Every material claim in your report carries what would prove it wrong.
An independent examination of your business model — where it holds, where it strains, and the single assumption everything rests on — read against the conditions your business will actually meet. Delivered within two weeks.
Six situations, one question underneath.
You’re preparing to raise, and you’d rather find the weak joint yourself than have an investor find it for you.
You aren’t sure the model holds, and everyone close to it is too close to say.
You’re about to commit another year, and you want to know what you’re committing it to.
You’re scaling what works, and you want the machine examined by someone with no stake in the answer.
The business earns, and you’re about to change it — a new product, a new market, a new model. Your numbers describe the business you have. The bet is on one that doesn’t exist yet.
You’re choosing between two directions, and everything written about either was written by someone who prefers one.
In every case the question is the same: does this hold, and how would I know?
Seven things you leave with, whatever else you decide.
Each is an artefact of the method — it exists because of how the examination is built, and it leaves with you whether the findings please you or not.
The strongest version of your own argument
Where the model holds is stated with the same rigour as where it strains. A finding that your position is real and defensible carries the same evidence standard as one that says it isn’t.
In practice: you’ll know which of your claims survive scrutiny, and you can stop hedging them.
Your keystone assumption, named
Most founders can name their product and their market. Few can name the one assumption whose failure would end the business. Yours is identified, written down, and set beside the one you named yourself.
In practice: you’ll watch one variable instead of worrying about all of them.
Your evidence agenda, in priority order
Every material finding names the observation that would settle it — a question to put to the market, or a figure to pull from your own data. Read together, ordered by how much rests on each.
In practice: you’ll know exactly what evidence would change the analyst’s mind — and yours.
An exit on every finding
Each finding names the event that resolves it — datable, verifiable, recorded when it occurs.
In practice: every finding in the report is a door, and each one states what opens it.
Reasoning traceable to your own words
Each finding cites the answers it rests on. Every claim can be followed back to the passage that produced it.
In practice: if you disagree with a conclusion, you’ll know exactly where to challenge it.
Independent scrutiny, permanently referenced
The report carries a permanent reference, verifiable against the platform — ready for a data room, a bank file, an application.
In practice: a business plan states what you believe. The record shows what has been tested.
A register that stays alive
The report is a snapshot. The findings register lives in your workbench and accrues its record as findings are addressed and resolve.
In practice: the work you do after the report gets captured too.
Two practical questions probably arrived before you got this far.
usually sooner.
Founding price for the first fifteen examinations, Standard or Advanced. Standard examines a business that exists as reasoning; Advanced examines one that exists as evidence — revenue, users, retention, a record that can be read. An operating record multiplies what must be cross-examined, and the fee follows the record: a record that shows paying customers or revenue is examined as Advanced. Every examination includes a year of full access to the register. It opens on payment, and the year counts from delivery. Both are explained below.
Reserve a founding examination now, and write the record after →
Everything above describes what leaves with you. Everything below shows the instrument that produces it.
Every finding arrives in the same shape.
Whether it concerns your evidence, your positioning or your execution, a finding carries the same fields. Learn to read one and you can read the whole report.
“Zero switching cost is presented as ease of adoption; it equally means nothing holds a user who doesn’t find what they came for.”
An illustration of the form.
The report speaks in six kinds of finding:
The assumption you identify yourself, before anyone names one for you.
The assumption your own filed answers point to.
Where the two differ, the gap is written down. It is usually the most useful sentence in the report.
The report runs eight sections, §0 to §7. The keystone, in §0, and the fragility ranking exist only when a person writes them; the platform refuses to draft either.
The instrument is the same for every business. What differs is the evidence it has to read.
§5The two examinationsTwo different question sets, one instrument.
Standard runs outward-in: the world first, then your offer, then what defends it. Advanced runs evidence first and interpretation last — because when customers have already arrived, the argument has been settled by them.
Standard
For a business that exists as reasoning — pre-launch, prototype, or early build, with little or no operating data yet.
$500
- It runs outward-in: the world first — the incidents that prove the problem is real, what people use instead today, who the first hundred actually are, and why this is possible now.
- Then the offer: your delta against the substitute in use, what it costs, and how the transaction works.
- Then what defends it: what stops a funded competitor copying it, and the smallest version worth building.
- The keystone comes last, and unassisted — you name the assumption everything rests on before anyone names one for you.
Advanced
For a business with operating evidence — revenue, users, retention, a funnel, or a delivery record that can be read.
$1,000
- It runs evidence first: who arrived, who stayed, what they displaced to use you, and where you’ve lost people.
- Then the machine: whether one unit makes money, where the funnel leaks, what the growth engine actually is, and where revenue concentrates.
- Then what breaks as it grows — the limits that only appear at the scale you’re planning for.
- Customers arriving already answered whether the problem is real. Your own case closes the examination instead — interpretation last, after the evidence has spoken.
Both receive the same instrument. Eight sections, the findings register, what-would-prove-it-wrong on every material claim, the keystone divergence, the question channel, the second reviewer’s signature, and the citable record.
The report ends here. The work it creates begins here.
§6After the reportEach finding changes something in the business.
The report closes; its findings stay open. Every finding converts into something operational — evidence to gather, a metric to watch, a decision to make, positioning to correct, work to remove. What the conversion produces depends on the evidence your business already has: the same finding hands a pre-launch founder a question for the market and an operating founder a metric to instrument.
Your evidence agenda, in priority order
Read in severity order, the falsifiers are the evidence worth gathering next — questions to put to the market, or metrics to pull from your own data. Most founders leave with a fortnight of specific, answerable work.
Your milestone plan, already written
Each resolution is stated as an event: datable, verifiable. For an operating business, those events double as thresholds. Either way, it is the milestone plan you were going to have to write — for an investor, a partner, or your own quarter.
An early-warning system for the business
Named, the keystone becomes a metric with a definition, an owner, and a place in whatever you already review weekly — a dashboard, the CRM, the Monday meeting. You know what could break the business, and you know early when it begins to.
Your positioning, corrected this week
A claim you present as a strength that the evidence — or a careful reader — takes as a weakness becomes an immediate change to your deck, your pricing, your next customer conversation. These findings need no new evidence to act on, and they pay back fastest.
The gaps in your measurement system
Each absence names a figure the record couldn’t source, why it matters, and what producing it would settle. Read together, the absences map what your current data cannot answer — and what to change in how you collect it, so it can.
What comes off your roadmap
Patterns surface only when the whole record is read at once — every answer against every other, and each against the evidence behind it. They most often reveal effort aimed at a problem the evidence leaves unsupported. What a pattern removes from the roadmap is usually worth more than anything it adds.
Independent scrutiny, permanently referenced
The examination leaves a permanent, verifiable record: what was examined, what was found, and the evidence each finding rests on. It enters a data room as an independent examination, a bank file as an external assessment, a programme application as examination already completed. A business plan states what you believe. The record shows what has been tested.
You’re about to hand over the complete record of your business. Here is exactly who can see it.
§7ConfidentialityWhat you file is held under conditions that are built, not promised.
- A mutual non-disclosure agreement stands before any review begins — automatic, not on request, and not a clause you have to ask for.
- Your submission is content-hashed at filing: a provable record of what existed, when it existed, and that it hasn’t been altered since.
- A submission marked confidential is excluded at the database level. It doesn’t enter a public query for anyone — not redacted in one, absent from all of them.
- The detail view of your project answers to exactly one account: yours.
- Nexum’s revenue is examination fees and seat fees. There’s no venture arm, no portfolio and no fund — no mechanism by which your model becomes anyone’s asset here, and no interest in one.
Nothing travels.
Your report, your findings and everything you filed are visible to you and to the desk that produced them — and to no one else, at any point.
Most analysis assumes the business exists in ideal conditions. Yours won’t.
§8Examined where it actually isA business doesn’t exist in the abstract, and isn’t examined as though it did.
Generic analysis reads a model against an imaginary market with frictionless payments, uniform regulation and unlimited capital. That analysis is cheap to produce and worth what it costs. Yours is read against the conditions it will actually meet.
Jurisdiction
The regulatory perimeter your model sits inside or beside, and what operating there requires — licensing, permissions, thresholds you may be near without knowing it.
Rails
The payment and banking infrastructure genuinely available where you operate, and what it costs. A model that assumes rails you can’t access is a model with a hole in it.
The real market
The population that exists, the frequency it buys at, and whether the arithmetic clears — computed, not asserted.
Capital and cost
What raising means in your geography, alongside the talent and cost structure you actually build within.
An ambiguity about which market you’re in is never treated as a detail. Two jurisdictions can turn the same model into two different businesses — different rails, different perimeter, different arithmetic — and an examination that skips the question is examining neither.
The output only makes sense because the order never changes.
§9How it is producedOne analyst takes responsibility for your examination.
Your case has a named analyst from filing to signature. They hold the full record, construct the analytical frame, put the questions to you, and sign the findings. The method they run is Nexum’s own — a controlled sequence, enforced by the platform, that holds every case to the same standard and lets the evidence decide what each one produces.
- The whole record, examined at once.Before a word is written, the analyst reads the complete record as a system — each answer against every other, each claim against the evidence behind it, the stated model against the conditions it depends on. The findings that matter most live in the relationships: a tension between pricing and channel, a pattern across five answers that no single answer shows.
- The analytical frame, filed before the draft.The analyst then constructs the frame the examination will run against: how the business makes money, what it depends on, where it is most likely to break, and what evidence would test each point. The platform holds the draft locked until the frame is filed. Every finding that follows answers to it.
- What the record can’t answer is asked, not assumed.A channel opens between the analyst and you: up to five questions, put directly, answered in your own words. Your answers join the case and are cited in the findings that follow. An ambiguity that matters is resolved by asking you — and a finding grounded in your own answer is one you can argue with rather than one you have to accept.
- Machine-scale analysis, under the analyst’s judgment.AI gives the examination a capability no reader has: the entire record held in view at once, every answer, claim and figure cross-referenced continuously against the frame. The frame directs the machine. And the two sections that carry the report’s judgment — the keystone assumption and the fragility ranking — accept only the analyst’s hand; the platform returns “analyst input required” to anything else.
- A second signature.A second reviewer, uninvolved in producing the report, reads the record and the report and signs before delivery. Reviewers are analysts of the house. Fellows read the register; they do not sign verdicts.
“A report with zero uncomfortable findings is presumptively defective.”
The reviewer’s standard.We can’t show you a track record. We can show you the instrument.
Nexum is new. No client counts, borrowed logos or testimonials appear on this page — they don’t exist yet, and a service whose whole premise is telling you checkable things shouldn’t open with something you can’t check.
What we can do instead is uncommon here: show you exactly what arrives — the form of a finding, the keystone pair, the sections, the method and the order it runs in — and let you judge the instrument itself. Every claim on this page is inspectable. Read it that way.
The first examination is published in full.
A complete examination — the report, its findings register, the keystone divergence and the question record — delivered to a founder and published with the founder’s consent. This page shows the instrument’s form; that one shows it used.
One practical question remains: what happens afterwards, and what if it isn’t for you.
§11The rest of the termsTiming
Your report is delivered within fourteen days of filing. Most arrive sooner.
The window depends on the complexity of the business and on how much your own record raises — a case with several ambiguities takes longer than one with few, because the questions have to be asked and answered before findings can rest on anything. The window for your case is stated when you commission it.
If it isn’t for you
Refunds: a full refund at any time until your record is accepted for examination. After delivery, if the report misses its fourteen days or the standard set out in the method, the house corrects it or refunds it. Refund policy →
What we don’t refund is discomfort. The report may find that the model as it stands doesn’t yet hold — that’s the instrument working, and the falsifiers and exits are exactly where its usefulness sits at that moment.
Bring the model. We’ll bring the questions.
It begins with one question, and what you answer determines which examination your case takes. It ends with a signed record of where your model holds, where it strains, and what would settle either.
Begin the examination