A Value-First Summer · the findings

We wrote down how we could be wrong, before we knew the answer.

In June this season committed to three claims and, for each one, the exact thing that would prove it false. That is the part that makes it a test instead of a story. The claims and their losing conditions are below. The findings land when the season closes on the equinox — not before, because deciding early is how a test stops being one.

Read the three claims The season’s story

How to read this page

A claim you cannot lose is not a claim.

Most things written about a way of working are shaped so they cannot fail. They describe a method, list what went well, and leave no way for a reader to check. This season tried the opposite: three claims, ordered from the most grounded to the riskiest, and for every one of them a stated condition under which we would have to say it broke.

The claims stack. If the first one fails, the two above it fall with it. If the second fails, any gain from the first is real but nobody can say why, so it does not travel. The third is the one most likely to break, and the team said so in June rather than afterwards.

Each claim below carries how strong its evidence can possibly be. None of it is proof. One person over one season, and a small group running a simplified version, is evidence — and saying which is which is the difference between a finding and a pitch.

The three claims

Each one, and the way it loses.

The findings are blank on purpose. They fill on September 22, 2026 — the equinox, written from what the season showed — and a tier that broke gets published as a tier that broke. A season that only publishes what worked is an advertisement.

H1 · Does it work, for one person

Running twenty threads through the board realizes more value, and is healthier to live inside, than the way the same person worked before it — where "before" means ranking everything by importance and picking the top three.

Who and what it covers: One operator — Chris — across one season, solstice to equinox.
How strong this can possibly be: One person, one season, no control group. This is evidence, not proof, and it cannot be generalized from this tier alone.

This is the most grounded of the three and the easiest to check, which is why it goes first. If it breaks, the two above it fall with it.

What would prove this wrong

F1aThe count of things actually realized does not improve across the season — August or September fails to beat June, measured from the baseline read on the day the season opened.
F1bThe way the operator engages trends unhealthy — rolling out of streak-guilt rather than interest, returning from gaps with "falling behind" language, treating the board as pressure rather than support. This breaks the claim on its own, whatever the count says. A high count reached compulsively is not a win; it is the slot-machine failure wearing a good result.

What gets read to decide it

The monthly count of what was realized — one quiet figure, read at the end of each month, compared to the month before.
The five strands, read as a shape rather than a score: which threads got energy, which kept coming back, which were confirmed, which surprised.
A qualitative read of engagement health at the end of June, July, August and at the equinox — an assessment, never a grade.

The finding

Not written yet. It lands at the close — September 22, 2026 — the equinox — and it gets written from what the season actually showed, including a “this one broke” if that is the answer. Deciding it early would make the whole exercise worthless.

H2 · If it worked, why

If the first claim holds, the gain comes from three specific things and not from novelty: the board holds the queue so the person does not have to; the roll reaches threads that importance-ranking would never have picked; and months get judged by what was realized rather than by how much was done.

Who and what it covers: Same person, same season — asking what actually did the work.
How strong this can possibly be: Pattern analysis with no control condition. It can show that the data is best explained by these three mechanisms; it cannot prove cause.

Without this tier, a real gain is a black box — true, but not transferable, because nobody knows which part caused it.

What would prove this wrong

F2aThe gain fades as the novelty wears off. If engagement and realization were strong in July and declined through August and September with no other explanation, the simplest reading is that the newness was doing the work.
F2bThe gain does not track the three mechanisms — for instance the count rises while attention narrows to the same few threads, or the surprising rolls produce nothing confirmed. The gain may be real and still not be this.
F2cThe operator's own account of the season does not describe these mechanisms. If, asked at the close, he does not describe the queue being held for him, threads surfacing he would not have chosen, or months felt by what realized — the mechanism story is not supported, whatever the numbers did.

What gets read to decide it

A comparison of engagement quality at week four against week eight — the test for whether novelty was carrying it.
Whether the roll spread attention across many threads or orbited a favourite few.
The operator's own words at the close: does "a productive month" mean realized, or busy?

The finding

H3 · Does it work for anyone else

The three mechanisms carry over to other people in proportion to how much their work already runs the way this one does — many threads at once, activated by interest rather than importance, with the structure held outside their head rather than in it.

Who and what it covers: A small group of practitioners running a simplified version of the same three mechanisms — not the full system, and not a random sample.
How strong this can possibly be: Small group, observed, no control. The riskiest tier by a wide margin, and the season exists partly to test it.

This is the difference between "here is how one person works" and "here is the kind of person this works for, and why". It is also the one most likely to break, and the team said so in June rather than after.

What would prove this wrong

F3aThe group shows no pattern — some benefit, some do not, with no relationship to how they already work. Then the mechanisms are personal, or the profile is too narrow to name.
F3bThe pattern is real but tracks something else entirely — prior frustration with tools, say, or a taste for novelty. The claim survives and its framing is wrong, which is a partial loss and still has to be said out loud.
F3cThe benefit only appears inside the full system. If a simpler version of the same three mechanisms does not work on its own, this is not a transferable way of working — it is a product.

What gets read to decide it

Whether each person reports spending less effort rebuilding context at the start of a session.
Whether threads surfaced that they would not have chosen by importance.
Whether their sense of a good week shifted toward what was realized.

The finding

What the season carried

Twenty-one threads, held at once, on purpose.

Twenty faces carried the season, and each one is named on the board. A twenty-first joined the roster at the close — it is Chris’s to announce, so it is counted here and not named yet.

They are listed in the order the die numbers them, which is not an order of importance — that is the entire design. The die does not know which thread matters most, and neither, on any given morning, does anyone else.

The AI-Native Consultancy
The AI-Native Shift
Compass
AI-Native Humans
The Value-First Collective
Value-First Studios
Value Creation Protocol
Value-First Books
Value-First Design Studio
Value-First Agent
Agent QMS
The Road to UNBOUND
Another Orange Morning
The Value Path Game
The Curation Engine
Chris Carolan
Value Points
HubSpot Academy Bootcamps
HubSpot Mentor Program
Value-First Markets
Named when Chris announces it

What happens next

The close, then the next season’s board.

The season closes on the equinox with a live stream and a published retrospective, and an autumn season on the same model is declared at that close. The findings above fill that day.

After that this page turns over: the summer’s findings stay, and the next season’s board opens beneath them. Same shape, same rule about writing down how you could be wrong first.