Schedule compression: crashing vs fast tracking for the PMP
Crashing adds resources to time-dependent activities to finish them faster, which raises cost. Fast tracking turns sequential links into partial ones, for example finish-to-start into start-to-start, so work runs in parallel, which adds risk and can reduce quality.
The techniques and their side effects
GAO's table of recovery and acceleration techniques lists crashing (add resources to time-dependent activities; costs more and may reduce quality if work is rushed with less-experienced labor) and fast tracking (reduce sequential dependencies to partial ones; resources may become overallocated, quality may drop and risk is introduced when work best done in sequence runs in parallel).
The same table lists splitting long activities so parts run in parallel, reviewing constraints and lags, revisiting duration estimates against actual progress, adding overtime (costs more and lowers morale over time), reducing scope (some requirements cannot be cut), and using schedule contingency (using it too early lowers the chance of finishing on time).
Where to apply them
GAO says recovery should focus on critical activities, and in particular on long-duration critical activities. Critical activities drive the key milestone dates, so the overall duration falls only when the critical path is shortened. Shortening many small tasks on non-critical paths does not move the finish.
Before any change, archive the schedule. After it, document the technique used, recalculate the critical path and check it is reasonable.
Why more people is not always faster
Duration is roughly the work divided by the people available: 16 hours of work takes 2 days for one person and 1 day for two. GAO warns that doubling resources does not necessarily halve the duration. Complex activities may need more time as people are added, because communication and coordination grow.
Fixed-duration activities do not shorten at all. Testing a satellite in a vacuum chamber takes the same number of days however many engineers watch it. Adding people to such an activity raises its cost and leaves its duration unchanged.
Where the exam catches people
Believing crashing has no downside if the budget allows it: GAO lists higher cost and possible loss of quality.
Believing fast tracking is risk-free because no money is added: running sequential work in parallel can reduce quality, add risk and overallocate resources.
Believing any shortened activity brings the finish date in: only shortening the critical path does.
Believing twice the people means half the time: coordination grows on complex work, and fixed-duration work does not shorten at all.