Critical Path Method (CPM)

The critical path is the sequence of activities that, if any one of them is delayed, will push back the project end date. It is found by calculating two passes through the network: the forward pass (which gives the earliest possible start and finish for each activity) and the backward pass (which gives the latest allowable start and finish without extending the project). Activities where these two calculations produce the same dates have zero total float — they are critical. The chain of zero-float activities from start to finish is the critical path.

The mechanics work like this. In the forward pass, the planner walks through the network from the project start, calculating each activity's Early Start (ES) and Early Finish (EF). ES of the first activity is day 0; EF = ES + duration. For any successor activity, ES is the maximum EF of all its predecessors (because the successor cannot begin until every predecessor is done). When the walk reaches the project end, the EF of the final activity is the earliest possible project completion date. The backward pass then walks back from that completion date, calculating each activity's Late Finish (LF) and Late Start (LS) — the latest dates each activity can finish and start without pushing the project end out. For any predecessor, LF is the minimum LS of all its successors. Total Float for each activity = LS − ES (or equivalently LF − EF). Activities with zero total float lie on the critical path.

A short worked example makes it concrete. Imagine four activities: A (5 days, no predecessors), B (3 days, follows A), C (7 days, follows A), D (2 days, follows both B and C). Forward pass: A runs days 0–5; B runs days 5–8; C runs days 5–12; D cannot start until both B and C are done, so D runs days 12–14. Project completion = day 14. Backward pass: D finishes at day 14, starts day 12. C must finish by day 12, so LF=12, LS=5. B must also finish by day 12, but it only takes 3 days, so LF=12, LS=9. Total Float for A = 0, for C = 0, for D = 0 — these are critical. Total Float for B = 9 − 5 = 4 days. Critical path: A → C → D. B has 4 days of float; if B slips by less than that, the project still finishes on day 14.

CPM is the foundation of almost all schedule management. Once the critical path is known, the project manager knows where to focus attention, where to resource-load, and where delay is genuinely consequential. It also enables compression analysis: if the project end date needs to be brought forward, crashing (adding resources to critical activities to shorten their duration, usually at increased cost) or fast-tracking (overlapping activities that were planned sequentially, usually at increased risk) are the levers available. The choice between them depends on the cost-time tradeoff and the team's tolerance for rework. Understanding CPM is a baseline competency for any planner or scheduler — and a recurring weak spot on UK programmes that adopt agile delivery practices without preserving the underlying network logic.

The most important thing to understand about CPM is that the critical path is dynamic — it changes as the project progresses and as actual durations deviate from the plan. A path that starts with two weeks of float can become critical if work on it slips. Planners should monitor near-critical paths (those with small amounts of float, often 5–10 days on infrastructure programmes) just as carefully as the nominal critical path. Also be aware that in a risk-modelled schedule, the 'most critical' path in a probabilistic sense may differ from the deterministic critical path — which is why criticality index from a Monte Carlo run is a more useful indicator than deterministic float alone. A path with high criticality index (the percentage of Monte Carlo iterations in which that path drives completion) deserves attention even if it currently shows non-zero float on the deterministic schedule.

CPM has well-known limitations. The first is that activity durations are treated as point estimates, when in reality they are uncertain — which is why Schedule Risk Analysis layers three-point estimates and Monte Carlo simulation on top of the deterministic CPM network. The second is that CPM assumes resources are unlimited; in practice a schedule that respects logic but ignores resource availability is not executable, which is why resource-loaded scheduling and the Critical Chain Method (CCM) were developed as extensions. The third is the DCMA 14 issue that critical paths running through float — meaning the path the team thinks is critical is not actually the path that drives completion — are a recurring red flag on programmes that have been re-baselined repeatedly without testing the logic. A controls function that can confidently identify the live critical path, explain why it is critical, and articulate what would change it is doing CPM well. One that needs to refer the question back to the planning office is not.