Critical Path · CPM
Critical Path Method
What is the critical path?
- The longest path through zero-float tasks
- Only shortening this path shortens the schedule
- Crashing costs money; fast-tracking risks rework
Once the critical path is drawn on a single page, the language of your meetings changes. Instead of "everything is urgent," you hear "that task has four days of float, so even if it isn’t done this week, the finish date stays the same." The real payoff of this technique is that schedule debates move from a contest of loud voices to a calculation of paths. Owners also learn whether their task is directly tied to the finish date, which separates the stretches that call for all-nighters from the ones where you can keep your pace.
Float comes in more than one kind. Total float is how far a task can slip without pushing the project finish date; free float is how far it can slip without delaying the start of the very next task. Even with 5 days of total float, a free float of 0 means a one-day delay leaves the downstream team waiting. There is not always just one critical path, either — when parallel paths are equal in length, two or three arise at once, and near-critical paths with only 2–3 days of float effectively have to be managed alongside them. Note also that CPM assumes unlimited people and equipment; critical chain is the approach that handles resource constraints and pools the safety buffers.
Compression has a price. Crashing buys time with people, equipment, and expedited shipping, so you need to calculate what each day saved costs and start with the cheapest tasks. Fast tracking pays in rework risk instead of money: overlap follow-on work before the design is final, and a single drawing change wipes out the entire overlapped period. Applied outside the critical path, both techniques have exactly zero effect.
The most common failure in the field is framing the path and hanging it on the wall. Teams keep looking at the path drawn at kickoff three months later, and when a task with 5 days of float slips 8 days, they never notice a new critical path has formed and pour people into the wrong place. Paper paths also drift from reality when approval waits, material lead times, the Lunar New Year holiday, and vacations are not entered as tasks. Misunderstandings about crashing run deep too — doubling the headcount does not halve the duration, and people added late can even push the schedule back because of training and handover.
John KimAuthor of B2B Marketing Design and CEO of Nanum Consulting
Naha JinStartup leadership specialist with a social welfare backgroundIt was developed in the late 1950s by DuPont and Remington Rand to optimize equipment maintenance schedules, and together with the U.S. Navy’s PERT from around the same time, it became a foundation of modern project management.
Here is a hypothetical case. Company A, a home appliance parts maker (310 employees, KRW 88 billion (≈ US$65,838,700) in annual revenue), had planned a new assembly line expansion on a 12-week schedule. Then its customer asked for mass-production approval sooner, and the order came down: "Finish within 9 weeks." In every meeting, each team kept insisting that "our process is the most urgent."
- When the production engineering manager linked the dependencies among 47 tasks, 11 of the 12 weeks turned out to hang on a single chain: equipment order → receipt → installation → commissioning → quality verification.
- Sales had been pushing to finish the packaging line early. A Gantt chart showed that task had 9 days of total float, so pulling it forward would not change the finish date. The purchasing manager admitted, "So the problem was our order lead time."
- By crashing the schedule, we cut equipment delivery from 6 weeks to 4 using air freight and early purchase of spare parts. The extra KRW 28 million (≈ US$20,900) was backed by a calculation of KRW 2 million (≈ US$1,500) per day saved, and the plant manager approved it.
- With fast tracking, installation and electrical wiring overlapped by 2 weeks, and to prevent rework, a joint drawing review with the equipment vendor was set at twice a week.
- The PM pulled the 6 near-critical tasks with 3 days of float or less into a separate table and reviewed it every Monday, moving people the same week whenever the path changed.
- As a result, the schedule shrank from 12 weeks to 9.5 weeks at an added cost of only KRW 34 (≈ US$1) million, and not a single rework case occurred during commissioning.