Flow, process and networks

Cumulative flow diagram

Shows the number of items in workflow states over time to reveal work in progress and bottlenecks.

  • Chart
  • Workflow monitoring
  • Trend over time

What it answers

Where is work accumulating in the workflow?

Shape
Stacked workflow-state areas
What it shows
Backlog growth and queues.

Example

The question

Where is work piling up in our delivery process?

Work items by stage, last 12 weeks (cumulative flow)

Work items by stage, last 12 weeks (cumulative flow)Stacked bands by stage. The testing band widens from 4 to 24 items while other bands stay steady: work is piling up in testing.Work items (cumulative)04080120160123456789101112To doIn progressTestingDone
Work items by stage, last 12 weeks (cumulative flow)Stacked bands by stage. The testing band widens from 4 to 24 items while other bands stay steady: work is piling up in testing.DoneTestingIn progressTo doWork items (cumulative)040801201601357911
Stacked bands by stage. The testing band widens from 4 to 24 items while other bands stay steady: work is piling up in testing.

Illustrative data. It belongs to no company.

When to use it

Use it when

Monitoring work in progress and bottlenecks.

Avoid it when

When states overlap or snapshots are inconsistent.

How to read it

Read band thickness for each state's count at matching dates.

Note

Bands normally represent state counts despite the cumulative name.

The data you need

Layout
Long table
Fields
Date: date; State: text; ItemCount: nonnegative integer
One row is
One workflow-state count at one snapshot date. What is the grain?
Rules
Use mutually exclusive states and common snapshot dates. Define which items belong to the population.
Optional
Raw alternative: ItemID,State,ValidFrom,ValidTo or complete transition history.
Preparation
Validate the input grain and missing values. Use mutually exclusive states and common snapshot dates. Define which items belong to the population.
Units
Bands normally count items in mutually exclusive workflow states at each snapshot. Total counts arepopulation snapshots.
Example of the input layout: Long tableA long table: one row per observation, few columns. New data adds rows.DateStateItemCountAABBC
A long table: one row per observation, few columns. New data adds rows.

Worked example

Snapshot ToDo 20, Doing 5, Done 10 totals 35. At another date Doing 10 with stable inflow may indicate a growing in-progress queue.

DateStateItemCount
2026-01-01ToDo20
2026-01-01Doing5
2026-01-01Done10