A PMO dashboard that nobody uses to decide is a mural. Reporting earns its keep when a sponsor changes funding, scope, or sequence because of a number they believe — and can explain to someone else.
Few metrics, each with an action
Leadership will act on a short set: are we still doing the right work, will we hit the dates we promised, is the money still honest, and what decision is stuck? That is portfolio health, delivery confidence, financial variance, and open decisions. Everything else is optional colour.
Define each metric so two PMs would calculate it the same way. 'On track' that means 'the PM is optimistic' in one project and 'earned value is within 5%' in another is how dashboards become fiction. Write the rule. Train it. Reject packs that invent a private definition.
Attach an action to movement. A date confidence that dropped should come with a choice: descope, add capacity, or accept the slip. A metric with no implied decision is a trivia question. Executives stop reading trivia by the third month.
RAG that can be red
Red–amber–green is useful only if red is allowed. Publish the criteria: red means a committed outcome will miss unless something changes this cycle; amber means a risk or issue that will become red without action; green means the plan still holds. If red requires a public hanging, you will get a field of amber forever.
Do not average a project to green. A green schedule with a red benefit owner is not a green project. Show the split — time, cost, scope, risk, benefits — and let the worst material dimension set the conversation. Composite scores exist to hide the thing that will embarrass someone.
Explain the last change. 'Amber because the vendor slipped two weeks; decision needed: mock the interface or move UAT.' Movement plus a request is reporting. A static traffic light with a paragraph of plot summary is a novel.
Treat the pack as a product
One portfolio page for the board. A deeper cut for the PMO and sponsors. Do not send the same 30-slide deck to both and hope they scroll. If a director needs a briefing, write the briefing. If they need the register, link it. Attention is the scarce resource.
Cadence beats completeness. A fortnightly pack that is three days old beats a perfect monthly archaeology. Lock the data cut-off, name it, and stop chasing late returns that delay the whole set. A PM who cannot file on time is already telling you something.
Retire charts that have not changed a decision in two cycles. Beautiful burn-ups that nobody discusses are how reporting grows until it needs its own project manager. Ask, once a quarter, which slide we would miss. Delete the others.
Course complete — what's next
Report a few defined metrics that imply a decision. Let red be red. Split the dimensions, show movement, and keep a short pack for the people who fund the work.
You now have the spine of a PMO: a model with real authority, governance that can refuse, a portfolio that selects, a maturity habit that fixes one constraint, and reporting leadership can use. Put one of those into a live conversation this week — preferably the no you have been postponing.
Practice this
Status Report
Turn theory into a real deliverable. Generate a status report with the AI tool and adapt it to your own project.
Open the generator