PERT vs CPM: project schedule methods compared

What is PERT?

PERT is a three-point estimating method developed for the U.S. Navy Polaris program in the 1950s to plan research and development work with no useful history. For every activity you estimate an optimistic (O), most likely (M), and pessimistic (P) duration, then calculate a weighted expected duration: te = (O + 4M + P) / 6. The spread between O and P also produces a per-activity variance, which is what makes PERT probabilistic. The output is not a single date but an expected date with a defensible range around it.

What is CPM?

CPM is a network analysis method developed at DuPont in 1957 for plant maintenance and shutdown scheduling. You model activities, durations, and dependencies, then run a forward pass to calculate early start and finish dates and a backward pass to calculate late dates. Activities with zero float form the critical path: the longest chain of dependent work, and the only sequence that directly sets the project finish date. When a critical activity slips by a week, the project slips by a week.

PERT vs CPM: the key differences

The core difference is what each method treats as uncertain. PERT treats durations as ranges and produces probabilities; CPM treats durations as fixed inputs and produces structure: the critical path plus the float held by every other activity. That is why practitioners describe PERT as an estimating technique and CPM as a scheduling network technique. The decision table below summarizes which model fits which situation.

When to use PERT

Use PERT when duration uncertainty is the dominant risk: first-of-a-kind work, new technology, unproven vendors, or any activity where honest estimates differ widely. PERT is also the better choice when a sponsor needs a date with a confidence level, because the variance math produces statements like "19 weeks expected, roughly 85% likely to finish within 21." What PERT does not tell you is which activities can slip without moving the end date. That is CPM’s job.

When to use CPM

Use CPM when durations are reliable and the scheduling risk is structural: sequencing, dependencies, and hand-offs across known scope. Construction, migrations, event planning, and recurring project types usually fit this profile. CPM’s forward and backward passes expose float, which is the day-to-day management surface: it shows which late activities need attention and which can absorb delay. If your team argues about who is blocked by whom rather than about how long tasks take, start with CPM.

PERT vs CPM vs Monte Carlo simulation

Monte Carlo simulation extends the PERT idea by running thousands of schedule scenarios instead of calculating one expected value, which matters when a network has many interacting uncertain activities. A practical rule of thumb: PERT for quick three-point estimating on a handful of risky activities, CPM for sequencing and float on the full network, Monte Carlo when stakeholders need statistically robust schedule risk analysis on complex programs. All three work on the same activity network, so they are stages of maturity rather than rivals.

Worked example: a five-activity product launch

One small network shows both methods at work. Five activities drive a product launch: A, requirements (optimistic 2, most likely 4, pessimistic 6 weeks); B, design (3, 5, 13); C, prototype (1, 2, 9); D, build (4, 6, 8); and E, test and launch (2, 3, 4). Dependencies: A precedes B and C; B and C both precede D; D precedes E. Note that B carries the widest range, which is typical for design work, while D is a tight, well-understood build.

Dependencies

A → B, C; B and C → D; D → E

Critical path

A → B → D → E = 19 weeks (variance ≈ 3.78, standard deviation ≈ 1.9)

Alternate path

A → C → D → E = 16 weeks (C holds 3 weeks of float)

Step 1: calculate PERT expected durations

Applying te = (O + 4M + P) / 6 to each activity gives A = (2 + 16 + 6) / 6 = 4.0 weeks, B = (3 + 20 + 13) / 6 = 6.0 weeks, C = (1 + 8 + 9) / 6 = 3.0 weeks, D = (4 + 24 + 8) / 6 = 6.0 weeks, and E = (2 + 12 + 4) / 6 = 3.0 weeks. Notice how B’s pessimistic 13 weeks pulls its expected value well above its most likely 5: PERT weights uncertainty honestly instead of letting a single guess drive the schedule.

Step 2: find the critical path with CPM

Sum both paths through the network using the expected durations. Path A → B → D → E = 4 + 6 + 6 + 3 = 19 weeks. Path A → C → D → E = 4 + 3 + 6 + 3 = 16 weeks. The design path is the critical path: it is the longest chain of work and it sets the launch date. The prototype path holds 19 − 16 = 3 weeks of float, so activity C can absorb up to three weeks of delay before it threatens the finish date.

Step 3: turn the critical path into a confidence statement

PERT’s variance math now applies along the critical path. Each activity’s variance is ((P − O) / 6)²: A = 0.44, B = 2.78, D = 0.44, E = 0.11. The path variance is about 3.78, so the standard deviation is its square root, about 1.9 weeks. Assuming durations are independent and roughly normal, you can state: 19 weeks is the expected launch, 21 weeks carries roughly 85% confidence, and a 95% commit date is about 22.2 weeks. That range, not a single date, is what a sponsor can plan against.

Using PERT and CPM together

The example shows the standard workflow. Estimate risky activities with three points and calculate expected durations (PERT). Feed those durations into the activity network and run the forward and backward passes (CPM). Read the critical path and float to focus management attention, then use the path variance to communicate a confidence range for the finish date. The methods answer different questions, and the answers compose: probabilities from PERT, structure from CPM.

How to choose: three questions

Ask three questions before picking a method. First, how wide are the honest estimate ranges? Wide ranges push you toward PERT. Second, is the dominant risk in durations or in dependencies and hand-offs? Duration risk favors PERT, structural risk favors CPM. Third, does the sponsor need a confidence level or an execution plan? Confidence levels come from PERT’s variance math; execution plans come from CPM’s float analysis. Most real projects need both, which is why the two techniques are usually taught as a pair.

Common mistakes when applying PERT and CPM

Three failure modes account for most bad experiences. Inflated ranges: when teams pad optimistic and pessimistic values, expected durations balloon and the confidence range becomes fiction. Ignoring near-critical paths: the 16-week prototype path in the example sits only three weeks from critical, and a modest re-estimate can flip it. Treating the expected duration as a promise: PERT’s output is a probability statement, and quoting "19 weeks" without its range quietly destroys the method’s credibility with the delivery team.

How Pocket PMO runs PERT and CPM

Pocket PMO implements both techniques on the same project data. The PERT analysis workspace captures three-point estimates per activity and computes expected durations and variance. The critical path method workspace builds the dependency network, runs the forward and backward passes, and highlights zero-float activities. Because both tools read the same schedule, you can move from estimating to network analysis without re-keying data into spreadsheets.

Run PERT vs CPM on your own schedule

Recreate this five-activity example, or import your real project, and see the two methods work together. Start with the PERT analysis feature to capture optimistic, most likely, and pessimistic estimates, then switch to the critical path method feature to watch the network, critical path, and float update as your estimates firm up. Getting started is free.

Choosing the right tool for the moment

Reach for PERT when durations themselves are the unknown: new work, new vendors, or new technology. Reach for CPM once durations are stable enough to trust and the question becomes sequence and float. Most real projects move between these two modes several times before delivery, so the methods are complementary rather than competing.

Common failure modes

The classic PERT failure is optimism bias dressed up as three-point data, where the pessimistic case is really just the likely case with padding. The classic CPM failure is treating the critical path as fixed; as estimates and dependencies change, the path can move, and stale path reporting hides exactly the schedule risk the analysis exists to expose.

How Pocket PMO supports both

Pocket PMO keeps three-point estimates on activities so PERT calculations stay current, and recomputes the critical path as dependencies and durations change. Because both live in the same workspace, the schedule you estimated and the schedule you report from never drift apart over the life of the project.

Related planning pages

• PERT analysis software
• Critical path method software
• Monte Carlo simulation software