2.3Deciding how to compress the schedule

When an activity on the critical path is delayed, or a customer moves the deadline earlier, the PM must choose the best compression option among limited choices. What matters here is not "can it be shortened" but the trade-off judgment of whether to accept added cost for a reliable shortening, or accept added risk from running work in parallel to keep cost down. This section covers the difference between crashing and fast tracking, and the criteria for choosing between them based on the situation.

2.3.1Crashing and fast tracking

• Crashing shortens an activity's own duration by adding cost—more staff, overtime, additional equipment. Since it does not change the sequence of activities, the increase in schedule risk is relatively small, but cost efficiency (additional cost / days saved = the cost slope) worsens, raising the risk of budget overrun. Adding staff also carries a Brooks's-law-like limit: onboarding new members and communication overhead can temporarily reduce productivity.
• Fast tracking reorders activities that would normally start after a predecessor finishes (an FS relationship) so they can run in parallel instead. It incurs almost no additional cost, but the successor proceeds while the predecessor's information is still unsettled, raising the risk of rework (such as a design change not being reflected downstream). It is effective when added cost must be kept down, but should be applied only between activities where the added risk is acceptable.

2.3.2Application principle and the trade-off decision

• Both crashing and fast tracking make sense only for activities on the critical path. Shortening a non-critical activity with float does not shorten the overall deadline; it only adds cost or risk for nothing. Compression targets must always be chosen from activities on the critical path.
• As successive compressions bring the critical path's duration below that of another path (a near-critical path), the critical path switches to that other path. Beyond that point, further shortening the original path no longer shortens the overall deadline, so the full set of paths must be recalculated after every compression to re-identify the new critical path.

In the previous section's example, the critical path was Path 2 (Activity A -> C -> D, 13 days total). The customer now requests the deadline be moved 3 days earlier, and the PM considers compression options. The only candidates are the critical-path activities A, C, and D; shortening the non-critical Activity B (1 day of float) by any amount would have no effect on the overall deadline, so it is excluded from consideration. For Activity C, adding one staff member is estimated to shorten it by up to 3 days at a cost of 200,000 yen per day (a cost slope of 200,000 yen/day). This is a textbook case of crashing: since the sequence is unchanged, the increase in schedule risk is small, but shortening it by 3 days incurs a definite additional cost of 600,000 yen. Meanwhile, Activity D was originally an FS relationship meant to start only after Activity C finishes, but it turns out that just the first half of Activity D (settling the overall design direction) can start in parallel even before Activity C's results are finalized. This is a case of fast tracking: it incurs almost no additional cost, but carries the risk that the first half of Activity D may need rework depending on Activity C's results. Because this project had already frozen its requirements and the risk of Activity C's results changing drastically was low, the PM made the hybrid judgment: "first fast-track Activity D to shave off 1 day at essentially no cost, then crash Activity C for the remaining 2 days to reliably shorten the schedule." This keeps the additional cost to a minimum (2 days' worth, 400,000 yen) while applying fast tracking only where the rework risk is judged low—achieving a balanced compression strategy that does not lean entirely on either added cost or added risk.

2.3.3Section summary

• Crashing = adds cost to shorten duration, with small added risk. Fast tracking = parallelizes with small added cost but added risk
• Both apply only to activities on the critical path; applying them to non-critical activities is wasted effort
• Since repeated compression can switch the critical path, recalculate all paths after every compression