What is a Gantt chart, and how do you make one that stays useful?

What a Gantt chart is

A Gantt chart has two halves. On the left is a list of tasks, often grouped under phases or a work breakdown structure. On the right is a timeline, with one bar per task. The bar's left end is the start date, its length is the duration, and its right end is the finish date.

The chart is named after Henry Gantt, who popularised it in the early twentieth century. It remains the most common way to show a schedule because anyone can read it without training: earlier is left, later is right, longer bars take longer.

A Gantt chart is a view of a schedule, not the schedule itself. The real schedule is the data underneath: tasks, durations, dependencies and calendars. When that data is right, the chart draws itself. When people draw bars by hand, the chart becomes a picture that goes stale the moment anything changes.

The parts of a useful Gantt chart

• Task bars: one per activity, with duration in working days.
• Summary bars: roll up a phase, from its earliest start to its latest finish.
• Milestones: zero-duration events, usually drawn as diamonds, such as "design approved".
• Dependency arrows: links between tasks. The usual one is finish-to-start (FS): B starts when A finishes. Start-to-start (SS), finish-to-finish (FF) and start-to-finish (SF) also exist, each with an optional lag.
• Critical path highlighting: the chain of tasks with zero total float, usually in red. Any delay on it delays the finish.
• Data-date line: a vertical line at the date progress was measured. Work left of it should be done; work right of it is forecast.
• Baseline bars: thin bars showing the original plan under the current bars, so slippage is visible at a glance.
• Progress: shading inside each bar for percent complete.

How to make a Gantt chart, step by step

1. List the deliverables. Break the work into a WBS, then into tasks small enough to estimate (a few days to a few weeks each).
2. Estimate durations in working days. For uncertain tasks, use a three-point estimate.
3. Link the tasks. For each task ask "what must finish before this can start?" Every task except the first should have a predecessor, and every task except the last should have a successor.
4. Set the calendar: working week, holidays, and the project start date.
5. Let the dates calculate. A forward pass gives early dates; a backward pass gives late dates and float. Avoid typing dates directly.
6. Add milestones for approvals, handovers and contract dates.
7. Review the critical path with the team. If it does not make sense, a link is usually missing.
8. Save a baseline once the plan is agreed.

Worked example: five tasks and a critical path

A small fit-out project has five tasks, all linked finish-to-start:

Task	Duration (days)	Predecessor	Early start	Early finish	Total float
A. Design	5	none	0	5	0
B. Procure equipment	10	A	5	15	0
C. Build partitions	8	A	5	13	2
D. Install equipment	4	B, C	15	19	0
E. Test and hand over	3	D	19	22	0

There are two paths. A, B, D, E takes 5 + 10 + 4 + 3 = 22 days. A, C, D, E takes 5 + 8 + 4 + 3 = 20 days. The longer one is the critical path, so the project needs 22 working days. Task C can slip by 22 minus 20 = 2 days without moving the finish: that is its total float.

On the Gantt chart, A, B, D and E are red, and C shows a two-day float tail. If procurement (B) slips by three days, the chart redraws: D and E move right by three days and the finish becomes day 25. Had you typed the dates by hand, you would have had to move three bars yourself and might have missed one. The critical path method guide walks through the forward and backward pass in detail.

How to keep a Gantt chart useful

• Update on a rhythm. Record actual starts, actual finishes and remaining durations at a fixed data date, weekly or every two weeks.
• Never move bars to match reality. Update progress and durations, and let the logic move the dates.
• Keep the level of detail steady. A few hundred well-linked tasks beat thousands of tiny ones that nobody updates.
• Use constraints sparingly. A fixed date stops the logic from working. Keep hard dates for real contract or regulatory deadlines.
• Compare with the baseline every period, so slippage is visible and explained.
• Show the right view to each reader. Sponsors need milestones and summary bars; the team needs the next three weeks of detail.

Common mistakes, and when not to use a Gantt chart

• Bars with no links. The chart looks fine until the first delay, then every date is wrong.
• Open ends. Tasks with no successor hide delays, because nothing downstream moves when they slip.
• Durations in calendar days on some tasks and working days on others. Pick working days and set the calendar.
• One giant chart for everyone. Filter and summarise instead.

A Gantt chart is not always the right tool. For short, repetitive or highly changeable work (a support queue, a content pipeline, a sprint) a Kanban board shows flow better than dates do. Many teams use both: a board for daily work and a Gantt chart for milestones and dependencies.

How to do this in Critova

In Critova you build the schedule in an editable activity grid, picking predecessors by name and choosing FS, SS, FF or SF links with lags. The critical path, total float and free float are calculated with a forward and backward pass using your working calendar and holidays, and the Gantt chart draws the result with the data-date line and your saved baselines. If you already have a plan in MS Project XML, Primavera XER or P6 XML, you can import it with a preview first. See the scheduling features.