly how much the answer depends on calls that could reasonably have gone the other way, and which ones. That is not hedging. It is the most honest and most decision-useful thing an FDD analyst can put in front of a buyer.
Every normalised EBITDA figure is built on a chain of judgement calls. Individually, each looks defensible. Collectively, they compound: two analysts working from the same data room, each entirely reasonable, can land €1m apart simply because they resolved four or five contested items in different directions. A single number hides that entirely - it presents the analyst's particular resolution of every call as though it were the only possible one.
A sensitivity analysis makes the uncertainty visible and, crucially, bounded. Instead of asserting €10.2m as a fact, you present a range - say €9.6m to €10.8m - and you attribute the range to the specific calls that could swing it. The buyer no longer has to trust that you got every judgement exactly right; they can see the judgements, see how much each one moves the answer, and form their own view. That is a fundamentally more grown-up conversation than "here is the number."
A single-point EBITDA says "trust my judgement." A sensitivity range says "here is my judgement, here is what it hinges on, and here is how much it would move if you disagreed with me." The second is worth far more to a buyer.
Not every line deserves this treatment. The goal is to highlight genuine judgement calls, not to manufacture false uncertainty around numbers that are actually solid - a sensitivity table cluttered with items that cannot really move is as misleading as false precision, just in the other direction. The candidates that genuinely warrant sensitising:
The discipline is to sensitise the three or four items that are genuinely contested and material, and to leave the solid numbers alone.
The mechanics matter, because a range built by waving your hands over "plus or minus 5%" is worthless. You build it item by item, quantifying the swing each contested call contributes, then combining them. Take a base case of €10.2m with three genuinely disputable items:
| Judgement call | Downside view | Upside view | EBITDA swing |
|---|---|---|---|
| Owner's "consultancy" cost add-back | Not recurring - don't add back (€0.0m) | Genuine one-off - add back (+€0.25m) | €0.25m |
| Retention of top customer (18% of revenue) | Contract at risk - haircut margin (−€0.30m) | Renews as normal (€0.0m) | €0.30m |
| Annualising a mid-year acquisition | Part-year only, conservative (−€0.15m) | Full run-rate (€0.0m) | €0.15m |
Resolve all three at the cautious end and you are €0.70m below base; resolve them all optimistically and you are slightly above. That gives a defensible, explained range rather than an arbitrary band:
| Scenario | Build-up | Normalised EBITDA |
|---|---|---|
| Conservative | €10.2m − €0.70m | €9.5m |
| Base case | Analyst's central view | €10.2m |
| Optimistic | €10.2m + €0.25m | €10.45m |
Now the range means something. A buyer can see it is driven overwhelmingly by one call - the top-customer retention - and can direct their remaining diligence budget accordingly, rather than worrying uniformly about a vague ±€0.5m.
Two refinements make the build more honest still. First, the individual swings should not simply be added on the assumption that they are independent, because in a downturn they rarely are: the same demand shock that puts the top customer's contract at risk is likely to soften the acquired business's run-rate too, so the true low case can be worse than the sum of the mid-point swings treated in isolation. Flagging that correlation, even qualitatively, stops the range from understating tail risk. Second, resist the reflex to make every driver symmetric. The owner's consultancy add-back is genuinely one-sided - either it is a defensible one-off worth €0.25m or it is nothing - whereas the customer call is mostly downside. A range that forces artificial symmetry onto asymmetric risks misrepresents the real shape of the uncertainty, and a sharp sponsor will notice. The point of the exercise is to describe the risk as it actually is, not to produce a tidy band that looks balanced on the page.
A sensitivity table that merely says "EBITDA could be between €9m and €11m" without saying why is close to useless to a buyer trying to set a price. The version that works does four things:
The fourth point is where analysts most often stop too early. A range in EBITDA terms is helpful; a range that follows the number through the multiple and into the equity value bridge is what a buyer can actually act on. Leaving the reader to do that translation themselves wastes the best part of the analysis.
A strategic acquirer funding mostly with equity can absorb a modest EBITDA surprise; a financial sponsor levering the deal at a fixed multiple cannot. As set out in the mechanics of an LBO, debt capacity is sized directly off EBITDA, so a downside case is not an abstraction - it is a smaller loan, a bigger equity cheque and a thinner return. A sensitivity range that says "take the conservative end of these three calls and EBITDA is €0.70m lower, which at a fixed leverage multiple reduces debt capacity by roughly €3.5m" is directly actionable for a sponsor in a way a single confident number never is.
| EBITDA case | EBITDA | Debt at 5.0x | Equity required (EV €51m) |
|---|---|---|---|
| Base | €10.2m | €51.0m | - (fully covered) |
| Conservative | €9.5m | €47.5m | +€3.5m more equity |
That table tells the sponsor exactly how much cushion they have and where the risk sits. It is the difference between diligence that informs a bid and diligence that merely describes the past - and it is why sensitivity work features so heavily in the red flags discussion for leveraged buyers, who feel every point of downside directly in their returns.
Building a good sensitivity analysis requires understanding your own numbers well enough to know which assumptions genuinely matter and which are noise. That is a real step up from producing a clean, defensible base case: it demands judgement about your own judgement. You have to be honest about which of your calls are robust and which are one reasonable interpretation among several - a kind of intellectual self-awareness that not every competent modeller has.
It is precisely the sort of work senior analysts get asked to lead, because it cannot be done mechanically. It requires knowing the business, knowing the deal, and knowing where your own analysis is standing on solid ground versus where it is standing on a judgement call that could tip. Master it and you signal a level of thinking that fits squarely with the broader hard skills the role is built on.
This topic tests whether you understand that FDD is about judgement, not just arithmetic.
Interviewer: "Why would you present EBITDA as a range rather than a single number?"
"Because a single number implies a precision the underlying analysis doesn't have. Normalised EBITDA is built on a series of judgement calls - which add-backs are genuinely non-recurring, whether a key customer renews, how you annualise a mid-year acquisition - and reasonable analysts can land differently on several of those at once. A range makes that honest. But the point isn't to hedge; it's to make the analysis more useful. I'd build the range from the specific contested calls, quantify how much each one moves EBITDA individually, and show the buyer that the range is driven mostly by, say, one customer's retention rather than spread evenly across everything. Then I'd give a clear recommendation on where within the range I'd land and why, because the buyer is paying for my view. This matters most on a leveraged deal: a sponsor sizes debt off EBITDA, so if the conservative case is €0.7m lower, that's real reduction in debt capacity and a bigger equity cheque. A single confident number hides that; a well-built range tells them exactly how much cushion they have and where the risk actually sits. So I'd rather give a buyer a range they can act on than a point estimate they have to take on trust."
That answer works because it separates honesty from hedging, shows the mechanical build, and connects the range to the buyer's actual decision - which is the whole reason the range exists.
A single number is comfortable and false; a well-built range is uncomfortable and true. The craft is not in widening the band - anyone can be vague - but in narrowing it to the two or three calls that genuinely matter, quantifying each, taking a clear view, and following the swing all the way through to the price. Do that, and you stop describing the past and start arming the buyer to make a decision. That is the moment FDD stops being bookkeeping about history and becomes advice about a deal.
The Transaction Services Interview Programme (€119.99, one-time) includes a full sensitivity-analysis module - how to identify contested judgement calls, quantify each driver, and translate an EBITDA range through the multiple into debt capacity and equity value, with the interview scenarios above. Enrol today.
Hundreds of candidates prepared their interviews with this programme. Those who landed the role have one thing in common: they worked the cases before walking into the room.
