In the critical chain approach, Gantt charts tell a story of precedence – each new set of tasks begins after the preceding tasks end, sequentially defining a critical chain of events, which ultimately determines the best-case project timeline. Well-managed time buffers are used compensate for noisy estimates of task durations. The grand result is the “buffered end date,” a so-called realistic estimate of project duration. Project tasks are managed to their best case duration, and time-buffers are consumed (or not) as the project team executes.
- No location available
-

In my experience, though, projects rarely hit their buffered end-date, because in reality, projects are capacity-limited not critical-chain limited. We simply fail to recognize that projects are capacity limited, because we so much want to believe in our own power to get things done. We attribute long project durations to vendor lead times, the need for successive iterations, and “unpredictability.”
- No location available
-

Resource-leveling the gantt would drive the timeline out. Stop and think about this. Every resource in the plan is fully loaded or overloaded. How many times have project managers told me that they won’t run resource-leveling because it would extend the timeline to an unacceptable date?
- No location available
-