WHAT WE BUILD

What we build,
and what we leave alone.

Every company the studio takes on has three things in common: a named buyer, a rule it has to satisfy, and a number that proves it works. Without all three it stays a thesis.

These are the shapes of company the studio builds — the kind of thing we take on, not a list of what we are building.

The shapes of company we build

Three things,
or it stays a thesis.

The line at the top of this page is easy to write and easy to mean nothing by. So here is what establishing each of those three actually involves, and when — all of it before anyone writes production code.

TEST 01 · THE BUYER

A person, not a segment

What it means

Not “mid-market treasury teams”, and not a persona with a stock photograph attached. Someone we can name, at an organisation we can name, who already pays for this problem — either to a supplier today, or to a person inside the building who does it by hand. A problem nobody pays to solve is usually not a problem. It is an inconvenience.

How we establish it

We go and ask, before the thesis is finished. The test is whether one of them will say what the problem costs them now. If a dozen conversations will not produce a single figure, then we have a hypothesis about a buyer rather than a buyer — and the thesis says so, in those words.

TEST 02 · THE RULE

The constraint is part of the product

What it means

In these five markets the hard part is rarely the technology. Moving money needs a licence, or a sponsor who holds one. Custody has a capital treatment. An event contract has to settle against a source that both the regulator and the losing side accept. Treat the rule as something to handle later and it will redesign the product for you, at the worst possible moment.

How we establish it

We write the constraint down in plain language in the first week, with the instrument it comes from, and design from it rather than toward it. Where the position is genuinely unsettled — and in digital assets it often is — that uncertainty goes into the thesis as a named risk with a cost against it, not into a footnote discovered in month nine.

TEST 03 · THE NUMBER

Measured net, and set in advance

What it means

One number, chosen before the work starts and measured after everything it costs to run has come out of it: fees, spread, failed settlements, the support ticket, the person who checks something by hand. Gross figures are easy to produce in every one of these markets and tell you very little. A great many promising ideas here turn out to be a real edge and a slightly larger cost.

How we establish it

The number, its definition and the level it has to clear are written into the thesis — together with the level at which we stop. Both are set before we know the answer, because a threshold chosen afterwards describes the result instead of testing it.

All three can fail, and most theses fail at least one. That is the cheap outcome: it costs weeks rather than years, and the next thesis starts knowing something. The cost is the finding works through one case of the third in full.

And what we leave
to somebody else.

Consumer apps that live or die on paid acquisition. Anything whose whole case rests on a token price going up. Marketplaces in markets we have never worked in. Nothing is wrong with them — we would just be no better at them than anyone else, and the studio only has one team.