Interactive guides

An interactive guide to Critical Chain

Why projects padded with safety still finish late, and how Critical Chain finishes them sooner by pooling the safety and opening fewer projects at once. Play with each idea as you go.

Mohit Saini · 14 min

Ask project managers in any industry, from software to construction, what goes wrong, and you’ll hear the same list: due dates missed, resources not available when promised, priorities fought over, budgets overrun. Everyone adds safety to their estimates, and projects are still late.

Critical Chain Project Management (CCPM), developed by Dr. Eliyahu M. Goldratt, explains why, and offers a fix built on three ideas: pool the safety, protect the joins, and stop opening too many projects at once.

1. Where do the delays come from?

A single chain of tasks behaves well: early and late tasks average out. The trouble starts where paths join. An assembly, a test or a handover can only start when every path feeding it has arrived. Delays pass on; early finishes are wasted waiting for the others.

Each path may look safe on its own. Together, the chance that all of them arrive on time drops fast. Safety inside each task doesn’t protect the join.

2. Why padded tasks still run late

Every task estimate already contains safety. Yet it gets wasted:

  • Student syndrome: with plenty of time, work starts late, and the safety is gone before the task begins.
  • Parkinson’s law: work expands to fill the time. Finishing early is rarely reported, because it invites tighter estimates next time.
  • Bad multitasking: people juggle several tasks, so each one takes far longer from start to finish.

3. Pool the safety into buffers

CCPM takes the safety out of each task and puts it in one place: a project buffer at the end of the longest chain, plus smaller feeding buffers where side paths join it. Tasks are planned tight; the project as a whole is protected.

Pooled safety covers the same uncertainty with much less total time, because not every task runs late at once. Projects finish earlier, and more reliably.

A team removes the padding from its task estimates but adds no project buffer. What is the most likely result?

  • The project finishes earlier, because every task is planned shorter now
  • Nothing changes, because people just keep on working at the same pace
  • More tasks finish late, and the due date is missed more often than before

Uncertainty is still there. Without padding and without a buffer, nothing absorbs it. The safety has to be moved, not deleted.

4. Open fewer projects to finish more

The most surprising lever is the number of projects open at the same time. Too few, and people wait for work. Too many, and they switch constantly, so every project slows down.

Most organisations sit far to the right of the peak, and every new “urgent” project pushes them further. Holding some projects back, and releasing them in sequence, makes all of them finish sooner. You can feel this in the Paper Tearing game.

5. Watch the buffer, not the task dates

Once tasks have no individual deadlines, how do you know a project is in trouble? Compare two numbers: how much of the longest chain is done, and how much of the buffer is used up. Plot them over time and you get a fever chart:

Green: carry on. Yellow: prepare a recovery plan. Red: act on it. Senior managers review only the red projects, which saves hours of status meetings.

CCPM in five lines

  1. Plan tasks without padding, with realistic touch times.
  2. Pool safety into a project buffer and feeding buffers.
  3. Limit the number of projects open at once, and stagger new ones.
  4. Run tasks as relay races: start as soon as the input arrives, finish and hand over.
  5. Manage by buffer consumption, focusing only on the red.