“When will it be finished?” is one of the most common questions an NGO project manager hears, from donors, directors, partners and communities alike. The critical path method (CPM) gives you a reliable way to answer it. It shows the shortest possible duration of your project, the sequence of activities that determines that duration, and exactly how much slack every other activity has. Once you understand it, you will know where to focus your attention and which delays really matter.
This article explains the method in plain language and then works through a complete example: a fictional health NGO organising a training programme for community health workers.
What is the critical path method?
The critical path method was developed in the late 1950s for planning large industrial projects and remains a foundation of modern scheduling, featuring in standards such as PMI’s PMBOK Guide. The core idea is simple:
- A project is a network of activities, each with a duration and with dependencies on other activities.
- There are several paths through the network from start to finish.
- The longest path determines the shortest possible project duration. That is the critical path.
- Activities on the critical path have zero float: any delay to them delays the whole project.
- Activities off the critical path have some float (also called slack) and can slip a little without affecting the end date.
Four inputs you need before you start
- An activity list, ideally derived from your work breakdown structure or logframe activities. Each activity should be a distinct piece of work with a clear start and finish.
- Durations for each activity, in a consistent unit (days or weeks). Estimate working time, and agree whether you are counting calendar or working days.
- Dependencies: which activities must finish before each activity can start. The most common type is finish-to-start.
- A start date and a working calendar that reflects public holidays, rainy seasons, school terms or religious festivals that affect your context.
The worked example: a community health worker training programme
Imagine a small health NGO preparing a residential training course for community health workers. The team lists eight activities, with durations in weeks:
| ID | Activity | Duration (weeks) | Depends on |
|---|---|---|---|
| A | Finalise training curriculum | 3 | None |
| B | Recruit and contract trainers | 4 | None |
| C | Train the trainers on the curriculum | 2 | A, B |
| D | Print training materials | 2 | A |
| E | Book venue and arrange logistics | 3 | None |
| F | Select and invite participants | 2 | A |
| G | Deliver the training course | 1 | C, D, E, F |
| H | Evaluate and write the training report | 2 | G |
If you sketch this as a network diagram, with each activity as a box and arrows showing dependencies, you will see five paths from start to finish:
- A → C → G → H: 3 + 2 + 1 + 2 = 8 weeks
- A → D → G → H: 3 + 2 + 1 + 2 = 8 weeks
- A → F → G → H: 3 + 2 + 1 + 2 = 8 weeks
- B → C → G → H: 4 + 2 + 1 + 2 = 9 weeks
- E → G → H: 3 + 1 + 2 = 6 weeks
The longest path, B → C → G → H, takes nine weeks. That is the critical path and the minimum project duration. Listing paths works for small examples, but for anything larger you need the forward and backward pass, which gives you the same answer plus the float for every activity.
Step 1: The forward pass (earliest dates)
Work from left to right through the network, calculating each activity’s Early Start (ES) and Early Finish (EF):
- Activities with no predecessors start at week 0.
- EF = ES + duration.
- An activity’s ES is the latest EF among its predecessors, because it cannot start until all of them are finished.
So A runs from 0 to 3, B from 0 to 4 and E from 0 to 3. C depends on A and B, so it must wait for the later of the two: it starts at week 4 and finishes at 6. D and F depend only on A, so each starts at 3 and finishes at 5. G depends on C, D, E and F, whose finishes are 6, 5, 3 and 5; it therefore starts at 6 and finishes at 7. H runs from 7 to 9. The project’s earliest finish is week 9.
Step 2: The backward pass (latest dates)
Now work from right to left, calculating each activity’s Late Finish (LF) and Late Start (LS): the latest it can finish and start without delaying the project.
- The final activity’s LF equals the project end date (week 9).
- LS = LF − duration.
- An activity’s LF is the earliest LS among its successors.
H must finish by 9, so it must start by 7. G must therefore finish by 7 and start by 6. C, D, E and F all feed G, so each must finish by week 6: C must start by 4, D by 4, E by 3 and F by 4. A feeds C, D and F, whose late starts are all 4, so A must finish by 4 and start by 1. B feeds only C, so it must finish by 4 and start by 0.
Step 3: Calculate float and identify the critical path
Total float = LS − ES (or LF − EF). Activities with zero total float are on the critical path.
| ID | Duration | ES | EF | LS | LF | Total float | Free float | Critical? |
|---|---|---|---|---|---|---|---|---|
| A | 3 | 0 | 3 | 1 | 4 | 1 | 0 | No |
| B | 4 | 0 | 4 | 0 | 4 | 0 | 0 | Yes |
| C | 2 | 4 | 6 | 4 | 6 | 0 | 0 | Yes |
| D | 2 | 3 | 5 | 4 | 6 | 1 | 1 | No |
| E | 3 | 0 | 3 | 3 | 6 | 3 | 3 | No |
| F | 2 | 3 | 5 | 4 | 6 | 1 | 1 | No |
| G | 1 | 6 | 7 | 6 | 7 | 0 | 0 | Yes |
| H | 2 | 7 | 9 | 7 | 9 | 0 | 0 | Yes |
The critical path is B → C → G → H, exactly as the path listing suggested. Now the table tells us much more:
- Trainer recruitment (B) is the activity to watch. If it takes six weeks instead of four, the training slips by two weeks, along with everything after it.
- The venue (E) has three weeks of float. The logistics officer can start a little later or work on other priorities first, without endangering the deadline.
- The curriculum (A) has one week of total float but no free float. If it slips by a week, D and F start a week later and lose their own float. A one-week delay to A is absorbed; a two-week delay pushes the project to ten weeks. Total float is shared along a path, so do not let two people each “use” the same week.
Tip: free float is how far an activity can slip without delaying the earliest start of any successor. It is calculated as the earliest ES among the activity’s successors minus its own EF. It is the float a team member can safely use without affecting anyone else’s work.
What to do when the schedule is too long
Suppose the donor needs the training delivered in eight weeks rather than nine. Two classic techniques help, and both should be applied only to critical activities, because shortening anything else does not change the end date.
Crashing
Crashing means adding resources to shorten a critical activity. In our example, the NGO might pay a recruitment agency to shortlist trainers faster, cutting B from four weeks to three. Compare the cost of each option per week saved and choose the cheapest. Watch out: once B is three weeks, the paths through A (eight weeks) are now equal to the path through B, so you have several critical paths and less room for error.
Fast-tracking
Fast-tracking means overlapping activities that were planned in sequence. The team could start training trainers on finalised modules while the last parts of the curriculum are still being written. This saves time without extra cost but increases risk, because later changes may mean rework.
From network to Gantt chart
The network is the engine; the Gantt chart is the dashboard most people prefer to read. Once you have ES, EF and float, draw each activity as a bar starting at its ES, highlight critical activities in a distinct way and show float as a thinner extension. Share the Gantt chart with your team and donor, but keep the network logic underneath, so that when an activity moves, the knock-on effects are recalculated correctly.
Handling uncertain durations
Durations in NGO projects are often uncertain: approvals take longer than expected, deliveries are held up, rains wash out roads. A useful technique from PERT (Program Evaluation and Review Technique) is the three-point estimate. For each activity, estimate an optimistic (O), most likely (M) and pessimistic (P) duration and calculate an expected duration as (O + 4M + P) ÷ 6.
If trainer recruitment could take three weeks at best, four weeks most likely and eight weeks at worst, the expected duration is (3 + 16 + 8) ÷ 6 = 4.5 weeks, half a week more than the single-point estimate. That simple calculation can prompt a valuable conversation about whether the deadline is realistic and whether a schedule buffer is needed.
Resources and the critical path
Classic CPM assumes resources are unlimited. In reality, the same programme officer may be responsible for both the curriculum and participant selection. If one person cannot do two things at once, the schedule must be adjusted through resource levelling, which may lengthen the project or change which activities are critical. Always check your critical path against who is actually doing the work, and include partner and government approval times as activities in their own right.
Common mistakes
- Missing dependencies. Activities with no successors (other than the final one) often signal a forgotten link. Every activity should lead somewhere.
- Activities that are too large. “Implement project” is not schedulable. Break work down until each activity can be estimated and has a clear owner.
- Treating float as spare time. Float is a buffer for the unexpected, shared along a path. Spending it early leaves you exposed.
- Forgetting external lead times. Donor approvals, visa processing, customs clearance and procurement thresholds are frequently the real critical path.
- Setting and forgetting. The critical path changes as the project progresses. Recalculate whenever an activity finishes late or early.
- Ignoring calendars. A four-week activity that spans a national holiday period or the rainy season may take considerably longer.
CPM checklist
- Derive activities from the work breakdown structure or logframe.
- Estimate durations with the people who will do the work; use three-point estimates where uncertainty is high.
- Record all dependencies, including external approvals.
- Run the forward pass to find early dates and the project duration.
- Run the backward pass to find late dates.
- Calculate total and free float; mark the critical path.
- Check resource conflicts and level if necessary.
- Present the schedule as a Gantt chart, and monitor critical activities most closely.
- Update the network whenever progress deviates from plan.
Key takeaways
- The critical path is the longest path through the network and sets the minimum project duration.
- The forward pass gives early dates; the backward pass gives late dates; the difference is float.
- Focus monitoring and risk management on critical activities, because any delay there delays the project.
- To shorten a project, crash or fast-track critical activities only, and watch for new critical paths.
- Recalculate the critical path regularly, and check it against real resource availability.
Critical Path Analysis, duration and budget estimating and schedule tracking are core topics in our one-day Advanced Project Management Tools & Techniques workshop, where you will work through network exercises based on real NGO projects. If you are new to project planning, start with Project Management Fundamentals, browse the full project management programme, or contact us about coaching on your own project schedule.
