ROWVIN
ROWVIN / METHOD

Four phases,
one judgment

running all four.

Understand, Decide, Build, Learn is not a novel sequence, and ROWVIN doesn’t claim it is. What changes the result is continuity: the judgment that framed the problem is the judgment that commits to a direction, builds the thing, and reads what comes back.

METHOD / 01–04 · cyclical · learn feeds understand

ROWVIN / THE FOUR PHASES

  1. 01 / UNDERSTAND

    The problem you arrive with is rarely the problem.

    Gather the context that already exists — the decks, the backlogs, the dashboards, what people say in the meeting and what they say after it. Separate signal from noise. Name the constraints, the assumptions, the stakeholders, the evidence and the unknowns, and keep the unknowns visible, where an assumption cannot quietly resolve them.

    DONE WHEN

    The problem can be stated in one sentence the people closest to it recognise, and what nobody knows yet is written down rather than glossed.

  2. 02 / DECIDE

    A direction nobody will commit to isn’t a decision.

    Determine what actually matters and what only appears to. Set priorities against the constraints rather than against preference. Evaluate the tradeoffs explicitly, including the ones that make the direction less comfortable to present. Then commit — in writing, with the reasoning attached to it.

    DONE WHEN

    A direction is committed, the tradeoffs it accepts are recorded, and it is clear whose call it was.

  3. 03 / BUILD

    Turn the decision into something reality can answer.

    Design, prototype, implement, integrate, deliver — with the decision and its reasoning still in the room. Execution is a thousand small choices that were never in the brief; they go right when the people making them share the understanding that produced the direction, and go wrong when the brief is all they have.

    DONE WHEN

    Something real exists — real enough that reality can disagree with it.

  4. 04 / LEARN

    Observe what happened, not what was supposed to.

    Watch what the thing actually does. Evaluate the evidence. Identify what changed, what didn’t, and which assumption from 01 turned out to be wrong. The output is an input: it goes into the next decision.

    DONE WHEN

    The next decision is different because of something that was observed.

04 / LEARN → 01 / UNDERSTAND · the loop is the method

ROWVIN / CONTINUITY

Every method has these four phases. Most lose the thread between them.

The phases are not the differentiator — the joints are. Work fails at the handoffs, and it fails the same three ways often enough to name them.

THE HANDOFF
Understanding is passed on before reality has tested it. The people who developed it move on; the people who inherit it get the conclusion without the reasoning, and cannot tell which parts are load-bearing.
THE DISTANCE
The decision is made where its consequences won’t be felt. Nothing corrects it, because the cost of being wrong lands somewhere else.
THE REPORT
Evidence arrives with nowhere to land. It gets read, filed, and changes no subsequent decision — which is the same as never having gathered it.

CLAIM / POSITION · for what exists today, see the record

METHOD / ENGAGEMENT

Every engagement starts with understanding the problem.

We define the shape of the work from there, then stay with it from decision through execution and learning.

ENGAGEMENT / STATED · how ROWVIN works, not a claim about outcomes

ROWVIN / POSITION

A judgment and execution company.

That phrase is doing real work. Judgment is the part that decides what matters; execution is the part that makes it exist. ROWVIN does both, deliberately, because separating them is the failure mode above. It also rules some things out, and the exclusions are more useful than the label.

NOT AN AI COMPANY
AI is leverage inside the work. The capability should show up in what gets delivered, and the vocabulary can stay ordinary.
NOT A CONSULTANCY
A recommendation that stops at the recommendation is where the thread gets lost. The judgment stays attached through the build.
NOT A DEV AGENCY
Work that begins at the ticket takes the hardest question — what should be built — as already answered.
NOT A DESIGN STUDIO
Design is one of the instruments. It is not the frame the problem gets defined inside.
NOT AN AUTOMATION COMPANY
Automating the wrong work makes the wrong work cheaper and more permanent.

POSITION / COMMITTED · what we decline is part of the definition

METHOD / LIMITS

There is a documented case where the method didn’t work. It stays on the record.

ROWVIN runs a structured discovery conversation — a real, repeatable way of working through a problem with the people who have it. It has produced resolutions that the participants themselves confirmed. It has also failed once, in a way worth being precise about.

The conversation runs at a fixed register. It assumes a participant who is already fluent in the language of their own problem. In one documented case that assumption was wrong: the register didn’t adapt to someone meeting the material for the first time, and the conversation didn’t reach a resolution. The case is documented next to the ones that worked.

DISCOVERY / RESOLVED
Used with real external participants. Three documented cases, each resolution confirmed by the participant.
CASES / DOCUMENTED
DISCOVERY / DID NOT RESOLVE
One documented case. The method’s fixed register did not adapt to the participant in front of it.
FAILURE / DOCUMENTED
LIMITATION / OPEN
That adaptation problem is not solved. This page does not claim otherwise, and the case stays written down until it is.
STATUS / UNRESOLVED

FAILURE / KEPT · a method with no recorded limits has not been used enough

The method starts with a conversation.

Not a contact form — the first phase, run on your problem. Eleven questions, and a brief that assembles itself as you answer. Most people finish it understanding their own situation better than when they started, which is the point whether or not it goes further.