Published method · for capital and for the companies it backs

We publish how we score, so you can argue with it.

A practice that sells judgment and will not show its working is selling assertion. These are the twenty-four dimensions we assess, what evidence satisfies each, how the scores reconcile, and what happens when they disagree. Two documents, the same method, written from opposite sides of the table.

Two editions, same method

Neither is a summary of the other.

The allocator edition is the method as we run it: what each of the twenty-four dimensions asks, what evidence satisfies it, the scale anchors, the evidence grading scheme, and the rule for what happens when two assessments disagree. It is written for the person underwriting the position.

The founder edition covers the same twenty-four dimensions from the other side of the table: what an allocator is actually testing, what your answer reveals when you do not have the evidence, and what to fix before the raise. It is written for the company being assessed.

Neither is a summary of the other. The allocator edition carries the scoring apparatus; the founder edition carries the preparation. A fund that reads only the allocator edition still does not know what its portfolio company will be told, and a founder who reads only the founder edition does not see the anchors their score will be set against.

Both are available on request. Neither is a public download — not to create friction, but because a method that is argued with is more useful than a method that is skimmed, and we would rather know who is reading it. Ask for either or both below; the link arrives by email and is good for seven days.

For capital

The document a fund manager reads before the first call.

v1.4.2The Assessment Methodology — allocator edition

Four assessments — technical readiness, commercial compulsion, execution capacity and capital readiness — across twenty-four scored dimensions. The evidence grades that constrain a score. The reconciliation rules for when the assessments disagree, and why we refuse to produce a single composite number.

The risk register that follows is built to ISO 31000 and COSO ERM structure, so a finding enters your own reporting without translation. Twenty-eight numbered references. Section 09a states when the full twenty-four apply and when a dimension is scored differently. Section 16 states plainly what the method does not do.

Request the allocator edition — PDF, 21 pages

False-color aerial image of a coastline.

False-color infrared · U.S. Geological Survey · public domain

For founders, and for the people who backed them first

The same questions, read from the other side of the table.

v1.4.2The Assessment Guide for Founders — founder edition

The same twenty-four dimensions, with what evidence satisfies each and what your answer reveals when you do not have it. Most fundraising advice teaches you to be persuasive. This is about being assessable — different skills, and after the first meeting only the second one matters.

Section 3 is written for the angel or seed investor already holding a position. It is the section most likely to be uncomfortable.

Request the founder edition — PDF, 14 pages

What the deliverable looks like

A worked assessment against a composite company.

v1.3.1The sample deliverable

Everything a fund receives at the end of a three-week Stage 1 engagement, shown against Terrafix Carbon — an invented company assembled from patterns we see repeatedly in agricultural-residue pyrolysis. The position page, all twenty-four scored dimensions with their evidence grades, the reconciliation, the dated comparison, the risk register formatted for a board pack, and the evidence register naming what we asked for and did not get.

Most sample deliverables show the provider being right about a good company. That is the easy case. Terrafix has real technology, a real dated trigger, a team that discloses honestly, and an underwriting problem invisible from the deck. It is the case where the method either earns its fee or does not.

Two pages from it

The position, in full. This is the first page a fund reads.

The pattern is a timing bet compounded by a financing decision — high technical readiness against undated commercial compulsion, with a capital plan that runs out fourteen months before the company’s own assumed repricing date.

Two dimensions determine the outcome, and they are not the ones the deck defends. CC-3, because the dated trigger the company cites compels its feedstock supplier and not its buyer. Fifty-eight percent of the revenue model sits on a credit line whose counterparty is a broker rather than an end buyer, with a discretionary budget rather than a program, and no trigger that can be dated. CR-3, because eleven months of cash reaches commissioning of the first commercial unit, and commissioning is only a fundable milestone if offtake exists against it. It does not.

The dated comparison. When the buyer becomes compelled, set against the month the cash runs out. This is the finding most often missed, and the one that turns a good company into a down round.

Open-burn prohibition	In force, dated and enforceable, and the reason the grower cooperative signed. Compels the feedstock supplier.	In force
Runway ends	$6.7M at $610k monthly. Reaches commissioning of the first commercial unit.	Jul 2027
CDR repricing assumed	Company model assumes $180/t from this date. A forecast, not a trigger. No obligation or contracted commitment produces it.	Q3 2028
Earliest dated compulsion	None identified. No instrument creating a durable-removal obligation is currently before a legislature in this jurisdiction.	Undated

The gap is fourteen months at minimum. That is not a commercial finding. It is a financing finding, and it converts the decision from an investment in a company into a commitment to a financing sequence — a different decision, requiring a different reserve, set before the position is taken rather than after.

The remaining eight sections, the twenty-four scored dimensions and the eleven risk register entries are in the document.

Request the sample deliverable — PDF, 10 pages

Who we work for

Stated here as plainly as it is stated everywhere else.

We are hired by capital, not by companies.

We publish the founder-facing version because the questions are better questions when everyone can see them, and because a company that has organized itself against them is a better company to underwrite. Nothing in it is advice to raise or not to raise, and nothing in it is written to make a company easier for us to sell to.

Where the method came from

A position taken in 1999, and what it cost.

In 1999 our founding principal took a substantial personal position in industrial-scale battery storage. The technology thesis was right — it is right today. He met the principals, saw a demonstration, read the plan and the projections, and had no instrument for assessing any of it.

The position failed. The larger cost was the reserve it drained and the positions that could not be taken for years afterward.

Each assessment in this method exists because of something that was not assessable that day.

The full account →

Argue with it

Both documents are reviewed at the end of 1Q2027.

Revision history is at the back of each document. If you think a threshold is wrong, they are built so you can say which one.

[email protected]

False-color aerial image of a mountain range in orange and blue.

False-color infrared · U.S. Geological Survey · public domain

We publish how we score, so you can argue with it.

v1.4.2The Assessment Methodology — allocator edition

v1.4.2The Assessment Guide for Founders — founder edition

v1.3.1The sample deliverable

Tell us what you want to hear about.