Gantt chart vs PERT chart: which one do you need?

How do they compare at a glance?

The short version: the Gantt chart is the better working document and the better thing to show people, and the PERT chart is the better thinking tool when the logic is complicated.

What does each chart actually show?

A Gantt chart shows time. A PERT chart shows logic. Nearly every other difference between them follows from that one.

On a Gantt chart the horizontal axis is the calendar. Each task is a bar, its length is its duration, and the bars stack in a list you read from top to bottom. Dependencies can appear as arrows between bars; the edges they connect identify the relationship. You can tell at a glance what is happening next week, which tasks overlap, and when the whole thing ends. What it is less good at is showing the shape of the logic: on a long plan the arrows cross each other, and following one chain from start to finish takes some squinting.

A PERT network usually places activities in boxes or on arrows, without a scaled time axis. A two-day task and a three-week task may take up the same space. Dates and earliest or latest times can be written on the diagram, but you read those labels rather than measuring a bar. The layout emphasises branching, merging and the paths through the work.

A useful comparison is a timetable and a metro map. The timetable tells you when each train leaves. The metro map tells you which stations connect, and deliberately ignores how far apart they are. You need the map to plan a route and the timetable to catch the train.

One office move, drawn both ways

Here is the same eight-task project as a PERT network and as a Gantt chart. The network makes the critical path easier to trace, and the Gantt chart makes the dates obvious. Each can be built from the other, but each makes a different thing easy to see.

The project is a small company moving to a new office. Durations are in working days. Each task has three estimates, and the expected time in the last column comes from the PERT formula explained in the next section.

As a PERT network

Drawn as a network, the project starts at one point and fans out into three tasks that need nothing before them: planning the layout, ordering the internet line and packing. The layout then splits into two branches, furniture and painting, which merge again at installing the furniture. That merges with the internet line at the setup task, which merges with packing at move day. Adding up the expected durations along every route from start to finish gives four totals:

• A, B, F, G, H: 4 + 11 + 2 + 2 + 2 = 21 days
• A, E, F, G, H: 4 + 6 + 2 + 2 + 2 = 16 days
• C, G, H: 10 + 2 + 2 = 14 days
• D, H: 3 + 2 = 5 days

The longest route is 21 days, so the critical path is layout, furniture, installation, setup and move day. That is a useful surprise. It is tempting to worry about the internet line first, and here it has seven days of float. The furniture order is what sets the date. Painting has five days of float, and packing has sixteen.

As a Gantt chart

Put the same tasks and links on a calendar, starting on a Monday, and you get eight task bars and a milestone, In the new office, after move day. The day numbers here count working days only, which in Ganttile assumes Hide weekends is turned on. Layout runs days 1 to 4, furniture days 5 to 15, painting days 5 to 10, the internet line days 1 to 10, packing days 1 to 3, installation days 16 and 17, setup days 18 and 19, and move day days 20 and 21. With Critical path turned on, the same five tasks are highlighted as in the network.

The calendar view makes an awkward choice easier to notice: packing on days 1 to 3 means living out of boxes for weeks. The network already gives packing sixteen days of float; the bars make the case for moving it later more immediate. They also show three tasks starting on day 1, which needs a capacity check if one office manager owns them all.

What the network showed more clearly was the merge at installation. On the Gantt chart, two arrows run into that bar, and which one actually holds it back only becomes clear when you notice painting finishes five days before the furniture arrives. Turning on Critical path does that reading for you, which is the main reason to use it.

What is the three-point estimate for?

A three-point estimate combines an optimistic duration (O), a most likely duration (M) and a pessimistic duration (P). The usual PERT formula is (O + 4M + P) / 6: a weighted average that gives M four times the weight of either extreme. These operational research course notes explain the calculation and its assumptions.

Take the furniture order. The supplier's best case is 8 days, their usual is 10, and when something goes wrong it is 18. That gives (8 + 40 + 18) / 6 = 11 days. The most likely figure is 10, but the expected time is 11, because the bad case sits much further from 10 than the good case does. That lopsidedness is normal. Work rarely goes much faster than expected, and it can go a great deal slower.

Across this critical path, the most likely durations total 19 days and the expected durations total 21. The difference comes from the larger downside estimates in this example. The formula does not add a safety margin automatically, and neither figure is a guaranteed completion date.

The same notes estimate each activity's standard deviation as (P - O) / 6. Assuming independent durations and a fixed critical path, adding their variances gives a standard deviation of about 2.2 days for this example. A normal approximation then puts completion by 23 days at roughly 82 percent. Those assumptions are restrictive: shared risks and another branch becoming critical can change the odds. J. E. Beasley's operational research notes show simulation as a check on the approximation and warn that the calculated probabilities may be inaccurate.

For a small team, the three numbers are more useful as a conversation than as maths. Asking "what is the realistic worst case?" brings out the supplier who is always late or the approval that sits in someone's inbox, which a single number hides. When that conversation pays off, and when one estimate is enough, is covered in estimating task durations.

When should you draw a separate PERT chart?

Less often than you might expect, and that is fine. Draw one when the sequence is the thing you are unsure about. Skip it when you already know the order and just need dates.

A network sketch earns its keep in a few situations:

• The order is disputed: putting tasks on sticky notes with no dates and arguing about the arrows keeps a planning session on logic, before anyone debates dates.
• Many branches merge: a product launch where hardware, packaging, certification and marketing all have to land before one date is easier to reason about as a network than as forty bars.
• Uncertainty needs discussion: first-time work, research or outside approvals may justify three-point estimates. You can use those estimates without drawing a separate network.
• You have to explain why the date is the date: a one-page network with the critical path highlighted is a quick way to show a sponsor that the furniture, not the internet line, is the problem.

And you can usually skip it when:

• the project is mostly one sequence of tasks, one after another
• the plan is short enough that the arrows on the Gantt chart are still readable
• the people you are planning for want to know when, not why

After sketching the network, enter its tasks and dependency links in the schedule. In Ganttile, Auto schedule recalculates linked dates and Critical path highlights the controlling chain. There is no separate PERT view or three-estimate field: calculate an expected duration on paper or in a spreadsheet, then enter that duration.

Where do PERT and CPM fit in?

PERT grew out of the US Navy's Polaris programme in the late 1950s. The 1959 PERT paper treated development durations as uncertain. CPM, developed by Kelley and Walker for industrial projects, used network planning to examine duration and cost.

Both methods analyse paths through a network. PERT traditionally uses optimistic, most likely and pessimistic durations; CPM commonly starts with one duration per activity. Modern scheduling tools combine ideas from both. A dependency network is not, by itself, a probabilistic PERT analysis.

Questions people ask about PERT charts

Why do some PERT charts put the tasks on the arrows?

Because the originals did. The 1959 PERT and CPM papers drew each activity as an arrow between events, the points where work starts or finishes. Most modern drawings put each task in a box and use arrows only for the order, which is easier to read and is how the office move above is drawn.

Is a PERT chart the same as a network diagram?

In everyday use, yes. People use PERT chart, network diagram and CPM diagram for the same drawing. Strictly, PERT also implies three estimates per task, so a network with one duration per task is closer to CPM.

Do you need special software for PERT?

No. The formula works on paper or in a spreadsheet, and the network can be sketched on a whiteboard. What software adds is recalculation: change one duration and the dates and the critical path update, which is the part worth automating.

Try the two views on one project

Use the network to work out the sequence, then the timeline to check dates, overlaps and availability. Keep uncertainty visible whichever view you choose.