Adding up the schedules of a programme’s projects does not give the programme’s schedule. The resulting document can run to several thousand lines and changes every week; the dates that commit senior management get lost among the tasks. The master schedule is the view that is then missing: the overall schedule of a programme or a large project, reduced to the milestones that carry a decision or a commitment, to the dependencies between projects and to the margins that protect them. The expression comes from common usage, and its content is specified in each organisation.
We suggest building it in this order: scope and level of detail, decision milestones, dependencies, margins, then the update rule. It does not replace the roadmap, which links milestones to objectives and grades the certainty of each date, whereas the master schedule sets side by side, from one project to another, need dates, forecasts and margins; our article on the roadmap of an IT project deals with the former.
A master schedule is not the detailed schedule of any project
Each project keeps its detailed schedule: tasks, sequences, assignments. The master schedule is neither their sum nor their summary. It carries what projects owe one another and what the programme owes the body that funds it.
Between the two levels, dates flow in a precise direction. Commitment dates go down: a release stage, a cutover, a need date set by the programme are imposed on the projects’ schedules. Forecasts go up: at each cycle, each project manager announces the date on which they expect to reach their milestones. The master schedule shows both series, and the gap between them is the information it provides.
It is also distinct from the IT master plan, which sets the orientations of the information system and the projects to be undertaken for several years. Our page on programme management lists among its typical pieces of work an integrated programme plan (major milestones, dependencies, critical path), which we liken here to the master schedule.
Scope and level of detail are set before the first line is drawn
The scope says what the document covers: the programme’s projects, and what weighs on them from outside. A supplier whose delivery conditions a release stage, a regulatory deadline, a closing or a peak in activity that keeps users busy appear in it on the same basis as the projects. The horizon is that of the programme, up to the last release stage.
The level of detail says what deserves a line. We suggest a criterion: a milestone enters the master schedule if it meets one of these conditions:
- a decision is taken at it by the programme or by the body that funds it;
- another project, or a third party, is waiting for what is delivered at it;
- an external commitment depends on it, contractual or regulatory.
A milestone that only concerns its own project stays in that project’s schedule. As an assumption, a programme of six projects that kept between five and eight milestones per project would obtain thirty to forty-eight lines, a document that can be read in a meeting. The unit of time follows from this: the week or the month, depending on the length of the programme.
Decision milestones, linked by dependencies between projects
Decision milestones are placed first, because everything else is ordered in relation to them: ends of release stages, launch of a build, signature of a contract, green light for a cutover. Each carries the date, the level that decides, the documents expected and the date on which the file is to be submitted. The deciding level is not chosen when the schedule is drawn: it comes from the rules described in our article on programme governance.
Dependencies come next. Each links two milestones: the one at which a project makes an object available, the one at which another needs it. The master schedule therefore shows two dates per dependency, and the name of the person committed on each side. A need date that the supplying project has not confirmed is flagged as such.
Once the dependencies have been drawn, one chain stands out: the sequence of milestones that leads to the last release stage with the least margin. This is the programme’s critical path, a term borrowed from the scheduling technique of the same name. It crosses several projects, and the schedule of any one of them cannot show it.
When programme teams work in iterations, the content of each iteration does not appear in it. Their objectives are set in their own events, such as the PI planning of the SAFe framework; the master schedule only keeps their releases and what other projects expect from them.
Margins are shown and have an owner
The margin of a dependency is the interval between the date of availability and the need date. Suppose that a project commits to delivering an interface on 6 March, and that the project waiting for it needs it on 3 April to open its acceptance testing: the margin is 28 days, or four weeks. If the delivery forecast moves to 20 March, 14 days remain. The need date still holds; half of the margin has been consumed, and the master schedule shows it.
Three choices need to be written down. The first concerns location: margins spread across each dependency, a single reserve placed before the last release stage, or both. The second names the owner. A margin hidden in each project’s estimates remains invisible and can be consumed without a decision; shown in the master schedule, it belongs to the programme director, who grants it. The third sets the alert. One possible convention is to refer the matter to the programme director as soon as a forecast eats into the margin, and to the committee when it exceeds it; a variant adopts an intermediate threshold, for example half.
A consumed margin is recorded as a decision: who granted it, to which project, for what reason.
Who maintains the master schedule, and how often to revise it
We suggest the following split. The programme director is accountable for the content: release stage dates, margins, the arbitrations that move them. The programme office maintains the document. It collects the project managers’ forecasts, checks that they agree with their detailed schedules and flags discrepancies; this work forms part of the responsibilities presented in our article on the role and responsibilities of the PMO, and setting up this office falls under our page on setting up a PMO. Project managers provide their forecasts, and none of them changes a date on which another depends.
Revision follows the management cycle: the document is up to date before each programme committee. Two series of dates remain distinct in it. The baseline, approved by the committee, only moves by a decision of that committee. Forecasts move at each cycle. Correcting the baseline so that it matches the forecasts erases the gap, and with it what the committee had to read.
The medium matters less than this discipline: a spreadsheet or a scheduling tool will do, provided that a single version is authoritative and that it bears its date.
A first version can start from two lists requested from each project manager: what they must deliver to others, with the date promised, and what they expect from them, with their need date. The lines that appear in only one of the two lists are the dependencies to be confirmed before anything is drawn.