Network Diagram: How to Map Project Activities

A network diagram gives every project activity a place in the sequence -- and shows exactly which tasks have to finish before the next ones can start. Without one, you're scheduling from a flat list with no visibility into the chain reactions that cause delays.

Project managers use network diagrams to find the critical path, compress timelines, and communicate scope to stakeholders who need more than a spreadsheet. The diagram doesn't replace a Gantt chart -- it works alongside one, each showing a different view of the same plan.

What Is a Network Diagram?

A project network diagram is a visual flow of all project activities and the dependency relationships between them. Nodes represent activities (or milestones), and arrows represent the sequence in which those activities must occur. The result is a directional graph from project start to project finish.

Network diagrams are the backbone of two major scheduling methods: the critical path method (CPM) and the Program Evaluation and Review Technique (PERT). Both depend on having a correctly sequenced network before any duration estimates or probability ranges can be applied.

Key Facts

• Only 47% of projects are completed on time, according to McKinsey -- making schedule control one of the most persistent challenges in project management.
• One in six IT projects experiences a schedule overrun of 70% or more, often because dependency chains were never fully mapped.
• Projects that actively use a defined scheduling method, such as CPM with a network diagram, consistently outperform those that rely on list-based tracking alone. (PMI, Pulse of the Profession)

AON vs AOA: Two Types of Network Diagrams

There are two formats for drawing project network diagrams. Knowing which one you're looking at -- and which one your tool generates -- matters because they read differently.

Activity-on-Node (AON) places each activity inside a node (box). Arrows between nodes show dependencies only; the arrow itself carries no meaning beyond "this must come before that." AON is the format used by modern project management software and by the critical path method as most practitioners apply it today.

Activity-on-Arrow (AOA) places activities on the arrows, not in the boxes. The nodes (circles) represent events -- the start or finish of an activity. AOA is the older format, closely associated with PERT charts, and occasionally still used in civil engineering contexts.

For most projects today, AON is the right choice. It's simpler to read, requires no dummy activities, and is what tools like MS Project, Asana, and Smartsheet generate automatically.

If you're building a PERT chart for a research-heavy project with uncertain durations, you may encounter AOA notation in older references -- but AON still works.

Network Diagram Symbols and Terms

Before building a network diagram, get comfortable with the core symbols and scheduling terms. These appear in every forward and backward pass calculation.

Float is the number that tells you how much breathing room you have. Activities with zero float are on the critical path -- any delay there becomes a project delay.

How to Create a Network Diagram

Building a network diagram is a five-step process. The quality of the diagram depends entirely on the quality of the inputs you gather before drawing anything.

Step 1: List all activities

Start from your work breakdown structure (WBS). Every work package that produces a deliverable becomes an activity in the network. Assign each activity a short label (A, B, C or a code like WBS 1.1.1). Don't skip activities because they seem obvious -- missing nodes create false dependency chains.

Step 2: Define dependencies

For each activity, ask three questions:

• What must finish before this activity can start?
• What can start at the same time?
• What can't start until this activity finishes?

Document the dependency type (FS, SS, FF, SF) for each relationship. Most real-world projects are 80-90% finish-to-start, but complex engineering or software work often has parallel start-to-start chains.

Step 3: Sequence the activities

Arrange activities left to right based on their dependencies. Activities with no predecessors go at the left (project start). Activities with no successors go at the right (project finish). Draw arrows to connect them. Every activity should have at least one incoming and one outgoing arrow, except the start and end nodes.

Verify there are no circular dependencies -- if Activity A depends on Activity C and C depends on A, you have a loop that makes scheduling impossible.

Step 4: Add duration estimates

Write each activity's estimated duration inside or beside its node. For uncertain durations, use PERT's three-point estimate: optimistic (O), most likely (M), and pessimistic (P). The PERT formula gives you a weighted average: (O + 4M + P) / 6.

For straightforward projects with reliable estimates, a single deterministic duration per activity is enough.

Step 5: Calculate the critical path

Run the forward pass to get early start and early finish for each activity. Then run the backward pass to get late start and late finish. Calculate float for each activity. The path where every float value equals zero is the critical path -- and its total duration is the project's minimum completion time.

See project planning for how to connect this output to your baseline schedule.

Network Diagram Example

Here's a simple software release project with six activities.

Forward pass (Early Start / Early Finish):

• A: ES=0, EF=3
• B: ES=3, EF=8
• C: ES=8, EF=16
• D: ES=8, EF=14
• E: ES=16 (driven by C), EF=20
• F: ES=20, EF=22

Project duration: 22 days.

Backward pass (Late Start / Late Finish, working from EF of F=22):

• F: LF=22, LS=20
• E: LF=20, LS=16
• C: LF=16, LS=8
• D: LF=20, LS=14 (D has 6 days of float: LS 14 - ES 8 = 6)
• B: LF=8, LS=3
• A: LF=3, LS=0

Float calculation:

Critical path: A --> B --> C --> E --> F (22 days). Activity D has 6 days of float, meaning the UI development can start up to 6 days late without delaying the release.

Network Diagram vs Gantt Chart

These two tools solve different problems and work best together.

Build the network diagram first. Once you're confident the sequence and durations are correct, transfer the output into a Gantt chart for day-to-day tracking and stakeholder reporting. The Gantt chart shows when; the network diagram shows why.

Common Mistakes

Skipping the dependency conversation. Many teams draw network diagrams after the schedule is already set, which makes the diagram a decoration rather than a planning tool. Use it at the dependency-definition stage, before durations are locked.

Treating every path as critical. When teams see a complex network, they sometimes manage every activity with the same urgency. Float exists for a reason -- let non-critical activities absorb minor delays while you focus attention on the true critical path.

Missing merge activities. An activity that has multiple predecessors (like Activity E in the example above) is a merge point. Managers often underestimate risk at merge points because they only watch the critical predecessor and ignore the others. A delay in any predecessor delays the merge.

Using the wrong dependency type. Defaulting everything to finish-to-start when the work actually overlaps means your schedule is longer than it needs to be. Review SS and FF dependencies actively; they're where schedule compression opportunities hide.

Forgetting to update the diagram. A network diagram built at project initiation and never touched again is worse than useless -- it gives false confidence. Update it when scope changes, when activities are added, or when a predecessor relationship turns out to be wrong.

Frequently Asked Questions

What is the difference between a network diagram and a PERT chart?

A PERT chart is a specific type of project network diagram that uses activity-on-arrow (AOA) notation and three-point duration estimates (optimistic, most likely, pessimistic) to handle schedule uncertainty. A network diagram is the broader term -- it includes both AON and AOA formats, and it can use single-point or three-point estimates. Every PERT chart is a network diagram, but not every network diagram is a PERT chart.

How many activities should a project network diagram have?

There's no fixed rule, but diagrams with more than 50-60 activities become hard to read on a single page. For large projects, break the master network into sub-networks by phase or work package, then connect the sub-networks at key milestones. The work breakdown structure is the natural way to group activities for this purpose.

Can a project have more than one critical path?

Yes. If two or more paths through the network have the same total duration (both equal to the project's minimum completion time), they're both critical. Multiple critical paths mean higher schedule risk -- a delay anywhere on either path delays the project.

What software generates network diagrams automatically?

MS Project, Primavera P6, Smartsheet, and ProjectLibre all generate network diagrams from task lists and dependency data. Lucidchart and Miro support manual drawing. Most project management platforms (Asana, Monday.com, Wrike) show dependency views that approximate a simplified AON network.

What happens to the critical path when a change order is added?

New activities shift float values across the network. A change order that adds duration to a previously non-critical path can make that path critical -- or create a new critical path entirely. Recalculate forward and backward passes after any scope change that adds or removes activities or changes dependency relationships.

Related reading

• Critical Path Method -- how to calculate CPM from a network diagram
• What Is a PERT Chart? -- probabilistic scheduling with three-point estimates
• What Is a Gantt Chart? -- the timeline view that follows your network diagram
• Work Breakdown Structure -- building the activity list your network depends on
• Project Planning -- connecting the network to a full project baseline

A network diagram is where a project schedule actually gets built. The Gantt chart and the status reports come after -- but the dependency logic that makes those documents reliable lives here, in the node-and-arrow map you draw before the first task starts.