Building a Remit With No Mandate

Overview

Module 1 stated the mechanism: if you do not define the role, it gets defined by accumulation. This topic is how to make the accumulation deliberate when nobody has granted you anything.

The central claim, and it changes the tactics entirely:

A remit is not granted. It is conceded — people stop contesting your involvement in a class of decision after you have been useful in it several times.

Which means remit-building is not a negotiation. It is a sequence of small, visible, low-risk usefulnesses, followed by writing down what has already become true.

Five moves, in order

graph LR
  A["<b>1 · Find a vacuum</b><br/>nobody owns it,<br/>nobody is defending it"] --> B["<b>2 · Be useful</b><br/>answer the question<br/>nobody could answer"]
  B --> C["<b>3 · Be cheap<br/>to involve</b><br/>a message, not a meeting"]
  C --> D["<b>4 · Write down<br/>what is already true</b><br/>precedent, not a claim"]
  D --> E["<b>5 · Decline<br/>deliberately</b><br/>so the remit has edges"]

  classDef early fill:#7D9A4E,stroke:#5F7A38,stroke-width:2px,color:#ffffff
  classDef mid fill:#5B8FB9,stroke:#4A7A9E,stroke-width:2px,color:#ffffff
  classDef late fill:#E8A33D,stroke:#C4832A,stroke-width:2px,color:#ffffff
  class A,B early
  class C mid
  class D,E late

Five moves. The order matters — step four fails if attempted first, because asking for authority invites a refusal that then exists on the record.

1 — Find a vacuum, not a contest. Contested territory costs credibility you have not yet earned. Uncontested gaps are free to occupy, and most organisations at the size that hires a first architect are full of them: nobody records decisions, nobody owns cross-team interfaces, nobody can describe the whole system, nobody has done the arithmetic on the business case.

Occupying a gap annoys no one. Taking a decision away from someone who had it annoys them permanently.

2 — Be useful before being authoritative. Do the unglamorous thing well. Write the document nobody wrote. Answer the question nobody could answer. Bring the number nobody calculated — Module 1’s arithmetic is exactly this move, and it works because it is useful to the person who holds the budget rather than impressive to engineers.

3 — Be cheap to involve. If consulting you costs a meeting and a week, people will decide without you and tell you afterwards. If it costs a message and you answer usefully within a day, you get consulted — and being consulted is how you see decisions while they are still cheap to change. This is unglamorous and it is most of the job in the first six months.

4 — Write down what is already true. After you have done something three times unchallenged, put it in the charter. Now you are describing a state of affairs rather than requesting a power, and agreement to a description is vastly easier to obtain than agreement to a claim.

5 — Decline deliberately. Module 1’s “what I do not decide”, and it is not politeness. A remit with no edges absorbs everything until you are accountable for everything and empowered for nothing. Declining a class of work also buys something concrete from the people most inclined to resent the role: the engineers who feared you would take their autonomy.

The antipattern: asking for authority

The move that feels correct and fails.

“Can I be the approver for cross-team technical decisions?” invites a no — and a no is now a recorded position that somebody has to reverse rather than merely update. Worse, it frames you as someone seeking power rather than solving a problem, which is the frame you least want in an organisation with no architecture practice.

The same outcome arrives free by doing the work and writing it down afterwards. Slower, and it does not fail.

What this costs

Being honest, because the costs land before the benefits.

It is slow. Quarters, not weeks. A remit built this way is real at six months and solid at a year.

It looks like you are not doing architecture. For the first months you are writing documents, doing arithmetic, answering questions and being present. Someone will ask what you have actually decided, and the honest answer is “very little, deliberately”. Module 1 warned that building domain understanding looks like slow starting; this is the same cost in a different currency.

You will be judged on that in month three. Which is why your sponsor needs to understand the strategy in advance — not as a defence, but so that your visible outputs are legible to them as progress rather than as preamble.

The vacuums, and the first three usefulnesses

The case states the position plainly: no architecture board, no practice of recording decisions, no mandate describing what you decide, no formal authority over any team.

Every one of those is a vacuum, which is the good news.

What is uncontested and valuable: nobody records decisions. Nobody owns the interfaces between the monolith and the Go estate. Nobody can describe what the whole system does. Nobody had done the arithmetic on the business case.

What is owned and should be left alone for now: the monolith’s internals, which are Ivan Petrov’s. Service implementation choices, which belong to the teams. Product scope, which is Priya Shah’s. Taking any of these early would spend credibility to acquire territory you do not need.

The first three usefulnesses, each answering a question nobody could answer:

1. The cost arithmetic. The finding that the company-wide $3.00 target cannot be reached by deflecting 38% — residual human cost alone exceeds the blended target. That is useful to Dev Raghunathan on day one, it costs nobody anything, and it establishes that you bring numbers rather than opinions.
2. The residency analysis. Three gates rather than one, with the third — can it be evidenced — producing a requirement on the architecture. Useful to Lena Fischer, who has blocking authority and whose criteria nobody had written down.
3. The noise-floor correction. The regression trigger as specified sits inside the measurement’s resolution. Useful to Maya Ortiz, who owns the accuracy floor, and to Ivan, because it demonstrates that the safeguard against a quiet repeat is being checked rather than assumed.

None of those required authority. All three are the kind of thing that, done three times, makes “the architect does the arithmetic on initiatives” a description rather than a request.

What to decline, explicitly and early: implementation choices inside team boundaries. Written into the charter as I do not choose which libraries or storage engines teams use; I set the constraints they must satisfy. It costs nothing, it is true, and it is the sentence that makes Tomas Beck’s engineers stop treating the role as a threat.

Feeds the workbook

This is the operational half of your positioning statement, and the rubric requires a concrete plan of action rather than an aspiration. Name the vacuums, the first three artifacts you would produce, and what you will decline in writing.

Check your understanding

1. Name three vacuums in your own organisation. Which is cheapest to occupy, and who would object?
2. What is the first useful thing you could produce that nobody else can? Why can nobody else produce it?
3. Someone asks in month three what you have decided. What is your honest answer, and who needed to understand the strategy beforehand?

---

→ Positioning and remit assume an organisation that merely lacks a practice. Some have shapes that actively defeat the role. — Organisational antipatterns