Mini-MBA Part 1: The numbers I stopped outsourcing

Financial literacy is being able to read the report. Financial fluency is being able to rebuild it – and therefore to refuse it.

There is a moment familiar to anyone who has carried a mid-market P&L. The management report lands. Gross margin is a point below last year. You know, roughly, why. You say so, the meeting agrees, and nobody in the room can decompose that point into mix, price, freight and purchase timing while the meeting is still running. So the question moves to next month, where it will arrive one month older and no better answered. Field Notes #008 took one of those components apart in detail; this piece is about why the decomposition so often cannot be done at all.

That is not a competence failure. It is a fluency failure, and it has a cost that never appears as a line item: every decision the company defers because the argument cannot be constructed in the room is a decision taken by default.

Of everything a general management programme puts in front of you, this material transferred most directly and least glamorously. It is not new theory – all of it sits in a good corporate finance textbook. What changes is the speed at which you can produce a defensible number under pressure, in front of people who are also numerate. That speed is the whole product, and since it is the one part of an MBA that genuinely can be handed over in writing, the rest of this piece does exactly that.

All figures below are illustrative, chosen to make the arithmetic visible. Use your own inputs.

Tool 1 – The hurdle rate, and the spread that tells you whether you are creating value

Ask a mid-market management team what the company's cost of capital is and you will usually get the interest rate on the last credit facility. That answer is not slightly wrong; it is structurally wrong, and it misallocates capital in a predictable direction.

The number you want is the weighted average cost of capital: the blended cost of every euro financing the business, weighted by how much of each you use.

WACC = (E/V × Re) + (D/V × Rd × (1 − t))

where E = equity, D = net debt, V = E + D, Re = cost of equity, Rd = pre-tax cost of debt, t = corporate tax rate.

Debt is multiplied by (1 − t) because interest is deductible, so the state pays part of it. Belgium's standard corporate income tax rate is 25%, with 20% on the first €100,000 of profit for qualifying small companies, so a Belgian mid-market company borrowing at 4.5% is really paying about 3.4% after tax.

The cost of equity is the part people skip, because there is no invoice for it. The textbook route is the capital asset pricing model:

Re = Rf + β × ERP (+ a size or illiquidity premium for a private company)

Rf is the risk-free rate – take the yield on a long-dated government bond in your own currency. The equity risk premium is the extra return equity investors demand over that; published estimates for Western Europe have sat broadly in the 5–6% range in recent years, and you should source a current figure rather than trust mine.

Beta measures how much your industry's returns swing relative to the market. A private mid-market company has no observable beta, so you borrow one from listed comparables in your sector and then add a premium – commonly 2–4 percentage points – for size and for the fact that nobody can sell your shares on a Tuesday.

Worked, with illustrative inputs: risk-free 2.5%, equity risk premium 5.5%, beta 1.0, size premium 3.0% gives a cost of equity of 11.0%. Debt at 4.5% pre-tax becomes 3.4% after tax at 25%. Financed 60% equity and 40% net debt, the WACC is 8.0%.

That single number does three things immediately.

It sets a floor no project may duck. Every investment inside the company is competing against 8.0%, whether or not anyone has written it down. When it is not written down, approvals default to payback period – and payback systematically picks the wrong projects, because it ignores everything that happens after the money comes back.

Take two proposals. Project A costs €500,000 and returns €100,000 a year for ten years: payback five years. Project B costs €300,000 and returns €150,000 a year for three years: payback two years. Payback ranks B first, comfortably. Discounted at 8.0%, A is worth +€171,000 of net present value and B +€87,000. Payback has just told you to prefer the option worth half as much. Repeat that ranking error for a decade and you have built a company that hums along and never compounds.

It prices the owner's own money. Retained earnings feel free because they never send an invoice. They cost 11.0% in the example above, which is more than the bank charges. Any cash sitting in a current account earning less than that is not prudence; it is a deliberate loss with a rate attached.

It gives you the only value test that matters. Compare your return on invested capital with the hurdle:

ROIC = EBIT × (1 − t) ÷ invested capital, where invested capital = net working capital + net fixed assets (equivalently equity + net debt).

Economic profit = (ROIC − WACC) × invested capital

A company with €4.0m of EBIT and €30m of invested capital earns a ROIC of 10.0%. Against an 8.0% WACC, the spread is +2 points, which on €30m of invested capital is about €0.6m of genuine economic profit a year. Growth is only worth having while that spread is positive. A company growing revenue with a negative spread is destroying value faster the better it sells – which is the single most common expensive misunderstanding in a mid-market boardroom, and the reason "we grew 12%" is not by itself good news.

Tool 2 – The growth-payout identity, and why the dividend conversation is a strategy conversation

In owner-managed companies the dividend discussion is usually the shortest item on the agenda and the one with the longest consequences. It is treated as a tax and liquidity question, at the end, often outside the board. There is an identity that makes the trade-off unavoidable:

g = ROE × (1 − payout ratio)

The rate at which a company can grow without raising external capital equals its return on equity multiplied by the share of profit it keeps. It is arithmetic, not strategy, and it settles arguments.

A company earning a 15% ROE and distributing 20% of profit can self-fund 12% growth. The same company distributing half its profit can self-fund 7.5%. If the plan on the table calls for 12% growth and the shareholder wants a 50% payout, the plan requires external capital and the board has just discovered that before the year rather than after it. That is the entire value of the identity: it converts two separate wishes into one visible constraint.

Underneath ROE sits the decomposition that explains where your returns actually come from – the DuPont identity:

ROE = net margin × asset turnover × financial leverage

Its real use is diagnostic. A distributor running a 3% net margin, turning assets three times, with leverage of 2.0, earns an 18% ROE. A manufacturer running a 10% margin, turning assets 0.9 times, with the same leverage, also earns 18%. Identical returns, opposite machines. It follows that a distributor benchmarking itself on margin is measuring the wrong term, and that anything which slows asset turnover – a warehouse of slow stock, a receivables book drifting out – attacks its economics far more than a point of price does. Knowing which of the three terms your business actually runs on tells you which meetings matter.

Two ratios bound how far the leverage term can be pushed, and both are worth having on a standing basis rather than discovering them in a covenant discussion: net debt / EBITDA, and interest coverage measured as EBITDA / net interest. Lenders set thresholds on both, they vary by sector and cycle, and the number to know is not the market convention but your own headroom against the covenants you have actually signed.

Tool 3 – Throughput, the constraint, and what one day of working capital is worth

The most useful operations idea I met is also the least fashionable, and already forty years old: Eliyahu Goldratt's Theory of Constraints, still best absorbed through the novel The Goal. It replaces conventional cost accounting with three measures:

• Throughput = sales revenue − truly variable cost (the cost that genuinely varies with the unit sold: materials, freight, commission – not allocated overhead)
• Inventory / investment = money tied up inside the system
• Operating expense = money spent turning inventory into throughput

And it delivers one conclusion that contradicts most management reporting: optimising any part of the system that is not the constraint does not improve the system. Goldratt's illustration is a set of robots that are always busy, always hitting utilisation targets, and destroying money – because work released faster than the constraint can absorb becomes stock, not sales. Two corollaries follow, and they are worth memorising: an hour lost at the constraint is an hour lost by the entire company, and an hour saved anywhere else is a mirage.

The method is five steps, in order: identify the constraint, exploit it (get everything possible out of it before spending), subordinate everything else to it, elevate it (now spend), then repeat, because the constraint will have moved. Most mid-market improvement programmes skip straight to elevate, which is why they cost money and change nothing.

In distribution the constraint is almost never the warehouse, though the warehouse is where the KPIs live. Usually it is working capital, one supplier relationship, or the fact that four decisions a week route through one person's calendar. Working capital has the advantage of being measurable in a single line:

CCC = DSO + DIO − DPO – days sales outstanding, plus days inventory outstanding, minus days payables outstanding.

The reason to put it on the board agenda is what one day is worth. One day of cash conversion cycle is roughly annual revenue divided by 365: at €50m of revenue, about €137,000 of cash per day. Taking nine days out of the cycle releases €1.2m – money that arrives with no dilution, no covenant negotiation and no interest, and which most boards will spend three meetings trying to borrow instead. That is the whole argument for treating working capital as a financing source rather than an operational hygiene metric.

Tool 4 – Scoring a judgement so it can be argued with

The habit that outlasted the individual models is more general. When a decision rests on judgement – a country's macro risk, a supplier's fragility, a market entry – you can make the judgement auditable without pretending it is a measurement.

The method is deliberately crude and takes an afternoon. List the factors that would actually change the decision, grouped into four or five themes; aim for twelve to twenty lines, because fewer hides the trade-offs and more becomes an exercise. For each line write a one-sentence definition of what a high score and a low score mean – this is the step people skip, and skipping it makes the scores incomparable between people. Score each line 1–5 for probability and 1–5 for impact, multiply for a risk score out of 25, sum, then divide by the theoretical maximum to normalise to a percentage.

The version I built for an investment question had sixteen lines across four themes, maximum 400, and came out in the mid-forties. As a number that is almost useless. As an argument it did three things a discussion could not. Two lines scored at the top of the range and dominated the total, so the conversation went immediately to those two. Anyone disagreeing with the conclusion had to disagree with a specific line and say why, on the record. And because the method was written down, the same question could be re-scored a year later and the movement compared. A scored judgement is not a measurement. It is a disagreement made specific, and repeatable.

The toolkit in one table

What did not transfer

Honesty about this matters more than the toolkit, because these failure modes are what make MBA-trained operators tedious in their first year back.

The eighty-minute case. A business school case arrives with clean exhibits and resolves before the bell. Real decisions arrive with the exhibit missing, and building it costs three weeks of a controller who has a month-end. The skill that actually matters – deciding what precision a decision requires before commissioning the analysis – is not what the case method trains. It trains the opposite.

Valuation precision theatre. Most of the value in a discounted cash flow sits in the terminal assumption, which means most of the apparent rigour is a growth rate you chose. In mid-market transactions price is set by a multiple, a negotiation and the seller's alternatives. The model is necessary to know your own walk-away point. It is not the reason the price lands where it lands.

The assumption of infrastructure. Every formula above silently assumes a data layer that can produce its inputs. In the mid-market the binding constraint is usually that the ERP cannot produce margin by customer by product without a week of manual work. Frameworks do not fix that. Two years of dull data discipline does, and it comes first.

What to do Monday morning

Compute your WACC on one page – inputs, method, date – then ask your CFO to do the same without seeing yours; if the two numbers differ materially, you have found why capital allocation feels arbitrary. Calculate your cash conversion cycle and multiply one day by revenue over 365, then put that euro figure in front of the management team, because it reframes a hygiene metric as the cheapest financing available to you. And run the growth-payout identity against next year's plan and the shareholder's dividend expectation; if g comes out below the plan, you have found a capital conversation that is better held in January than in October.

The five numbers for the financial-fluency dashboard

One page, reviewed annually at board level.

1. WACC: one number, with method and date, agreed between CEO and CFO.
2. ROIC minus WACC – the spread, in percentage points, and the resulting economic profit in euros.
3. Cash conversion cycle in days, against the same month last year, with one day expressed in euros.
4. Self-funded growth rate g, next to the growth rate the approved plan assumes.
5. Percentage of board decisions in the last year with a number attached in the minutes.

A management team that can produce these five without a project has fluency. One that needs three weeks to produce them has just discovered its constraint, and it is not in the warehouse.

None of this makes the decision for you. It makes the decision contestable by someone other than the person proposing it. Numbers tell you what should be done; they tell you nothing about whether anyone will do it – which is where a rollout I was certain was correct met a regional leader who was not persuaded, and where part two starts.

Photo: Unsplash