A baseline is the approved schedule you agree to be judged against. Until you freeze it, every date is a draft and nobody is late. After you freeze it, changes go through control — or the baseline is a mood board you update when the news is bad.
What you are freezing
The schedule baseline is the agreed dates, logic, and milestones at a moment in time — usually the end of planning, once sequencing, critical path, and leveling have been done once for real. In a full performance measurement baseline it sits with scope and cost. For this course, treat the schedule baseline as the thing you will compare 'actual' to when someone asks if you are late.
A baseline is not a forecast. The live schedule will move as you update progress, remaining duration, and approved changes. The baseline stays put so variance means something. If you overwrite the baseline every Friday to match reality, your variance is always zero and your credibility is a rounding error.
Get the people who own the date to sign something short: this is the logic, these are the milestones, this is the assumption set. A baseline nobody accepted is a PM hobby. A baseline the sponsor initialled is a conversation you can have when week nine looks different from week one.
Freeze, then change-control — in that order
Updating progress is not changing the baseline. Mark actual starts, remaining work, and slipped successors. That is hygiene. Changing a committed milestone, inserting a workstream, or rewriting logic because a vendor 'needs more time' is a change. It needs impact on the end date, cost, and risk, and an approver who can live with the new promise.
Re-baseline only when an approved change makes the old baseline a misleading story — a formal descoping, a funded delay, a killed workstream. Do it in the open. Keep the old baseline as history if the tool allows; people will ask what you originally promised. A quiet re-baseline after a bad month is how organisations learn not to read your charts.
Thresholds help. You may slip a non-milestone activity by two days inside float without a committee. You may not move a go-live without the sponsor. Write the thresholds before the first panic. Inventing them during the panic produces either paralysis or a rubber stamp, depending on who is in the room.
Defending the baseline without becoming a clerk
Report against it. Start with milestone variance and critical-path slack, not a tour of every bar that moved. 'Integration is six days late; total float on the go-live path is now two days; decision needed on overtime versus scope' is a baseline earning its keep. A 40-page Gantt in the appendix is optional. The variance sentence is not.
When someone wants work that breaks the baseline, show the network, not your feelings. The new request has predecessors, a duration, and a resource. Park it on the path and watch the milestone move. People argue with PMs. They argue less with a successor that suddenly starts after the conference.
Do not defend a baseline that was a lie. If you froze dates you already knew were impossible, change control will feel like bureaucracy and the team will be right. Baseline the schedule you would take a bet on after leveling and a first risk pass. Optimism belongs in the benefits case, not in the dates you will be quoted in six months.
Course complete — what's next
Baseline the approved logic and dates after the network, path, and resources are honest. Update progress freely. Change milestones and logic through control. Re-baseline in the open when the story has formally changed. Report variance, not a chart you keep forgiving.
You now have a working schedule practice: sequence the work, find the critical path and float, level the people, compress only with a named trade, then freeze and control. Apply it to one live plan this week — even a small one. A network with real links beats a date in a slide.
Practice this
WBS / Project Schedule
Turn theory into a real deliverable. Generate a wbs / project schedule with the AI tool and adapt it to your own project.
Open the generator