A RAG study shows where the bottleneck’s time goes
We watch the bottleneck and sort every minute into Red, Amber or Green. That gives us the data to find and rank what holds the line back.
RAG study
A RAG study sorts every minute at the bottleneck into Red, Amber or Green
A RAG (Red, Amber, Green) study is a production line study. We watch the bottleneck, log every stop by type and by asset, and record speed changes.
We find the bottleneck first. Studying a step that is not the constraint tells you little about the output of the whole system.
The study works for people and for machines. We use it most to see what stops a team at the bottleneck from producing more.
Four colours cover all the time we observe
| Class | What it means | Examples |
|---|---|---|
| Green | Productive: the person or machine is making output. | Running at speed |
| Amber | Essential but non-productive: needed, but it makes no output. | Changeovers, cleans, quality checks, mandated safety steps |
| Red | Non-productive: neither productive nor essential. This is the improvement opportunity. | Breakdowns, waiting for materials or people |
| White | Not studied: time nobody observed or recorded. | Gaps in the study |
Efficiency = Green time ÷ (Green + Amber + Red time)
White time stays out of the calculation, because nobody saw it. Red time, plus any Amber time you can reduce, is the most you can gain.
We run a study in five steps
- Find the bottleneck. We use capacity analysis across the process steps, and look at where work in progress builds up.
- Set up the observation. We agree what Red, Amber and Green mean for this operation. The observer needs a clear view of the bottleneck, usually for several shifts.
- Collect the data. We classify the time as it happens, log each stop’s type, asset and reason, and record every speed change.
- Analyse the time. The share of Green time is current efficiency. Amber is essential non-productive time, and Red is pure waste.
- Find and rank the barriers. We list what takes time away from Green, then rank each barrier by how often it happens and how long it lasts.
An illustration: crew size. Five people run a bottleneck, and the study shows 80% Green, 10% Amber and 10% Red time.
If the causes of Red time are removed, four people could make the same output. That is 20% fewer people and 25% more output per person.
This only works once the barriers are really gone. Cut the crew before that and output falls, so we prove it with a trial first.
The tool times the line, logs stops by type and asset, logs speed changes and exports the data.
When a study shows a loss but not why it happens, we use Split Solving to find the root cause.