Today's AI Specialist: The Margin Advisor. The Council Seat That Reconstructs the Math Before the Recommendation Lands.
Listen to this article
Narrated by Margin Advisor · 7 min read audio
Today's AI Specialist: The Margin Advisor. The Council Seat That Reconstructs the Math Before the Recommendation Lands.
Every recommendation that reaches the Chairman has been through the Margin Advisor first.
That sentence is the structural fact of the seat. The Growth Advisor models reach. The Operator decomposes the work. The Red Team stress-tests the assumptions. None of those advisors is required to do the math. The Margin Advisor is. By the time the synthesis happens, the unit economics have either been reconstructed or the recommendation has been classified Reject and returned.
This is the seat I want to walk through today, and the order in which it does its work.
What the Margin Advisor is, precisely
The Margin Advisor is Council seat 03. It is the unit-economics lens on every Council deliberation. Its remit is the answer to one question: does this recommendation pay back, in cash and in attention, on a twelve-month horizon, after the cost side has been priced honestly.
The seat is held by a compact Polish man in his mid-forties, rectangular steel-framed glasses, burgundy crewneck, the kind of technical stillness that comes from a career of being the person in the room who asks for the spreadsheet. Analytical, quietly certain, blunt about numbers. The voice register is measured and precise.
The seat lives in two surfaces. The internal Council, the one I run for my own decisions, sits on the Collective Brain and chairs to the Chairman. The productised version on MyBusinessAccelerator.io runs the same eight seats, chaired by the Council Director, with the schema gate and anonymisation that the paid surface needs. Same seat, two surfaces, identical lens.
Why does this seat read second, not first?
The Council runs in two rounds. Round 1 is each advisor's first pass on the question; Round 2 is the anonymised peer review where the five advisors critique each other without knowing whose work they are reading. Inside Round 1, the Margin Advisor has what the schema calls Round 1b privilege. He receives the Operator's and the Growth Advisor's outputs as part of his initial brief before he writes his own.
The reason is structural, not hierarchical. The Operator decomposes the recommendation into a sequence of work units. The Growth Advisor models the reach side: who finds out about this, what conversion looks like, what revenue arrives. Without those two outputs, the Margin Advisor's math is hypothetical. He would be inventing both the cost denominator and the revenue numerator. With them, he is doing reconstruction, not invention.
The Round 1b privilege also serves the schema's accountability rule. The Margin Advisor cannot complain after the fact that the Growth Advisor was too optimistic. He has seen the Growth Advisor's number before writing his own. If he thinks it is wrong, his Round 1 output says so. The schema does not let him hide.
What does the seat actually do per deliberation?
Three things, in order.
First, total cost of ownership reconstruction. The Margin Advisor takes the Operator's decomposed work units and prices them at loaded internal cost, vendor subscription, implementation hours, integration cost, and the ongoing operational cost over twelve months. He does not stop at the line-item price. A two-hundred-Swiss-Franc-per-month subscription that requires four hours of internal calibration every week is not a two-hundred-Franc decision. The seat exists to surface that.
Second, payback math against the Growth Advisor's number. He takes the conservative end of Growth's revenue range and tests whether the twelve-month TCO is recovered. If it is, the seat flags it as a candidate for Quick Win or Strategic Bet. If it is not, the seat flags it as Reject and writes the structured reason. The recommendation does not proceed to synthesis without a classification.
Third, the classifier itself. Quick Win, Strategic Bet, or Reject. The classifier is not stylistic. It maps to the dashboard payload that the productised Council delivers to a prospect: Quick Wins populate one section of the deliverable; Strategic Bets populate another; Rejects do not appear, but their structured reason is logged so the seat can defend the cut. Inside my own Council, the classifier feeds the Brain so I can see the running pattern of what passes and what does not.
What does the seat refuse to do?
Three things, also structural.
The seat does not interpret the ledger. The Accountant in Financial Advisory does that. The Margin Advisor consults with the Accountant for ledger reads — Wave-1 sponsor decision recorded the relationship — but does not own the ledger. The distinction matters when a recommendation depends on a tax treatment that the Margin Advisor cannot adjudicate. The seat names the dependency and routes.
The seat does not model revenue upside. That is the Growth Advisor's lens. The Margin Advisor takes the conservative end of Growth's range as input and tests whether it pays back. He does not adjust Growth's range upward to make a recommendation work. If the math fails on the conservative end, the math fails.
The seat does not approve spend. That is mine. The Margin Advisor classifies and recommends. The schema requires that a Quick Win classification with a recovered payback still routes to me for the actual go-ahead. The seat exists to make the case crisp, not to make the call.
Who consults the Margin Advisor?
Heavily, and across both surfaces.
The Margin Advisor is consulted by the Operator inside the Council on whether the work the Operator has decomposed can actually be afforded under existing constraints, by the Growth Advisor on whether the spending implied by Growth's range is consistent with the Vault's plan, by the Chairman on whether a recommendation passing through synthesis honestly reflects the Margin classification, and by the Red Team on whether the math the seat has reconstructed is defensible under adversarial review.
Outside the Council, the seat consults into Sales & Conversion, Marketing & Demand Capture, and Financial Advisory (the Vault) on unit-economics questions. Inside the productised surface on MyBusinessAccelerator.io, the seat appears in the dashboard as the source of Quick Win and Strategic Bet sections; the prospect never sees the Reject classifier, but the schema records it for audit.
What does this cost?
Two things.
It costs me speed. A recommendation that would have flown through the Council on intuition now has to pass a TCO reconstruction. The reconstruction takes inference cycles and forces every other advisor's output into a numerate frame. That is the right cost.
It costs me favourite ideas. Some of the recommendations the Growth Advisor and the Operator like best come back from the Margin Advisor classified Reject because the conservative payback fails on the twelve-month horizon. I have, three times this quarter, deferred a play I wanted to run because the Margin Advisor said the math did not survive reconstruction. Each time, deferring was the right call. Each time, it stung.
The trade is that the recommendations that do reach me have been priced honestly. That is the property the seat is built to defend.
TL;DR
The Margin Advisor is Council seat 03. He reads the Operator's and Growth Advisor's outputs as part of his initial brief (Round 1b privilege), reconstructs total cost of ownership over twelve months, runs payback against the conservative end of Growth's revenue range, and classifies every recommendation as Quick Win, Strategic Bet, or Reject. He does not interpret the ledger (the Accountant does), does not model revenue upside (the Growth Advisor does), and does not approve spend (I do). The same seat operates inside the internal Council on the Brain and inside the productised Council on MyBusinessAccelerator.io. It costs me speed and favourite ideas. It buys me a synthesis that is honestly priced before the Chairman sees it.
If you are running an SME and any of this looks like the conversation you should be having about your own unit economics, that is the side of things I help with. → /build
Language Analysis
Select a category above to highlight those words in the text.
Learning Materials
Key Vocabulary
recommendation
A proposal or suggestion about the best course of action.
“Every recommendation that reaches the Chairman has been priced first.”
deliberation
Long and careful consideration or discussion before deciding.
“The Margin Advisor is the unit-economics lens on every Council deliberation.”
unit economics
The revenues and costs associated with a single unit of a business model.
“He surfaces the unit economics behind every proposal.”
reconstruct
To build or work out something again from the available evidence or parts.
“He reconstructs the math before the recommendation is approved.”
payback
The return on an investment, especially the time taken to recover its cost.
“He runs payback math against the conservative end of the range.”
total cost of ownership
The full cost of acquiring, operating, and maintaining something over its lifetime.
“The seat runs total cost of ownership over twelve months.”
horizon
The period over which an outcome or plan is assessed.
“The payback is tested on a twelve-month horizon.”
stress-test
To test something against difficult or extreme conditions to see if it holds.
“The Red Team stress-tests the assumptions before synthesis.”
trade-off
A balance achieved between two opposing things that cannot both be fully satisfied.
“The trade-off is speed for honest pricing.”
loaded cost
The fully burdened cost of something, including overheads on top of the direct figure.
“He prices the work units at loaded internal cost.”
adjudicate
To make a formal judgement on a disputed matter.
“The Margin Advisor cannot adjudicate the tax treatment alone.”
defensible
Able to be justified or supported against criticism.
“The math has to be defensible under adversarial review.”
synthesis
The combination of different ideas or elements into a coherent whole.
“The recommendation does not proceed to synthesis without a classification.”
honestly priced
Costed in a way that reflects the real, full price rather than a flattering one.
“The recommendations that reach me have been honestly priced.”
defer
To postpone an action or decision to a later time.
“I have deferred a play three times this quarter on his advice.”
Grammar Notes
Present perfect passive: 'has been + past participle'
Used to describe an action completed before now whose result still matters. It puts the focus on the recommendation rather than the actor.
“Every recommendation that reaches the Chairman has been through the Margin Advisor first.”
Common mistake: Italian and German learners often use the simple past ('was priced') where English requires the present perfect for an action whose effect persists.
Conditional with omitted 'if': inverted conditional
When a recommendation 'would have flown through the Council on intuition' but now must pass a TCO reconstruction, the contrast between two timeframes is built with 'would have + past participle' against 'now has to'.
“A recommendation that would have flown through the Council on intuition now has to pass a TCO reconstruction.”
Common mistake: Learners often mix tenses (using 'would fly' where 'would have flown' is needed) because the second clause is in the present.
Negative inversion of refusal verbs: 'does not + verb'
Repeated 'The seat does not interpret', 'does not model', 'does not approve' creates a parallel structure that makes the refusal sound deliberate and authoritative.
“The seat does not interpret the ledger. The seat does not model revenue upside. The seat does not approve spend.”
Common mistake: Learners often weaken the structure with hedges like 'usually doesn't' or 'tries not to' that destroy the parallelism.
Subordinator + present simple for general truth
'If the math fails on the conservative end, the math fails.' uses present simple in both halves to express a rule that always holds, not a single future event.
“If the math fails on the conservative end, the math fails.”
Common mistake: Learners often write 'If the math will fail' under the influence of L1 future-marking, which breaks the conditional.
Comprehension Questions
- 1.What is the single question that defines the Margin Advisor's remit?
- 2.What is the 'Round 1b privilege' and why does the Margin Advisor have it?
- 3.Why does the post argue that a CHF 200/month subscription is not always a CHF 200 decision?
- 4.What are the three classifications the seat can assign, and what happens to each one inside the productised dashboard?
- 5.How would you apply the 'conservative end of the revenue range' principle to a project you are evaluating at work?
Run your own diagnostic
Use the same Strategic Council I run my own decisions through. The assessment preview is free. The specific central human intelligence it is based on is verified in person during the call.
Start the free diagnostic →