How Many Activities Should a Schedule Have?

There is no correct number of activities. A two-story retail shell and a hospital tower should not have the same schedule, and anyone who gives you a fixed count without asking about the job is guessing. What there is, though, is a right level of detail for a given project, audience, and update cycle. This post gives you a way to find it, a few tests to tell if you have overshot or undershot, and a worked comparison of the same scope at three levels of detail.

The tradeoff

More activities give you more control and more visibility. They also cost more to build, more to update every month, and more to explain. Past a certain point the schedule stops being a plan and becomes a data entry job, and the updates start getting skipped or faked.

Fewer activities are easy to maintain but hide problems. A single "MEP Rough-In, Level 2 — 45 days" bar cannot tell you that the plumber is on track and the electrician has not started.

The right level is the one where:

• Every activity is something you can actually status in an update.
• Every activity has one responsible party.
• Progress or delay on any activity is visible within one update period.
• The person updating the schedule can get through it in a reasonable time.

Levels of schedule detail

Schedulers often talk about schedules in levels, roughly like this. The exact definitions vary by organization, but the idea is consistent.

Most construction contracts that require a CPM schedule are asking, in practice, for something at Level 3, with some areas drilled to Level 4. Level 5 detail usually belongs in a three-week lookahead or weekly work plan, not the master schedule. Keeping it out of the CPM is what keeps the CPM maintainable. How to Build a Three-Week Lookahead from the CPM Schedule covers that handoff.

The same scope at three levels

Here is interior work on one floor of an office building, about 25,000 SF, shown three ways.

Level 2 (one activity):

Interiors, Level 2 ............................. 60 days

Level 3 (seven activities):

Layout and Top Track, Level 2 ................... 3 days
Metal Stud Framing, Level 2 ..................... 10 days
MEP In-Wall Rough-In, Level 2 ................... 12 days
In-Wall Inspection, Level 2 ..................... 2 days
Drywall Hang/Tape/Finish, Level 2 ............... 15 days
Ceiling Grid and Above-Ceiling Inspection, L2 ... 8 days
Paint, Flooring, and Trim, Level 2 .............. 15 days

Those durations add up to 65 days end to end. With framing and rough-in overlapped using start-to-start relationships, the floor still finishes in about 60, which is where the Level 2 number came from.

Level 4 (split by zone, roughly 20+ activities): each of the Level 3 activities split into East and West halves, with MEP rough-in further split into plumbing, electrical, and fire sprinkler, and the inspections tied to each zone.

Which is right? For an owner summary, Level 2. For the contractual CPM, Level 3 is usually enough on a straightforward floor. Level 4 makes sense where floors are big enough that two zones are genuinely worked in parallel by different crews, or where the floor is on or near the critical path and you need to see which zone is holding things up.

The general principle: add detail where it changes decisions. Critical and near-critical work deserves more detail. Work with lots of float can stay coarse. You do not need the same granularity everywhere.

Tests for too little detail

You probably need more activities if:

• Activities span multiple update periods with no visible progress pattern. A 60-day bar that sits at 50% for two updates tells you nothing.
• One activity has several responsible parties. "MEP Rough-In" covering three subs means none of them own it.
• Inspections and approvals are buried inside work activities. If the in-wall inspection is not its own activity, a failed inspection is invisible to the schedule.
• The critical path runs through long, vague activities. If the critical path is "Structure — 120 days" followed by "Interiors — 180 days," it is not telling you anything you could act on.
• You cannot write a meaningful predecessor relationship. If drywall can start when "some" of the rough-in is done, you probably need to split rough-in by area.

Tests for too much detail

You probably have too many activities if:

• Updates are not getting done, or are getting done badly. If the scheduler is making up percentages to get through it, the schedule is too big for the update process.
• Activities are shorter than the update interval and there are hundreds of them. One-day activities in a monthly-updated schedule are mostly noise. Each one will be either not started or complete at every update.
• Nobody outside the scheduler can read it. If the superintendent prints a filtered view and still gets 40 pages, the master schedule is doing the lookahead's job.
• Logic is being written to the hour. "Electrician pulls wire in rooms 201-208, then 209-216" is crew planning, not CPM.
• Activity names are distinguished only by a number. "Framing 1," "Framing 2," "Framing 3" with no location means someone split for the sake of splitting.

Duration as a guide

Activity duration is a decent proxy for detail. A common rule many schedulers use is that most construction activities should fall somewhere between a few days and about a month. The DCMA 14-Point Assessment flags long-duration activities as a quality concern for exactly this reason: they hide status.

Exceptions are normal and fine:

• Procurement and fabrication can run months. You cannot see steel being detailed in the shop anyway; track it with intermediate milestones like shop drawings approved and material released.
• Level-of-effort activities like general conditions or project management span the whole job by design. They should not drive logic.
• Milestones are zero duration. See Milestones That Actually Mean Something.

The maintenance budget

A practical way to set detail is to work backward from how much time you have for updates. If you update monthly and you have a day to do it, including statusing with the field, you can maintain far fewer activities than if you have a dedicated scheduler with a week.

Do the rough math honestly. If each in-progress activity takes a minute or two to status with the field and enter, and a few hundred activities are in progress or about to start in any given month, that is already several hours before you have looked at the results. Most of the schedule is either complete or not yet started at any point, so it is the active window that sets the workload, not the total count.

If the budget does not fit, cut detail on high-float work first.

Detail in pieces, not all at once

Rolling wave planning is a reasonable way to keep detail manageable: build the near-term work at Level 3 or 4, keep the far-term work at Level 2 or high Level 3, and add detail as each phase approaches. Interiors on the top floor of a building that has not been dug yet does not need zone-level splits.

Two cautions. First, if the contract requires a fully detailed baseline, rolling wave may not be allowed for the submittal, so check the spec. Second, when you add detail later, make sure the new activities replace the old ones cleanly in the logic, with no orphaned summary activities left behind. Open Ends and Dangling Logic covers the cleanup.

Key takeaways

• There is no right activity count. There is a right level of detail for the job, the audience, and the update budget.
• Add detail where it changes decisions: critical work, multi-trade areas, inspections, and handoffs.
• Keep crew-level and day-level tasks in lookaheads and pull plans, not the master CPM.
• Most construction activities should be statusable within one update period. Split anything that is not.
• If updates are being skipped or faked, the schedule is too detailed for the process. Cut high-float detail first.