Haute Lumière
Commerce · II.03 · MMXXVI · daylight
For the person working inside a gainshare arrangement — where a defined share of verified improvement returns to the people who created it. This chapter is about boundaries, and a gainshare is nothing but boundaries: what counts as improvement, what counts as the unit, what counts as yours.
A gainshare has four parts: a baseline, a measure, a share, and a period with a verifier. Every one of those four is a boundary, and the arithmetic of your payment is decided by where each one sits.
The chapter's claim, applied to you, is exact: none of those four was discovered. Each was elected, each has permitted alternatives, and somebody is paid by where each one sits. Sometimes that somebody is you. Sometimes it is not. Either way, the only way to know is to read the document.
This workbook teaches one skill: reading the perimeter of your own scheme, and then asking for one specific change to it, with the arithmetic attached.
Exercise 1.1 — Find the four boundaries in your scheme document (2 hours)
Take the scheme document, or your last statement, and write down the answers.
Exercise 1.2 — Find the perimeter clause (90 minutes)
This is the exercise that is specific to this chapter and it is the one almost nobody does.
Ask, in writing: what happens to the measure if the perimeter changes?
The chapter's arithmetic makes the stakes concrete. A firm that sells a plant and buys back the same output can report a 100 percent emissions reduction while emitting exactly the same tonnage. The identical move is available in a gainshare measure, and where the scheme is silent, it is available without anybody breaking a rule.
If nobody can tell you the answer, that is the most valuable thing you will produce this month, and it is the thing to ask for.
Exercise 1.3 — Find where the gain actually comes from (2 hours, appreciative)
Ask three colleagues:
"Tell me about a time something here got genuinely better — not the biggest improvement, the one that surprised you. What made it possible, and who would not have shown up in the numbers?"
Take notes on who would not have shown up. That is the population your measure's boundary excludes, and identifying it precisely is the strongest argument you will ever make for widening the scheme.
Exercise 2.1 — Price your own baseline reset (90 minutes)
Model two schemes over five years, on the same underlying improvement.
improvement of 4 % a year on a base of 100
fixed baseline: gain measured against 100 each year
year 1 4.0 year 2 8.2 year 3 12.5 year 4 17.0 year 5 21.7
five-year total measured gain 63.3
ratcheting baseline: gain measured against last year's level
year 1 4.0 year 2 4.2 year 3 4.3 year 4 4.5 year 5 4.7
five-year total measured gain 21.7
Same work, same physics, same improvement. Roughly a third of the measured gain, because of where one boundary in time was drawn. Multiply by your share percentage to get your own number, and write it down.
Exercise 2.2 — Price the perimeter move (60 minutes)
Take your measure and ask what it would show if the largest cost in it were outsourced tomorrow. Compute two figures:
measured improvement, with the cost outside the perimeter ____
real improvement in the thing the measure is a proxy for ____
If the first is large and the second is zero, your scheme has the gap the chapter calls the outsourcing move, and the fix is one clause.
Exercise 2.3 — Read the chapter's numbers as a worker (45 minutes)
Three figures from the chapter, translated.
| Chapter figure | What it means inside a scheme |
|---|---|
| Scope 1+2 captures 5–8 % of a footprint | A measure drawn at your own team's edge captures a small share of what your work causes |
Chain overcount (n+1)/2 | If several teams claim the same improvement, the sum of claims exceeds the improvement — and finance will eventually notice and tighten everyone's share |
| €4m of payroll moved to distribution raises EBITDA 40 % | The same reclassification available to a board is available in reverse: costs pushed into your measure that were not there at baseline |
The honest negative, and it applies to you. Widening a measure's boundary cuts both ways. A measure that captures more of what you cause also captures more of what you do not control. Ask for a wider boundary only where you can also name the lever. A scheme that measures you on something you cannot move is worse than a narrow one, and asking for it is the most common way a well-meant widening ends badly.
Ask for these in this order. They rise in difficulty and each one makes the next easier to grant.
Clause one — the perimeter statement. The measure is computed on the organisational perimeter as at the baseline date, which is set out in Schedule [x]. Costs nothing, refuses nothing, and makes every later conversation possible. This is the one to get first and it is rarely refused.
Clause two — mandatory restatement. Any change to the perimeter requires the baseline to be restated on the new perimeter before the next computation. This is the boundary lock from the chapter's Operationalize movement, in the language of a scheme rather than a facility. It is the same clause, and pointing that out helps: your employer may already be negotiating it with a lender.
Clause three — a stated baseline term. The baseline is held for [three to five] years, or ratchets on the published schedule in Schedule [y]. Use the arithmetic from Exercise 2.1. A ratchet nobody can see coming is the single most common way schemes die quietly.
Clause four — the named verifier and the right to the working. The computation and its inputs are provided to scheme members with the statement. A share you cannot recompute is a discretionary bonus wearing the word.
Exercise 3.1 — Draft your own one-pager (90 minutes)
One page. Four parts, in this order, and no others:
Give it to one person who can act on it. Not the team, not the forum, not the survey. One person, one page, one clause.
Exercise 4.1 — Get it into the standing document (one action)
A clause in an email is applied by the person who wrote the email. A clause in the scheme rules is applied by whoever has the job next. Push for the rules.
Exercise 4.2 — Recruit the second reader (one conversation)
Find one colleague who will read the annual computation independently and check it against the rules. Two people reading it is a practice. One is a hobby, and the hobby ends when you change roles.
Give them the credit for the first correction you find together. That is how it becomes theirs, and something that is theirs survives you.
Exercise 4.3 — Notice the delight (ongoing)
There is a specific pleasure here and it is worth naming, because it is what keeps people doing this.
It is the moment a long-running grievance turns out to be a definition. Two teams who each believed the other was being credited with their improvement discover that the measure's boundary ran between them and neither had ever read it. The resentment does not need to be resolved. It evaporates, because it was never about either team.
Go and cause one of those. They are more available than you think, and the room afterwards is a different room.
A gainshare is a ledger before it is a payment, and the ledger has four entries. Learn to read them in this order and you can reconstruct any statement you will ever be handed.
What to claim, and what not to. Claim the improvement you can trace to a lever you moved. Do not claim an improvement that arrived because a perimeter changed — not because it would be dishonest, though it would, but because the first time finance traces one of those the whole scheme's share percentage comes under review and everybody pays for it. The chain overcount in the chapter is the mechanism: when several parties truthfully claim the same underlying gain, the sum of claims exceeds the gain, and the correction is always applied to the share rather than to the claims.
The strongest position available to you is to be the person whose claims have never had to be revised. It takes one cycle to establish and it is worth more than any single payment.
Tick what you can answer from a document rather than from memory.
| Yes | Not yet | |
|---|---|---|
| I can state the baseline period and the date it was set | ||
| I know whether the baseline ratchets, and on what schedule | ||
| I can write the measure as a formula | ||
| I know whether the share is of gross or net improvement, and which costs | ||
| I know what happens to the measure if a function is outsourced | ||
| I know whether the baseline is restated when the perimeter changes | ||
| I know who verifies, and when | ||
| I can recompute my own last statement from the inputs |
The fifth and sixth are the ones this chapter adds, and in most schemes they are the two nobody has ever asked. Being the person who asks them, in writing, politely, with arithmetic attached, is a genuinely strong position.
For the meeting where you ask for clause one.
"The scheme is working — the measure is clean and the last computation matched what I could reconstruct. There is one thing I could not find and I think it is a gap rather than a decision: the document does not say which perimeter the measure is computed on.
It matters because if any part of what the measure covers moves outside the business — outsourced, sold, reorganised — the measure changes without anything actually improving. I have run it: on our numbers that would show as a [ ] percent improvement with no change in the underlying.
I have drafted one sentence that closes it. It does not change the share or the baseline. It just states which perimeter we are computing on."
Then stop talking. The clause is one sentence; the arithmetic is on the page; you have asked for nothing that costs anybody money.
Do not ask for four clauses in one meeting. Clause one makes clause two a technicality, and clause two makes clause three obvious. Sequence is the whole skill.