The charter and governance: who decides what

Station one established what you inherited. This station is the machine you will run everything else through.

Start from an observation that explains most stalling: projects are not delayed by mistakes. They are delayed by waiting.

“Do we change the chart of accounts?” waits three weeks because it is not clear who owns it. Three weeks in an eight-month project is 9% of the duration, for one decision.

Governance is not bureaucracy. It is the machine that turns a question into an answer within a known period.

The charter — two pages, no more

A charter longer than two pages is not read, and a charter that is not read does not exist.

Eight items, three lines each.

1. Why this project — in the customer’s words rather than yours, with one measure to be taken a year later.
2. What is in scope by name — modules, entities, sites, processes.
3. What is out of scope by name — worth more than the item above it.
4. Milestones and dates — five, not fifty.
5. The budget and what it is spent on.
6. The team: who, and for what share of their time — as a percentage, not just a name.
7. The decision structure — who decides what, within how many days.
8. The success criterion — what would make us say this project succeeded, written now rather than at the end.

Item eight is the one that gets skipped and argued about at the end. A project with no pre-written success criterion is judged by the mood of whoever judges it, and usually by the last complaint they heard.

The decision structure

Three tiers, and every decision drops to the lowest tier that can settle it.

Tier	What it decides	Deadline
The working team	Configuration detail inside approved scope	48 hours
Both project managers	Schedule, resources, priorities	5 days
The steering committee	Scope, budget, major dates	Its next meeting

The rule that makes this work: every decision has a deadline, and breaching it escalates automatically rather than personally.

Escalation is not a complaint. It is a mechanism written into the charter and agreed by everyone in advance, so that when it happens it does not read as an accusation.

The steering committee — meets to decide

Most steering committees meet to listen to a status presentation. That wastes the most expensive hour of the month.

Three rules make it work.

One: the agenda is decisions rather than updates. Instead of “migration status”, “we need a decision: do we migrate two years of history or five? Two options with their schedule and cost impact.”

Two: papers 48 hours in advance. Anybody who has not read does not vote. Write that into the charter and enforce it at the first meeting, or it is never enforced.

Three: authority in the room. Somebody who attends and cannot decide makes the meeting pointless. The customer’s representative must be able to approve scope and budget rather than relay.

Cadence: fortnightly on a six-to-nine-month project. Monthly is too slow, and weekly turns the committee into a working team.

The status report — one page

A long status report is not read, and one that is not read protects nobody.

Six items are enough, weekly.

• The five milestones: on time, or late by how many days — as a number, not a colour.
• Decisions waiting, who they are waiting on, and since when. That item alone is the report.
• The top three risks and what changed this week.
• Key user hours: planned against actual.
• Change requests: the count, their total in days, and the contingency left.
• What I need from you this week — two items at most.

The last item separates a report that is read from one that is filed. A project manager who asks for two specific things every week gets read; one who lists achievements gets read once and then ignored.

And do not use colours alone. “Amber” means nothing, and “six days late with four days of contingency left” means everything.

The decision log

One table, started on day one, never closed.

#	Decision	Alternatives rejected	Who decided	Date	Impact

Its second column is its whole value. Seven months in, somebody asks: “why didn’t we do it the other way?”. With the log you answer in thirty seconds with the alternatives and their impact; without it you reopen the discussion from the beginning and lose a week.

Decisions get logged as they happen, not at the end of a phase. A log filled in retrospectively is filled with what its author remembers, which is not what happened.

What you actually do in this station

• Write the charter in two pages with the eight items, and get it signed.
• Write the success criterion as a number and put it to the sponsor before kickoff.
• Write the decision structure with its deadlines into the charter itself.
• Name the steering committee members, and verify that each one holds the decision expected of them.
• Open the decision log on day one and record today’s first decision in it.

The three commonest mistakes

One: a twenty-page charter. Written and unread, so it governs nothing.

Two: a steering committee with no authority. Decisions get presented and deferred, and the meeting becomes a ritual.

Three: deferring the decision log. Settled questions get reopened, again and again.

The acceptance test for this station

1. Show your charter with the eight items in two pages.
2. State the success criterion as a number and who approved it.
3. Show the decision structure with its deadlines and the escalation mechanism.
4. Name the steering committee and what each member can approve.
5. Show the decision log with at least three decisions and their alternatives.

What comes next

The machine is built. The next station protects what you build with it: scope and change control — why scope expands through small requests, and how a small request is measured before it is accepted.