Most organisations privately accept that RAG ratings tell them very little. Few remove them. This is the case for a deliberate retirement, and a practical way to carry it out.
The column nobody deletes
Every reporting review I have taken part in has ended in the same place. The templates are rewritten, the dashboards are rebuilt, the governance tiers are redrawn. And somewhere on the new template, usually on the right-hand side, there is a column headed RAG.
It survives everything. New owners, new systems, new consultants, new PMOs. It survives programmes that have failed with every line on the dashboard showing amber. It survives the private acknowledgement, which I have heard from almost every sponsor I have worked with, that the colours do not really tell them anything.
I have come to the view that the column should go, and that removing it is a decision a board has to take deliberately. It will not fade away on its own. This piece is about why it lingers, what it costs, and how to retire it without the programme noticing much beyond better meetings.
1. Why RAG survives
RAG is not kept because anyone believes it is accurate. It is kept because it is useful to almost everyone in the room, in ways that have little to do with the programme.
It is cheap. A colour takes seconds to assign and needs no evidence. It is familiar. Every board member has read a RAG report, and most have chaired a meeting about one. It rolls up neatly, so a portfolio of forty projects fits on a single slide. And it protects people. A project manager who reports amber has said something without committing to anything. A sponsor who receives amber has been told something without being obliged to act.
Nobody has ever lost their job for reporting amber.
That last point is the real reason. RAG is a system in which the cautious answer is always available and always safe. Any replacement has to take that safe answer away, which is precisely why replacements meet resistance.
2. What keeping it costs
The cost of RAG is rarely visible, because it shows up as things that did not happen.
It costs time. I have watched status meetings spend longer debating whether a line is amber or red than discussing what to do about it. The colour becomes the subject, and the decision becomes an afterthought.
It costs warning. A colour changes only when someone is willing to change it, and people are rarely willing until the evidence is overwhelming. By then, the problem is several periods old. The board learns last, and learns late.
And it costs meaning at portfolio level. Roll nine green projects and one red into a programme view and the summary is often amber, or worse, green. The one project that needed the board’s attention has been averaged into the others. You cannot average opinions and get a fact.
3. The objections, and the answers
Propose retiring RAG and you will hear the same five objections. They deserve straight answers, because each contains something true.
“The board likes it simple.” So do I. But a colour is not simple; it is vague. “Forecast to breach time tolerance in March” is a single line that a board can act on. “Amber” is a single word that it cannot.
“We would lose professional judgement.” Judgement is not lost. It moves. It goes into setting tolerances, defining evidence and choosing verifiers, at the start and in writing. A project manager who thinks the calculated status is wrong can still say so, with their name beside the override. What disappears is unsigned judgement made every month by the person with most to lose.
“Our data is not good enough.” Then the RAG was not good enough either. It was simply concealing the fact. A status mechanism that requires real data exposes poor records quickly, which is a benefit rather than a reason to delay.
“People will game the numbers.” Some will try. But a gamed number leaves a trail: a verifier who signs everything, a forecast that moves every month, commitments that shrink to fit. A gamed colour leaves nothing at all.
“Our funder or regulator expects RAG.” Then give them one, derived mechanically from the calculated state and never chosen by hand. The colour becomes an output of the rules rather than an input to the meeting. In my experience, funders are rarely attached to the colours. They want to know the truth early, which is the whole point.
4. How to retire it
The replacement I use is set out in Status by rule: four measured values, four states set by fixed rules, and a board pack sorted worst first. But the mechanism is the easy part. The retirement is a small piece of change management, and it goes better if it is treated as one.
Agree the rules before the numbers. Tolerances, evidence standards and state rules should be settled with the board before anyone sees how their project will read. Agreeing them afterwards turns every rule into a negotiation about a particular project.
Run both systems for one period. Report the old RAG and the new status side by side, and show the board every line where they disagree. The disagreements are the argument. In my experience, they are rarely in the comfortable direction.
Expect an uncomfortable first month. Projects that have sat in amber for some time may appear as forecast breach for the first time. That is not the new system creating a problem. It is the new system reporting one that already existed. The board should be told this in advance, so the first honest report is not mistaken for a sudden collapse.
Delete the column. After the parallel period, remove RAG from every template, not just the board pack. If it survives in the project-level reports, it will find its way back up.
Do not let the colours sneak back. Somebody will shade the new states red, amber and green in a spreadsheet within a fortnight. It is a natural instinct and it should be resisted. Once the states are coloured, people start reading the colour rather than the words, and the old habits return with them.
5. What changes in the room
The most noticeable change is not in the reports. It is in the meetings.
When status is calculated, there is nothing to debate about it. The conversation moves straight to the question that matters: given where this project stands, what decision does it need, and who needs to take it? Meetings get shorter. Escalations arrive earlier, and arrive with a date attached. Project managers stop spending the day before the board crafting the right shade of reassurance.
There is a quieter change too. Teams that once treated reporting as presentation start treating it as record. Once the numbers decide the status, the only way to improve the status is to improve the project.
Once the numbers decide the status, the only way to improve the status is to improve the project.
Letting it go
RAG has had a long career. It made reporting accessible to boards at a time when many had very little visibility of their programmes at all, and that was worth something. But it has outlived its usefulness. It asks the wrong person the wrong question at the wrong moment, and it lets the cautious answer pass for information.
Retiring it costs a few weeks of agreement and one awkward month. Keeping it costs the warning a board needs most, at the moment it needs it. I know which I would rather explain to a board.