Haute Lumière
Commerce · I.02 · MMXXVI · daylight
One page each. A reader who reads only these ten pages has the chapter.
The idea. That sentence is how an organisation describes an asset it has stopped being able to see.
It is not dismissal and it is rarely laziness. It is the ordinary consequence of something working reliably for a long time: reliability makes things invisible. A practice that has never failed produces no incidents, no reports, no escalations and no reviews — and an organisation's attention is allocated almost entirely by exception.
What follows. Invisible things are not funded, not defended when someone proposes to cut them, and — most expensively — not reproduced anywhere else.
Worked example. A maintenance supervisor rebuilds a press instead of replacing it. Eleven years, no incidents. Because there were no incidents, there was never a review; because there was never a review, the other four plants never learned of it.
The move. When you hear the sentence, do not argue. Write down exactly what "that" refers to, and go and cost it. The sentence is a marker: it tells you precisely where to dig.
The idea. The value of a regenerative practice sits almost entirely in what did not happen, and management accounts have no place to record a non-occurrence.
V = (C_alternative − C_actual) + R_captured + O_option
Of those four terms, only C_actual — what you spend on the practice — has a line in the accounts. And it is the one that looks like a cost.
Worked example. The press: C_actual is the £22,000 rebuild programme, visible in the maintenance budget. C_alternative is £50,000 a year of deferred capital, visible nowhere, because deferred capital does not have a line.
Why this matters more than anything else in the chapter. It explains under-adoption without appealing to ideology, short-termism or indifference. The accounting shows the cost and hides the return, and people respond rationally to what they can see. Fix the visibility and much of the behaviour follows on its own.
The idea. Before you celebrate a finding, ask whether the underlying stock is rising, flat or falling.
R(t+1) > R(t) regeneration — the asset is improving
R(t+1) = R(t) maintenance — honest, and worth costing
R(t+1) < R(t) liquidation — a cost being reported as a saving
The trap this catches. A maintenance programme that keeps a machine running by deferring a rebuild it genuinely needs is not regeneration. It is drawing down a stock and reporting the draw-down as a saving — and it is very easy to celebrate by accident when you are hunting for good news.
How to run it. For a machine: vibration, tolerance, downtime, defect rate. For a team: capability, not headcount. For soil: organic carbon. For a supplier: their capacity and their financial health, not your unit price.
The rule. Ask for the trend, not the level. A practice that cannot produce a trend is a practice you have not yet measured, and it should be set aside — not rejected — until it can.
The idea. Some of what a protected relationship gives you is optionality, and optionality is real, valuable and genuinely hard to price.
The supplier you never re-tender typically provides four things the unit price does not capture: out-of-hours response, held stock, early warning when you are about to make a mistake, and a rush order without renegotiation. Together those are an option on your own operational continuity.
The rule. Name it. Do not monetise it. An executive who puts a number on everything is discounted entirely; one who says "this is real, here is how we have used it, and I am not going to pretend I can price it" is trusted with the numbers they did give.
If you must put a floor under it. Use the cheapest defensible method: what would the same optionality cost in the market? Held stock is warehousing. Rush capacity is a call option. Both have prices. Quote the market price of the substitute, say plainly that it is a floor and not a valuation, and stop.
A floor you can defend beats an estimate you cannot.
The idea. The strongest counterfactuals are observed, not modelled.
If your organisation has six comparable sites and one of them does the thing, the other five are your control group. The comparison becomes empirical rather than hypothetical, and an empirical comparison is very much harder to argue away.
Worked example. One unit of two hundred people runs at 9 percent voluntary turnover while comparable units run at 22 percent. That is not a model. It is twenty-six fewer people leaving each year, observed, in the same organisation, under the same policies.
What to look for before you build a model. Multiple sites, multiple teams, multiple periods, or a before-and-after where something changed on a known date. Any of these gives you a comparison you did not have to assume.
Why it is under-used. Organisations with multiple sites are sitting on a natural experiment they have never run, and running it costs a query rather than a study. Nobody runs it because nobody's job is to.
The idea. The most common way to overstate a finding is to treat a deferral as an avoidance.
"Without this we would have replaced the press" assumes the replacement was necessary and imminent. Frequently it was neither.
The arithmetic. If the press would have limped on three more years anyway, you have not avoided £50,000 a year. You have deferred a £400,000 outlay by three years, which at a 9 percent discount rate is worth about:
400,000 − 400,000 / (1.09)³ = 400,000 − 308,900 ≈ £91,000 once
— or roughly £11,600 a year over an eight-year cycle. That is a real number and it is a fraction of the naive one.
The move. Compute both. Present both. Label the assumption beside each.
Why this is a strength and not a weakness. The person who presents the conservative case alongside the optimistic one is the person whose optimistic case gets believed. Volunteering your own strongest objection is the cheapest credibility available in any finance conversation.
The idea. Michael Polanyi's formulation is that we know more than we can tell — and in an organisation, much of what is most valuable is held that way.
The maintenance supervisor knows what the press sounds like when it is about to need attention. The manager with low turnover knows which conversation to have in week three. The buyer knows which supplier will tell them the truth. None of it is written anywhere, and none of it appears in any system.
Why this makes the chapter's method work. Tacit knowledge cannot be extracted by a survey or a process map, because the holder does not experience it as knowledge — they experience it as just how you do the job. It comes out in conversation, in detail, when someone who can act on it asks a specific question and then stays quiet.
The practical consequence. The inventory's value is concentrated in the conversations, not the spreadsheet. Which is why the exercise cannot be delegated to a researcher: it is cheap in hours and expensive in authority.
The idea. The least obvious place to look is the decision your organisation declined to make.
A plant not closed. A line not discontinued. A restructuring proposed twice that quietly did not happen. A decision not taken produces no document, so these never appear in any review — but the person who argued against it usually had a reason, and the reason was usually a stock they could see and the proposal could not.
The characteristic mistake. Filing this under organisational politics rather than under information. It is information. The person who blocked the closure was frequently holding a supplier relationship, a process dependency or a customer fact that never made it onto a slide.
How to find them. Ask, in a room, without blame: "What have we decided not to do in the last five years, and who made that argument?" Then go and ask that person what they knew.
The idea. The inventory becomes financial the moment it enters the management pack — as an appendix, not a restatement.
| Column | Content |
|---|---|
| Practice | The plain name |
| Owner | A person, not a function |
| Basis | The counterfactual, one sentence |
| Conservative value | The number that goes in the pack |
| Trend | Rising, flat, falling — metric named |
| Last verified | A date, and who by |
Why an appendix. Restating the accounts is a two-year argument with your auditors that you will probably lose. An appendix is a management report, requires nobody's permission beyond whoever owns the pack, and does ninety percent of the work: it puts the number, with an owner, in front of the people who allocate, every month.
The one design rule. The conservative number goes in the pack; the optimistic one lives in the appendix's own footnotes. This is what stops the document inflating at every annual revision.
The idea. The inventory's highest-value output is not defending existing practice. It is reproducing it.
value at site A × number of comparable sites
Transfer ──────────────────────────────────────────────
cost of documenting and teaching it
Worked example. A retention practice worth £1.17 million a year in one two-hundred-person unit. Six comparable units. Transfer at only 30 percent effectiveness:
1,170,000 × 6 × 0.30 = £2.1 million a year
cost: documenting a method and funding six afternoons
Why this is the strongest number in the volume. Every other figure in the inventory defends something that already exists — worthy, and not exciting. This one creates something that does not. It is the difference between an audit and an investment case, and it is available to anyone who has done the inventory.
Put it on the front page.
All figures in these briefs are computed in lib/verify.py and sourced in the chapter's Works Cited.