← DolphinUniverse Body of Knowledge

The Second Number - How to Double Your EBITDA Without Doubling Anything Else

Wolfram Müller · 2026-08-06 · 10 min read
EBITDA Theory of Constraints The Second Number - How to Double Your EBITDA Without Doubling Anything Else

The Second Number - How to Double Your EBITDA Without Doubling Anything Else

Wolfram Müller · 2026-08-06 · 10 min read · CC-BY-SA 2026-08-06

Two Harmless Questions

It was supposed to be a quiet Q&A session in our community. Nik Baumann and I were the only two who showed up. So he used the opportunity.

He had two questions. Both sounded harmless.

Question one: in our Constraint Analytics Center, which profit measure did we settle on - EBITDA or net profit?

Question two: what are we actually going to do with Throughput Accounting?

Ten minutes later we weren't talking about profit measures anymore. We were talking about why a method that has demonstrably produced better decisions for forty years has all but vanished from corporate practice.

And the answer had nothing to do with methodology.

The First Surprise: The Machine Is Almost Never the Constraint

Nik is a financial controller. A good one. So he immediately went for the sharp edge of question one:

"What about depreciation-heavy companies? At some point, you face the question 'do we buy the machine, and when?' - and that machine's depreciation is Operating Expense. As a controller, I'd want it in there."

Technically? He's completely right.

But here's what struck me:

In all these years, I have never once run into that problem. Never ever.

Why not?

Because we almost never meet a company whose constraint sits on a machine. Walk into a serious manufacturer - automotive supplier, instrumentation, industrial automation - and ask about the one machine. It doesn't exist. They deliberately run 20–30% protective capacity on their lines.

Full utilisation would mean late deliveries and lost throughput. Nobody does that on purpose.

I remember exactly one real exception: a company painting steel artworks. A paint booth genuinely was the constraint. Except that the people who could spray so that the colour coming out was actually the colour that was specified were scarcer than the booth. The second booth existed mainly so the human could walk from one to the other without losing time.

So the constraint was the human being again.

In modern organisations, the constraint has moved. It sits in sales, in R&D, in industrialisation. Machine capacity is structurally in surplus. And when the constraint moves, so does every investment decision you thought you understood.

The Second Surprise: One Number Runs Everything

Then came question two - and this is where it got uncomfortable.

I asked Nik where our cost accounting actually comes from. Is it defined by physics? Is it justified?

His answer was fair: it is beautifully justified. It is built for financial reporting, inventory valuation and tax. Which costs must we capture, and how do we value them?

I showed him a book. J. O. McKinsey, 1923.

And in 1923 that book was fine! No computers. Barely any cars. The task was to supply an infinitely large market with uniform goods as cheaply and stably as possible.

But then Nik said the sentence that made the whole conversation click:

"If you have one number that is highly prominent - the full costs, the ones you also show the outside world - then for the second number you have to invest additional work. Reconciliations. Checks. Maintenance. And because that costs money, the entire management orients itself on number one."

There it is.

Every company has two numbers. Only one of them gets maintained, reported and believed.

The German controlling textbook is right, by the way. Contribution margin accounting, marginal costing, constraint resources - it's all in there. Nik has it in his bones. But what you actually have to deliver every month is the close. And in the close sits full absorption costing.

So you take what you have. And the second number quietly disappears.

Nik watched this live for three years as Head of Finance for Global Manufacturing. His verdict:

"The textbook is right. The practice has fallen away. It is genuinely scary how strongly that single number shapes the thinking."

I see the consequences in the field. One chemical company introduced Drum-Buffer-Rope on half its lines. Those lines deliver continuously, high output, highly profitable - because it flows. And every single year the finance person sits down in front of the numbers and says:

"These numbers cannot be right."

Because he cannot explain how a line can be profitable while it is not fully utilised.

The managers no longer even know that second world exists. And out of that come wrong decisions, again and again.

The Core: Throughput Accounting in Ten Seconds

Here's the good news. This is not a second set of books. It's the same numbers, pointed at a different question.

Cost accounting asks: what did this cost? A question about the past.

Throughput Accounting asks: what should we do next? Which project to start. Which product to push. Where to invest.

Three formulas, that's the whole toolkit:

1. Throughput

Throughput (T) = Sales (S) − Totally Variable Costs (TVC)

TVC is stricter than it sounds: material, licences, suppliers to pay - and only if you actually sold the thing. Labour is deliberately not in there, because you pay your people whether you sell or not.

2. The goal

EBITDA = T − Operating Expenses (OE)

3. And the one that changes everything

Octane number = Throughput ÷ constraint time consumed

The constraint sets the output of the whole organisation. So every constraint hour is spent exactly once. Octane tells you what each of those hours earns.

Let me show you why this matters. Two projects:

By ROI logic you look straight at B, obviously.

Now check where those 100 days sit. In project A, exactly one day touches the constraint. In project B, ten days do.

Suddenly A has five times the octane of B. And it gets better - you could run ten A's in the constraint time that one B consumes. That's 5 million instead of 2.

Same numbers. Completely inverted decision.

The Numbers Speak for Themselves

This is the part executives lean forward for. Take a company:

Now pull one of two levers.

Lever 1 - 10% more constraint capacity. No new people. No new machines. Just more flow through the constraint. If you can sell it: $110m sales, $55m TVC, OE unchanged.

EBITDA: $10m. Doubled.

Lever 2 - 10% richer mix. Same capacity. Sales simply favours the higher-octane work. Average octane goes from $100k to $110k per constraint day.

EBITDA: $10m. Doubled again.

And in practice, lever 2 is the easier one. The mix is always wide - some projects earn $400k per constraint day, others $10k. You don't have to double anything. You have to move the average by ten per cent.

Pull both levers: 550 constraint days at $110k → EBITDA $15.5m. Tripled.

Here's the asymmetry, and it's the whole point:

Neither move is heroic. Yet each one doubles the bottom line - simply because Operating Expenses don't move with them.

The Three Levers You Can Actually Pull

Everything above collapses into three rules:

  1. Internal projects that touch the constraint come first. Anything that permanently raises available constraint capacity - or permanently lowers the constraint time your work consumes - pays out across the entire future. Every single percent.
  2. When choosing between projects or products, take the highest octane. The average octane rises, and with it the EBITDA earned per constraint hour.
  3. Work that doesn't touch the constraint? Do it. Sell it. It costs you no constraint capacity. Just watch that it doesn't create a new constraint.

And if you only run internal projects? Throughput is easy to calculate wherever something gets sold. For a purely internal portfolio, use Cost of Delay ÷ constraint days as your proxy for octane. Same logic, same power.

The Real Gap: Nobody Can Build the Business Case

Now the part that actually keeps me up at night.

I'm currently in a large ten-year CAPEX project. The controllers understand all of this. But we're not allowed to talk to the controllers. And everybody else does very strange things - because they don't know what the game is about.

Tell a head of development that contribution margin 2 is nonsense and CM1 is the number, and you get:

"I'm an engineer. I don't understand these numbers."

Ask the top managers why they don't explain it, and you get:

"They won't get it."

"Have you ever tried explaining it?"

"No. They won't get it."

So what happens? A middle manager goes upstairs and says:

"Boss, if we buy the machine we can do more."

And the boss says: "Yes… hmm."

Now imagine the same manager said this instead:

"We're currently selling 10% less than we could, because machine X is the constraint. Put a second one next to it and sales becomes the constraint. Because we'd then deliver on time, sales can move 20% more. Revenue minus totally variable costs gives a payback on that machine of two weeks."

And the boss says: "Ah. Understood."

Same machine. Same money. Ten seconds of business language - and a completely different outcome.

Most business cases I see are worth exactly zero, because they never touch the constraint. Or they're cost-saving cases that scrape together a handful of euros. Somebody says "we have to improve software development with AI now"; the executive asks what it does to the bottom line, and the answer is "we'll be faster." What does it do to the bottom line? "We'll be faster."

We do it because everyone else is doing it. It just doesn't deliver anything.

Why This Is Career-Relevant

Here's my honest conclusion after that conversation with Nik.

Throughput Accounting isn't a calculation tool. It's a strategic tool. The moment somebody accepts the constraint and aligns their business thinking with it, all the other initiatives tumble off the list by themselves. They simply drop out.

And for a manager personally?

Being able to juggle these figures daily - and to always see what your action does to the bottom line - is career-relevant. Throughput Accounting is the simplest way to get there.

You don't need a new ERP. You don't need a transformation programme. You need to know your constraint, and you need the second number.

Everything else is arithmetic you learned at school.

Get the Full Story

Nik and I recorded that whole conversation and turned it into a booklet - the complete dialogue, edited for reading, with the primer on Throughput Accounting up front, the full worked example, and the parts of the argument I could only summarise here.

It's called "The Second Number" and it's free.

→ Download the booklet

Inside you'll find the full exchange on EBITDA versus net profit, why German controlling knows the right answer and still can't deliver it, the complete octane calculation, and the business case template that gets an executive to say "Ah. Understood."

So go find your constraint. Ask what each of its hours earns. And then go have the conversation with your CFO - you'll be surprised how simple their language actually is.

Stay sharp. Stay curious. And smile while you're at it.

Wolfram - Your Dolphin!

P.S.: You can join our free community to discuss this further with experts and Dolphins - that's also where this Q&A took place, and where the Throughput Accounting course lives.

P.P.S.: If you want to run the numbers on your own organisation, the Constraint Analytics Center walks you through it: find the constraint, quantify the potential, build the business case.