Warrant

I. A number that looks fine

A founder sends a model. Year three revenue: $4.2m. The arithmetic is clean, the tabs are labeled, the chart slopes the right way. Nothing about it is dishonest.

The useful question is not whether $4.2m is correct. Nobody knows that, including the founder. The useful question is what would have to be true for $4.2m to be reachable, and it almost always decomposes into three or four propositions doing all the structural work:

• that a sales cycle observed twice holds at fifty times the volume
• that churn stays flat while the customer profile changes
• that a named channel produces leads at a cost that has been assumed rather than measured
• that one person can carry delivery until month eighteen

Everything else in the model is downstream arithmetic. Those four lines are the building. If the second one is wrong by a factor of two, the $4.2m is not a forecast, it is a decoration.

This is the whole discipline in miniature. Find the propositions holding the weight. Establish what entitles anyone to assert them. That entitlement is what I call warrant.

II. Where the word comes from

Warrant is a legal instinct before it is a commercial one.

Before a court reaches the merits of a claim, it may have to answer prior questions. Does this claimant have standing or the right to bring it? Does this court have jurisdiction? Is the claim in time? Do the pleaded facts disclose a legally cognizable claim? A claim can be entirely true and still fail, because truth and entitlement to be heard are different properties. Warrant precedes merit.

The same structure appears everywhere serious people work:

Mathematics. A theorem’s warrant is its derivation. The conclusion is uninteresting on its own; the chain from the axioms is the object of value.

Physics. A quantity is warranted by measurement, by dimensional consistency, by conservation. A result with the wrong dimensions has failed before its numerical precision becomes relevant.

Economics. A market price is warranted by an executable transaction. A valuation is warranted by the assumptions and method from which it is derived. Confusing the two is how opinion acquires the appearance of fact.

Governance. Power is capacity — the ability to compel an outcome. Legitimacy is warrant — the grounds on which that capacity is properly exercised. Constitutional argument repeatedly turns on that distinction: who may exercise power, under what authority, subject to what limits.

Different vocabularies, one structure. Ask what entitles the assertion, and a great deal of confident material reorganizes itself into two piles: propositions with grounds, and propositions carried by momentum.

Business language conceals this better than most. A pitch deck compresses a sequence of claims into a format designed for rapid acceptance rather than premise-by-premise examination. A financial model is an argument wearing the costume of a calculation. A term sheet is a set of contingent entitlements that most founders read as a valuation.

III. Where warrant fails in a real business

A surprising amount of business failure begins with a premise that was admitted without inspection and then capitalized as though it were fact.

The recurring sites:

Revenue. A conversion rate borrowed from a case study about a different market, a different price point, and a different buyer.

Cost. Monthly operating cost estimated from the visible items. Rent, salaries, software. The invisible ones — payment processing, insurance, professional fees, the cost of failed hiring, the working capital trapped in a sixty-day receivable — arrive later and arrive together.

Funding classification. A founder says: my father is putting in £200,000. That single sentence describes four legally and commercially distinct events. A gift. A loan. An equity subscription. A revolving credit facility. Each produces different consequences for control, for tax treatment, for the ranking of that money in insolvency, and for what happens at the family dinner in year three. Until it is classified, the sentence has no warrant. It is a sound the business made.

Ownership. Percentages agreed verbally between friends, with no vesting, no leaver provisions, and no answer to the question of what happens when one of them stops working.

Permission. A business model that requires a license, an approval, a registration, or a professional indemnity policy that nobody has confirmed is obtainable in the jurisdiction where the revenue is booked.

Capacity. A plan that requires the founder to work at a rate they have sustained for six weeks and are assuming for six years.

None of these are exotic. They are ordinary, and they are expensive precisely because they are ordinary enough to skip.

IV. The method

The work is a sequence, and it runs the same way whether the subject is a launch, a raise, a restructure, or a dispute.

1. 1.Restate the decision as a claim. Not “we’re launching in Q2.” Rather: “we are entitled to commit £180,000 and eleven months to a launch in Q2.” Now it can be examined.
2. 2.Decompose the claim into premises. List them exhaustively, including the ones nobody has said out loud. Unstated premises are the dangerous ones, because they were never available for inspection.
3. 3.Sort by load. Which premises, if false, take the structure down. Which are cosmetic. Analytical effort belongs on the load-bearing ones, and most analysis fails by spreading itself evenly.
4. 4.Classify each premise by what would settle it. This is the step that determines the commercial shape of everything afterward. A premise is settled by arithmetic, by observation, by a market test, or by a qualified professional holding authority in a jurisdiction. Four different routes, four different costs, four different timelines.
5. 5.Build the structure in writing. A model, a review, a chronology, a risk register, an issue map. The written form matters because it is auditable. A client can inspect, challenge and revise a document. A conversation leaves no equivalent audit trail.
6. 6.Convert to sequence. Decisions have an order. Some questions must be answered before others can be asked, and spending in the wrong order is how capital gets destroyed by people doing sensible things at the wrong time.

Step four is where the method turns on itself, and where the honest part of this article begins.

V. The same test, applied inward

If warrant is the standard, it applies to the person applying it.

So: what entitles the analysis?

Where the work is structural and derivable, the warrant is the derivation. A financial model, a cost decomposition, a break-even, a risk register, a chronology, an issue map. These carry their own justification. Every assumption is visible, every calculation traceable, every conclusion falsifiable by a client who disagrees with an input. The client does not have to trust me. They can audit the argument, which is a stronger position for them than trust.

Where the work is interdisciplinary synthesis, the warrant is demonstrated competence. Economics, mathematics, law, governance, and the physical-systems habit of asking whether a quantity is even dimensionally coherent. This capability was built through sustained self-directed study — primary texts, statute and case law, formal certification including contract law through HarvardX, and the drafting of formal instruments submitted to senior institutions, prepared to the standard those institutions require. The warrant claimed here is therefore competence demonstrated in the work itself: the quality of the sources, the integrity of the reasoning, and the ability of another reader to test the result.

Where an answer must bind, the warrant belongs to someone else. Bar admission, a practicing certificate, a chartered qualification, a regulatory authorization: these confer something categorically different from capability. They confer recognized professional authority, together with the duties, regulatory framework and professional accountability attached to that status. That is not a soft distinction, and no quantity of study converts into it. Treating capability as a substitute for licensure is a category error of exactly the kind this entire method exists to catch.

Which produces the operating rule, stated plainly:

Sovereign Intelligence produces strategic intelligence for a decision maker. Where a question requires an answer that binds, the decision maker takes that question to a professional who holds the authority to answer it.

The article’s own logic demands that sentence. If I claimed warrant I do not hold, the method would be self-refuting, and a reader capable of buying this work would notice.

VI. Why this saves money rather than spending it

The obvious objection: if a lawyer or a tax adviser has to be involved anyway, what has been gained.

Consider how specialist engagements actually consume budget. A founder arrives at a senior adviser billing $400 to $1,000 an hour with a narrative — a story about the business, told chronologically, containing everything that felt important. A substantial part of the opening engagement may be consumed by structuring: converting narrative into answerable questions, reconstructing chronology from scattered documents, and identifying which concerns actually require that specialist’s expertise.

That work is necessary. It also consumes specialist-rate time before the specialist reaches the questions for which their authority is indispensable.

The alternative is to arrive with the structuring already done. A dated chronology. A document register with the gaps marked. Funding classified, or at minimum the classification question stated precisely. A list of questions written so that each one admits of an answer, sorted by which ones actually change the decision.

The specialist then does the irreducible thing: applies authority to a well-formed question. That engagement is shorter, sharper, and materially cheaper, and the answer that comes back is better. A precise question permits a precise answer. A vague question purchases qualification, assumptions, and caveats.

There is a second economy underneath the first. Many premises never need a specialist at all. They need arithmetic, or a phone call to a supplier, or a two-week test. Sorting those out of the pile before the specialist meeting is often the largest single saving available, and it is invisible to anyone who has not done the sorting.

VII. What this looks like on an engagement

The deliverables differ. The posture does not.

A review will tell a client which of their premises is carrying the weight, and will say so directly, including where the answer is unwelcome. A model will expose its assumptions on the first tab, because a model that hides its inputs is a persuasion device. A risk register will name the risks that have no owner. An adviser checklist will state the questions in the form a specialist can answer, with the jurisdiction identified, because a tax question without a jurisdiction is not yet a question.

And where the honest finding is that the structure does not hold — that the funding classification is unresolved, that the license is not obtainable, that the unit economics require a conversion rate nobody has ever achieved in that market — the finding is delivered as found. A client paying for analysis is entitled to the analysis, particularly the part they were hoping to avoid. The adviser is not retained to preserve the premise.

That is the offer, in one line: the grounds of the decision, made explicit, in writing, before the money moves.