What Is a Gantt Chart? A Worked Guide to Bars, Dependencies, and Schedule Delays

September 16, 2026Shusaku Yosa

A Gantt chart lists tasks against a timeline and uses horizontal bars to show their start-to-finish durations. It makes schedule overlap easier to see, but bars alone do not establish completion or workload. This worked guide explains how to read the legend, calendar, dependencies, baseline, and forecast together.

A Gantt chart places task durations on a timeline

A dated task list can manage a schedule. A Gantt chart adds a visual position and length for each period, making sequence and overlap easier to compare. The timeline may use days, weeks, or months, depending on the decisions the reader needs to make.

A long bar represents a long duration, not continuous work by one person. It may include waiting. Duration is time on the schedule; effort is the amount of human work. A task spanning three working days might require only two hours of a person’s effort.

Before reading individual bars, establish whether they show an agreed baseline or the latest forecast. Decide whether the purpose is communicating a deadline or coordinating handoffs. That determines how much detail belongs in the view.

Read bars, milestones, and dependency legends

Microsoft’s official guidance illustrates tasks, hierarchy, dependencies, and Gantt views. These provide useful reading concepts; the example here does not depend on a particular application. See the task and dependency guidance.

Element	What it conveys	What to verify
Bar	Start-to-finish duration	Working or calendar days; waiting included?
Milestone: ◆	A point such as a release decision	Completion rule and decision owner
Dependency: →	Relationship between tasks	Which start or finish is the condition?
Status line	The date used to assess information	Is the information current at that date?
Shading inside a bar	A defined percentage	Deliverables, effort, or duration?

Finish-to-start, or FS, means a successor starts after its predecessor finishes. Other relationship types exist. This article’s example uses FS throughout, with each successor starting on the next working day. Do not assume every arrow has that meaning without checking its legend.

Separate baseline dates, actuals, forecasts, and the status date

Keep the agreed baseline, actual dates, and remaining forecast distinct. Replacing the baseline with every new estimate hides what changed. Add a status date so comparisons use information from the same point in time.

An unfinished bar to the left of the status line is a prompt to investigate. If the record is stale, confirm it first. Fifty-percent shading could describe effort recorded or deliverables accepted; elapsed time is not automatically completed output.

When a task slips, check its forecast finish, successor conditions, and available margin. Shortening a bar by omitting required reviews does not make a schedule feasible. Record the reason for a proposed change and who may approve it.

Trace a delay to the deadline in a worked example

This fictional conceptual example uses consecutive working days D1–D8, excluding nonworking days. Tasks finish at the end of their shown day, and successors start on the next working day. Dependencies are A → B → C → D. The release decision occurs when D finishes; the external deadline is the end of D7. It is not a promise about external advertising-review times.

Task	D1	D2	D3	D4	D5	D6	D7 deadline	D8
A Requirements: baseline	■	■	—	—	—	—	—	—
B Production: baseline	—	—	■	■	—	—	—	—
C Internal review: baseline	—	—	—	—	■	—	—	—
D Release preparation: baseline	—	—	—	—	—	■◆	Margin	—
B Production: +1-day forecast	—	—	□	□	□	—	—	—
C Internal review: forecast	—	—	—	—	—	□	—	—
D Release preparation: forecast	—	—	—	—	—	—	□◆	—

Legend: ■ denotes baseline duration, □ the forecast after production extends by one working day, ◆ the release decision, and — no work for that row. Symbols distinguish scenarios, not completion percentages. The same A → B → C → D dependencies apply to the forecast.

Scenario	B finishes	C occurs	D and release decision	Effect on end-of-D7 deadline
Baseline	D4	D5	D6	One working day of margin
B extends 1 working day	D5	D6	D7	Within deadline; margin consumed
B extends 2 working days	D6	D7	D8	One working day late

Here, one day of delay does not immediately miss the deadline; two days do. That distinction supports a discussion about scope or timing. Recalculate if another constraint emerges, such as the reviewer being unavailable on D6.

Check effort before treating overlapping bars as overload

Two overlapping bars can fit one person’s capacity if each requires little effort. One bar can also exceed capacity. Use overlap to trigger a check of daily assignments, rather than as proof of overload.

Illustrative D4 assignment	Production	Other review	Total	Available	Interpretation
Owner A: initial	2 hours	3 hours	5 hours	6 hours	Within time capacity; also check skills and fixed times
Owner A: revised	2 hours	5 hours	7 hours	6 hours	1 hour short; change scope, timing, or ownership

Calendars, assignments, and slack can affect the schedule. Microsoft describes them separately. Check availability, other projects, and the people required for approval before accepting a revised date. Read about scheduling, calendars, and slack.

Use Gantt charts alongside a WBS, board, or task list

A work breakdown structure helps define scope; a Gantt chart places work in time; a board highlights state and flow. They can be combined rather than treated as mutually exclusive methods. Work can be broken down, scheduled, and then monitored through a state view.

Gantt charts are useful for fixed releases and handoffs. In frequently reprioritized operations, use a state view as the main working record and schedule only important milestones. Exploratory work can still have a timebox for reviewing a hypothesis, without pretending the conclusion is guaranteed.

Check the legend and working calendar.
Compare deadlines, milestones, baseline dates, and forecasts.
Trace dependencies and remaining margin.
Verify effort, reviewers, and constraints.
Record the required decision and next update.

Begin with the main stages, then add information when it supports a decision. A current, interpretable schedule is more useful than an elaborate chart whose assumptions no longer hold.