RAG status: how to define Green, Amber and Red
Red, Amber, Green is the most common status scale in project and portfolio reporting, and one of the least consistent. One team's Amber is another team's Green. This guide shows how to define each color so it means the same thing in every report.
By the Tarivra team · · 6 min read
Why RAG status drifts
Most organizations define RAG in a sentence: Green is on track, Amber is at risk, Red is off track. That leaves the hard part, deciding what at risk means, to each program manager. Two managers looking at the same facts will often choose different colors.
- Optimistic managers wait until a problem is certain before choosing Amber.
- Cautious managers report Amber for anything with an open question.
- Some teams rate effort and morale; others rate the date.
- Without written criteria, nobody can check whether a color is right.
Rate a few dimensions, not a feeling
A RAG status is more consistent when it is built from a small number of dimensions, each with its own thresholds. Four cover most programs:
Schedule
Will the remaining work finish by the target date at the team's real pace?
- How to check:
- Divide the open items by the average weekly finish rate and compare with the weeks left.
- Red flag:
- Amber when the forecast is within two weeks of the date or slightly past it; Red when it misses by more.
Scope
How much work has been added since the start, without a change to the date?
- How to check:
- Count items created after the start date as a share of the total.
- Red flag:
- Amber above 10% growth; Red above 20% with the date unchanged.
Flow
Is work still finishing at a steady pace?
- How to check:
- Compare items finished in the last two weeks with the two weeks before.
- Red flag:
- Amber when the finish rate drops by a quarter; Red when it halves.
Stuck work
How much of the open work is blocked, overdue or untouched?
- How to check:
- Count open items flagged blocked, past their due date, or not updated in 14 days.
- Red flag:
- Amber above 15% of open work; Red above 30%.
The thresholds above are a starting point. Pick numbers that suit your portfolio, write them down and use the same ones for every program.
Combine the dimensions into one color
- Green: every dimension is Green, or one is Amber with a recovery plan the sponsor has agreed.
- Amber: one or two dimensions are Amber, or one is Red with a credible plan to fix it within two weeks.
- Red: any dimension is Red without an agreed plan, or the target date cannot be met at the current pace.
Writing the rule down matters more than the exact rule. Once it is written, anyone can check a status against it, and a disagreement becomes a conversation about the facts rather than about optimism.
Make Amber safe
Clear definitions only help if people are willing to use them. If an Amber triggers extra meetings and blame, teams will find reasons to stay Green. Treat an early Amber as good management: ask what help is needed, and record the decision you make.
- Thank the first team that reports Amber under the new rules.
- Review the gap between reported and supported status, not the color alone.
- Track how often each team's status matched the work, and recognize the ones that did.
Where Tarivra fits
Tarivra applies checks like these to every program automatically, using the work in Jira, Azure DevOps, ServiceNow, GitHub and GitLab. It puts the color the data supports beside the color the team reports and scores how well they match, so a disagreement is visible the same week.
Questions
What does RAG status stand for?
Red, Amber, Green: a three-color scale for project and program status. Green means on track, Amber means at risk, and Red means off track or in need of help.
What is the difference between Amber and Red?
Amber means a problem is visible but recoverable with the current plan or a small change. Red means the target cannot be met without a decision: more people, less scope or a new date.
Should RAG status be set by a person or calculated?
Both. Calculate the status the work supports from a few written criteria, and let the program owner report their judgment beside it. The difference between the two is where the useful conversation is.