Every status report has a colour. Green, amber, red. It’s the most-shared, least-useful artefact in project delivery — because a single colour is a vote held in private, by one person, after the politics have already been applied.
A RAG status answers the wrong question. It asks “what should I tell the steering committee?” when the question that matters is “what does each person actually believe, and where do they disagree?” Delivery confidence answers that one.
A colour hides the disagreement. A score surfaces it.
Give five people a traffic light and they’ll converge on amber — it’s the safe answer, the one nobody gets fired for. Give the same five people a 0–10 scale across specific dimensions and something different happens: they diverge. The PM scores delivery an 8. The tech lead says 5. The CFO’s budget confidence is a 4.
That gap isn’t an argument to win. It’s the diagnostic.
This is the principle we call divergence as diagnostic: the distance between two honest scores is the most valuable signal you have. A single colour averages it away. A confidence score per dimension, gathered independently, puts it on the table where you can act on it.
Why “confidence”, not “status”
Status is backward-looking: where are we now. Confidence is forward-looking: will this deliver what it promised? That shift matters, because most projects don’t fail on a wrong decision — they fail on no decision, made too late, while a green dashboard quietly reported progress on the wrong thing.
Confidence is also harder to fake. “We’re green” survives a meeting. “I’m a 5 on benefits traceability, and here’s why” starts a conversation that changes the work.
What you actually score
The Delivery Confidence diagnostic scores eight dimensions, each against a concrete 10-out-of-10:
- Scope clarity — is it clear what this delivers, and what it doesn’t?
- Budget confidence — do you know what you’ve bought, not just what you’ve spent?
- Timeline realism — is the schedule honest, or hope with dates on it?
- Risk management — are the real risks on the register, or just the safe ones?
- Team capacity — can the people actually deliver what’s being asked?
- Stakeholder alignment — do sponsors and delivery agree what “done” means?
- Benefits traceability — can you trace the line from work to promised benefit?
- Governance & oversight — is someone actually steering, with evidence?
How to run it without it becoming theatre
The method only works if the scores are independent. Hand the Delivery Confidence Snapshot to each person separately — PM, tech lead, sponsor, finance. Don’t fill it in as a committee; a committee produces amber. Then lay the answers side by side and look for the widest spread. That dimension is where the next conversation lives.
The output isn’t a prettier dashboard. It’s a decision: continue, rescope, or stop — made with evidence, by people looking at the same picture for the first time.