Analysis graph
The network view: node size, badges, rework edges.
The analysis chart deliberately uses a completely different visual language from the BPMN diagram. It is not there to present the process but to answer: which step is heavy, which step repeats, where the flow concentrates.
How it differs from the BPMN diagram
| BPMN diagram | Analysis chart | |
|---|---|---|
| Shape | Per BPMN notation | Every step is a circle |
| Size | All the same | Proportional to time or cost |
| Colour | By notation kind | By role in the flow |
| Edges | Right-angled bends | Curved, thickness by probability |
| Answers | “How the process runs” | “Where the weight is” |
Reading the picture
- The circle radius is proportional to the square root of the weight, so the area reflects the magnitude correctly — the human eye compares areas more accurately than diameters.
- The number inside the circle is the weight of the step (minutes or money).
- The orange badge in the top right corner is the expected number of runs, shown only when it is greater than 1 — meaning that step is repeated.
- Edge thickness is proportional to probability: the main branch is thick, a rare branch is thin.
- The orange dashed edge looping around underneath is the loop-back flow.
Colour by role
| Colour | Role | Determined by |
|---|---|---|
| Green | Entry point | No incoming flow |
| Red | Exit point | No outgoing flow |
| Orange | Branch point | An exclusive gateway |
| Purple | Parallel point | A parallel gateway |
| Ocean blue | Merge point | More than one incoming flow |
| Blue-grey | A processing step | Everything else |
Switching the weight to cost
The Size by cost button redraws with the radius based on money. Well worth doing: the step that eats time and the step that eats money are usually not the same step. Compare the two pictures and you see immediately where to attack, depending on whether the goal is to shorten or to save.
The per-step metrics table
Right below the chart, sorted by actual time descending:
| Column | Meaning |
|---|---|
| Elapsed/run | The work minutes + waiting minutes you entered — not the work minutes alone |
| Runs | The expected number of times the step runs in one pass of a case |
| Actual elapsed | The product of the two columns above — the number most worth looking at |
| Actual cost | The same thing for money |
What a run count below 1 means
That step only runs on some branches. A value of 0,25 means only a quarter of the cases pass through it. A very expensive step that rarely runs can add up to less than a cheap step that every case has to go through — the “actual” columns exist precisely to show that.
That step only runs on some branches. A value of 0,25 means only a quarter of the cases pass through it. A very expensive step that rarely runs can add up to less than a cheap step that every case has to go through — the “actual” columns exist precisely to show that.
Anything unclear or missing?