The critical path method (CPM) is a scheduling technique that finds the longest chain of dependent tasks in a project, the chain with no room for delay. If any task on that chain slips by a day, the whole project finishes a day late. You list the tasks, their durations and dependencies, then run a forward and a backward pass to work out each task's float.
Key takeaways
- The critical path is the longest chain of tasks from start to finish with zero float; it sets the project's duration.
- The calculation has four parts: a task list, dependencies, a forward pass (earliest dates) and a backward pass (latest dates).
- Float shows how many working days a task can slip without moving the finish date; management attention belongs on the zero-float tasks.
- Panorama Consulting's 2026 ERP Report found that almost a quarter of organisations saw their project run over schedule.
- A critical path is never drawn once and left alone; it has to be recalculated after every slip, which is why a tool that reschedules dependencies automatically saves real effort.
Why does the critical path method matter, and what do delays cost?
Most projects are not derailed by one big mistake but by small slips that add up. Design sign-off arrives two days late, developers wait, the testing week gets squeezed and the delivery date moves. The real problem is that nobody can see, at the moment it happens, which delay actually threatens delivery.
This pattern is common in larger business projects too. Panorama Consulting's 2026 ERP Report found that more than a quarter of organisations went over budget and almost a quarter went over schedule, with a median project timeline of 9 months. The report puts organisational issues, rather than technical ones, at the top of the reasons for schedule overruns. We look at the same patterns in why ERP projects fail.
A delay costs more than a late handover. Idle teams, penalty clauses, deferred revenue and client confidence all go into the bill. That is why schedule risk deserves its own line when you estimate software project cost.
What are the key concepts in the critical path method?
A few terms need to be clear before you start. All of them are measured in working days, so weekends and holidays are excluded.
- Predecessor and successor: a predecessor is a task that must finish before another can start; the dependent task is its successor.
- Early start/finish (ES/EF): the earliest day a task can start and finish if its predecessors finish on time.
- Late start/finish (LS/LF): the latest day a task can start and finish without delaying the project.
- Float (or slack): the difference between LF and EF. If it is zero, the task is on the critical path.
- Milestone: a zero-duration event such as a sign-off or a launch.
Dependencies are not always "one finishes, the next starts". Some tasks start together, some finish together, and there may be a lag (a wait) or a lead (an overlap) between them. Getting this right at the outset decides whether the result can be trusted. Our guide to what a Gantt chart is shows how these links look on a timeline.
How do you calculate the critical path step by step?
The same sequence works on paper and in a project tool:
- List the tasks. Break the work into deliverable pieces, each small enough for one person or team to own.
- Estimate durations in working days. Agree estimates with the people doing the work, and write down a realistic figure rather than the best case.
- Define dependencies. For every task, ask: "What has to be finished before this can begin?"
- Run the forward pass. Starting from the first task, work out each task's ES and EF; a task with several predecessors starts after the latest of them finishes.
- Run the backward pass. Working back from the task that finishes last, find each task's LF and LS; where a task has several successors, take the smallest LS.
- Calculate float. For each task, LF − EF. Join up the tasks with zero float and you have your critical path.
A plan built before the requirements are settled will mislead you, however accurate the arithmetic. To pin down the scope your task list comes from, see our guides to software requirements analysis and writing technical specifications.
Critical path example: a website project
Take a simple website project with six tasks. Durations are in working days and the day count starts at zero.
| Task | Duration | Predecessor | ES | EF | LS | LF | Float |
|---|---|---|---|---|---|---|---|
| A. Requirements | 3 | – | 0 | 3 | 0 | 3 | 0 |
| B. Design | 5 | A | 3 | 8 | 3 | 8 | 0 |
| C. Content writing | 6 | A | 3 | 9 | 10 | 16 | 7 |
| D. Development | 8 | B | 8 | 16 | 8 | 16 | 0 |
| E. Content upload | 2 | C, D | 16 | 18 | 16 | 18 | 0 |
| F. Testing and launch | 2 | E | 18 | 20 | 18 | 20 | 0 |
The critical path is A → B → D → E → F, and the project takes 20 working days. Content writing (C) can slip by 7 working days without changing the launch date. If design (B) slips by a day, however, the launch slips with it.
The practical lesson: extra resource put into design and development shortens delivery, whereas extra resource on content does not. The critical path lets you say "protect this chain of tasks" instead of "everyone hurry up". When deciding which tool to run the calculation in, the criteria in our guide to choosing project management software will help.
What mistakes do teams make with the critical path method?
The most common mistake is drawing the plan once and pinning it to the wall. In reality tasks stretch, shrink and multiply, and the critical path can jump to a different chain after any change. A task with two days of float that slips by three is now on the critical path.
The second mistake is defining dependencies loosely. In projects where "everyone knows their own job", tasks look independent and the calculation wrongly produces a short duration. The third is ignoring resource clashes: if the same person is assigned to two critical tasks, they cannot run in parallel.
Doing the calculation by hand or in a spreadsheet is fine for small projects. As the task count grows, though, redoing the forward and backward passes after every change becomes impractical, as we explain in outgrowing spreadsheets. Manufacturers face the same problem at machine and work-order level, which is where production scheduling and APS comes in.
How does SmartPM make critical path tracking easier?
SmartPM is the web-based, Turkish-language project management software in Digital Bridge's Smart family. It does the calculation above for you and keeps it current as the plan changes. Tasks are managed as a list, a kanban board and a Gantt chart, all from the same data.
The Gantt chart supports four dependency types: finish-to-start, start-to-start, finish-to-finish and start-to-finish. Lag is entered in working days, a negative value pulls work forward, and the system rejects any link that would create a loop. Weekends are not counted as working days, and working days can be set per project for teams that work on Saturdays.
The critical path view switches on with a single click: critical tasks and their links turn red, and every other task shows its float. SmartPM uses the same logic as above, working back from the task that finishes last to find each task's latest possible finish and showing the difference as float.
Scenario: an agency opens a client website project from SmartPM's ready-made website template. The design task runs three days over because the client's sign-off is late. When the project manager stretches the bar on the Gantt chart, development, content upload and launch move forward with their durations intact, and the screen reports how many tasks have shifted. The content writing task, which does not depend on design, stays where it is.
Client approval runs from the same place. The team marks the design task as visible to the client and clicks "Request approval"; the client sees only the tasks opened to them and replies with Approve or Request changes. Before the weekly meeting, the AI-assisted status summary pulls together late tasks, the critical path and upcoming deadlines from the project's own data only; it is shown as a suggestion and is not saved. We cover other ways to catch delay signals early in AI as an early warning for project risk. You can explore every feature on the SmartPM product page.
How can Digital Bridge help?
Digital Bridge treats project management as a way of working, not just a software licence. When setting up SmartPM, we plan with you who works in which role (manager, member, viewer, client), each project's working days and which tasks are opened to clients. If you use other Smart products too, you work under Smart360 with a single account.
Where an off-the-shelf tool does not fit, we build planning screens around your own workflow through our custom web software service. For projects run with outside suppliers, our guide to supervising software contractors helps keep the schedule on track. For site-based work, see our guide to construction project management software.
Next step: make your critical path visible
Take a live project today and write it down as tasks, durations and dependencies; working out the critical path even once shows you where to focus. Rather than redoing the calculation by hand after every slip, you can rely on automatic rescheduling and the critical path view in SmartPM.
To buy SmartPM, fill in the SmartPM sales form and our team will plan the right set-up for your organisation with you. For questions, reach us through our contact page, and find more guides on our software development topic page. If you want to tie performance indicators to the plan, see our definition of a KPI.