Advanced Critical Chain · Lesson 7 of 8
Congestion management
Even with WIP control, full kit and daily management, work can pile up at one department. Measure congestion in days, map it, set limits, and balance capacity before projects stall.
Sequential priority, WIP control, full kit and high-frequency management cut most of the wastage in even the most complex project environments. So how do we take performance to the next level? Is there still hidden wastage?
Where flow still stops
Take ten projects in WIP, sorted by priority, with the rest in the waiting bucket. Full kit and daily management are followed by the book. Does that mean every one of the ten is progressing without interruption?
Zoom in. Every project passes through several departments, or work centres, and their capacity and speed differ. Each department allocates its people in order of sequential priority:
- Department 1 has people for projects 1 to 5. Projects 6 to 10 wait in its queue.
- Department 2 can staff all ten.
- Department 3 reaches project 6; 7 to 10 wait.
A project waiting in a department’s queue isn’t flowing: it is congested, or stagnated. Despite overall WIP control, work can pile up at local work centres, and if nobody notices, the waiting delays grow. That information is the next pot of gold.
Measuring congestion
Congestion is easy to measure in days:
Days of congestion = work waiting in the queue ÷ the department’s flow rate
If department 1 completes 5 units of work a day and 100 units are waiting, it has 100 ÷ 5 = 20 days of congestion. Do this for every department and plot it: that’s a congestion map.
Some congestion is normal. With one WIP rule across several functions, variation in capacity, scope and interruptions will always create small queues. Congestion beyond the noise needs attention, so set a congestion limit for each work centre and watch the trend.
Acting on congestion
A rising trend at one work centre is an early warning. Resource managers can respond before projects stall:
- Overtime at the congested centre
- Shifting resources across departments, where skills allow
- Temporary hiring
- Multiskilling, so people can move to where the queue is
- And where a critical resource has spare time, selling that capacity for extra revenue
Pause and think: three projects turn red at the same time, and all three are waiting at Fabrication. What should management do first?
- Run a separate recovery plan for each of the three projects
- Treat it as congestion at Fabrication: add or move capacity there
- Re-prioritise the three projects to the top
- Wait for the monthly review
When several projects are stuck at the same work centre, the problem is systemic. Fixing the queue at Fabrication helps all three, and every project behind them.
Check your understanding
Answer all five. Get four right to complete the lesson.
1. A department completes 4 units a day and has 60 units waiting. Its congestion is…
- 4 days
- 60 days
- 15 days
- 240 days
60 ÷ 4 = 15 days.
2. Why can congestion exist even with WIP control?
- Departments differ in capacity and speed, so work piles up locally
- WIP control doesn’t work
- Projects have no priority
- Full kit causes it
One WIP rule across functions with different capacities leaves local queues.
3. A congestion limit is used to…
- Stop work at a department
- Measure people’s efficiency
- Set task deadlines
- Separate normal noise from congestion that needs attention
Some queue is normal; beyond the limit, act.
4. Which is a good response to rising congestion at one department?
- Opening more projects
- Moving or adding capacity there: overtime, multiskilling, shifting people
- Ignoring it
- Raising WIP
Balance resources to the congestion.
5. Several red projects all waiting at one department call for…
- A separate rescue for each
- A change of priorities
- A systemic fix at that department
- Cancelling them
Check congestion before acting on single projects.