Abrigo·Marrón Systems

La Práctica The Practice

Research and implementation are the same work, done in a different order.

Research takes a question into the field and brings back a finding. Implementation takes a finding into the field and brings back a question. Running both is the only way we know to keep either one honest.

Dos Vías Two Tracks

One practice, two directions of travel.

Vía 01 — Investigación

Research

Structured inquiry into how real organizations govern themselves, conducted where the stakes are physical. Method is qualitative and systematic: collect what practitioners actually say and do, code it, compare constantly, and let the categories emerge instead of arriving with them pre-written.

Output is written and public — field notes here, and longer work where it belongs. We publish negative results. A framework that only ever confirms itself is not a framework.

Vía 02 — Implementación

Implementation

Diagnostic and design work inside an organization that already knows something is wrong and has usually misidentified it. We establish context, map which governance functions are genuinely being performed, name the deficiencies, and design into the constraints that actually exist.

We hand over something the client's own people can run without us. An intervention that requires its author to stay is a dependency, not a fix.

Dominio 01 Management

Does this organization find out in time?

That is the whole of it. Every management system is, structurally, a set of feedback loops: something is measured, the measurement travels, somebody acts, and the action changes what gets measured next. Loops have properties. They have gain, they have delay, they have a bandwidth beyond which they stop tracking reality, and they have failure modes that are fully predictable once you know the loop.

Most reporting systems fail on delay. The number is accurate and arrives after the window to act on it has closed, at which point the report's only remaining function is to assign blame. Others fail on gain: a small deviation triggers an enormous response, the response overshoots, and the organization oscillates — the familiar pattern of a company that is either ignoring a problem or in an all-hands crisis about it, with nothing in between.

What we look for

  • Where the measurement is taken versus where the decision is made, and how long the trip takes.
  • Whether anyone independently verifies what operations reports about itself.
  • Which decisions are being made by people who cannot see the problem, because the authority sits above the information.
  • What the organization has stopped measuring because the number was uncomfortable.

Dominio 02 Team-building

Fix the coupling and the morale usually follows.

Cross-functional teams rarely break down over personality, although personality is almost always the explanation offered. They break down at the handoff — where one group's output becomes another group's input, where the schedule assumes a response time nobody agreed to, and where two people each reasonably believe the other owns the next step.

Spend the budget on a facilitated offsite and you will get two good days and no structural change, because the structure was never the subject. Make the handoff explicit — who hands what to whom, in what condition, by when, and what happens when it slips — and the interpersonal friction frequently evaporates on its own. It was never about trust. It was about two competent people being set against each other by a seam nobody had drawn.

What we look for

  • Interfaces between groups: defined, assumed, or actively disputed.
  • Functions performed twice by rival parties who each believe they own it.
  • Decisions with no owner, which default upward until they default to no one.
  • The individuals holding the organization together informally — the surest map of where the designed system has a hole in it.

Dominio 03 Systems engineering

The document is where the agreement either exists or does not.

Requirements, interface definitions, cause-and-effect matrices, verification plans: these are governance instruments written in engineering language. They are the durable record of what an organization actually agreed to, and they are the only artifact that survives the reassignment of everyone who was in the room.

Which is also how they fail. A requirement written to satisfy a gate rather than to constrain a design documents a consensus that never existed, and the gap shows up in the field, during commissioning, at the worst possible moment and at the highest possible cost. Honest specification is slow, unglamorous, and the cheapest work in the entire lifecycle.

What we look for

  • Requirements that cannot be verified as written — which means they cannot be enforced either.
  • Interfaces owned by neither side of the interface.
  • Vendor commitments that live in a conversation rather than in a document.
  • Gate criteria that reward completeness of paperwork over resolution of risk.

El Orden del Trabajo How an Engagement Runs

Four stages, in this order, without exception.

  1. Etapa 1

    Context

    Establish the conditions that are actually true: constraints, history, what has already been tried, who has been burned by it. No analysis of a system that is required to hold still for the analyst.
  2. Etapa 2

    Function mapping

    Determine which governance functions are genuinely being performed, by whom, and with what information. Expect to find several performed by one exhausted person and at least one performed by nobody.
  3. Etapa 3

    Deficiency naming

    State the dysfunction in terms the organization can recognize and argue with. Written down, specific, and falsifiable. If the client cannot push back on it, it was not a finding.
  4. Etapa 4

    Design and hand-off

    Design into the real constraints, then transfer it. The measure of success is that the client runs it after we stop showing up.

Los Límites What We Will Not Do

Stated in advance, because it is cheaper for everyone.

  • We will not pre-write the conclusion. If an engagement exists to produce documentary support for a decision already made, we are the wrong firm and will say so in the first conversation rather than the last.
  • We will not price by headcount or by hours of presence. Both pay us for occupying a chair. Scope the finding, not the calendar.
  • We will not recommend before we diagnose. An early recommendation is an opinion in a deliverable's clothing, and it will be remembered as a commitment.
  • We will not build a dependency. Anything we design is documented well enough for the client's own people to operate, modify, and eventually discard.
  • We will not carry a client's proprietary information into published work without written permission and de-identification. The writing matters; it does not matter that much.