Why revenue is the wrong unit to present in
A leakage figure is naturally produced in revenue, because revenue is what leaked. But revenue is not the unit the decision gets made in at a company with a sponsor, a board or a credit agreement. Those readers think in EBITDA, and presenting a revenue number to them invites the correct objection that most of it is not profit.
Worse, presenting revenue makes the figure look larger than the value it represents, which is the specific way a good analysis loses its reader. An operating partner who mentally discounts your $3M to $600,000 while you are still talking has stopped evaluating the finding and started evaluating you.
So do the conversion yourself, in the report, and show it. The number gets smaller and considerably harder to argue with.
Incremental contribution margin, not gross margin
This is the step that is most often done wrong, and the error always runs in the same direction. Recovered demand is incremental, which means the right margin to apply is the contribution margin on one more unit of that specific revenue — not the blended gross margin across the business.
Sometimes incremental margin is higher than blended, because fixed costs are already covered and the recovered revenue drops through at near-contribution. Sometimes it is much lower, because serving more volume requires more capacity: another technician, another van, another shift. Which of those is true is a question about where you sit relative to capacity, and it is the single largest determinant of whether a leakage finding is worth acting on.
The honest version states which case applies and why. "We are at 70% utilisation on installation crews, so the first $1.4M of recovered revenue requires no additional capacity and the next tranche requires a crew" is a sentence that makes a report credible. Applying a single blended margin to the whole figure does the opposite.
The chain, with a worked example
Take the $3.0M annualised leakage figure from a diagnostic at a $40M services business. Every input below is either that company’s own or explicitly the reader’s to supply.
- 01Start with recovered revenue: $3.0M annualised, already net of a stated recoverability haircut.
- 02Apply incremental contribution margin. Suppose 45% on this revenue, with existing capacity absorbing it. That is $1.35M of incremental contribution.
- 03Subtract the annualised cost of running the recovered process. Software, licences, monitoring, and any partial headcount. Suppose $90,000. That leaves $1.26M.
- 04Subtract the amortised build cost if your board wants it in the first-year figure. Implementation at this scale typically runs $15,000 to $75,000 across all phases; take $60,000, and the first-year EBITDA effect is about $1.20M with the run-rate effect at $1.26M.
- 05Apply your sector’s EBITDA multiple to the run-rate figure to get the enterprise-value effect. The multiple is yours to supply — it is sector- and cycle-specific and nobody should take one from an article. At a hypothetical 7x, $1.26M becomes roughly $8.8M of enterprise value.
The four places this arithmetic gets inflated
- Using blended gross margin instead of incremental contribution margin. Almost always overstates, and it is the first thing a CFO checks.
- Ignoring the capacity step. Recovered demand that cannot be served is not recovered revenue, it is a waiting list and a service failure.
- Omitting the run cost. Every automated process has an ongoing cost, and a figure presented without it is a gross number pretending to be net.
- Applying a multiple to a found figure rather than a sealed one. Found is what the diagnostic identified; sealed is what was measured afterwards and is always smaller. Capitalising found rather than sealed is the single largest overstatement available in this arithmetic, and it is the one most often made.
What to do when the multiple is unknown or contested
Present the EBITDA effect and stop there, then show the enterprise-value effect as a sensitivity across a range the reader chooses rather than as a single figure. Three columns at three multiples is more persuasive than one column at your preferred one, because it demonstrates that the conclusion does not depend on the assumption you would most like to be true.
The same logic applies to the recoverability haircut and the margin. Any figure in this chain that you would be reluctant to see adjusted is a figure the reader should adjust first.
Why the conversion is worth doing even when it shrinks the number
Because it changes what the finding is for. A revenue figure is interesting. An EBITDA figure is fundable. An enterprise-value figure is the version that gets discussed at a board meeting where the operational detail is not on the agenda.
And because the conversion is the part a sceptical reader would otherwise do themselves, silently, less generously than you would, and without telling you the result. Doing it in the report means you control the assumptions on the record instead of losing the argument in somebody’s head.