Reading a construction schedule

The document you are usually given

A residential construction schedule typically arrives as a row of bars with month labels: foundation, framing, roof, mechanical, drywall, finishes. It is reassuring and it is nearly information-free, because it shows duration without showing dependency.

The question a schedule exists to answer is not how long each activity takes. It is what cannot start until something else finishes, and therefore what happens to the end date when one activity slips. A bar chart without dependencies cannot answer that, which is why the same document is equally consistent with a job that finishes on time and one that finishes four months late.

The dependency spine

Under the bars there is a chain, and on a detached house it is remarkably consistent.

Site work and the foundation come first, and everything waits on them because nothing can be built until there is something to build on. Framing follows, and it produces the thing every subsequent trade attaches to. Once the structure is up and the roof is on, the building is weathertight, which is the first point at which interior work becomes possible.

Then comes the part that governs the middle of the job: the rough-in trades work simultaneously inside open walls. Plumbing, electrical and mechanical are all in the same cavities at the same time, and they interfere with each other constantly. Their sequence among themselves is negotiated rather than fixed, though the trade with the least routing flexibility usually goes first, which in most houses is the plumbing drainage.

Rough inspections gate the next step. Insulation follows, then the air barrier, then drywall, and drywall is the point of no return: it closes the walls and converts anything left undone into a repair.

After drywall the finishing trades run in a long tail: interior finishes, cabinetry, flooring, painting, fixtures, appliances, and the mechanical, electrical and plumbing trims. That tail is where schedules go quietly wrong, because it contains dozens of short activities with many dependencies and very little slack.

Where the real risk sits

Three places, and none of them is where owners look.

The first is the front, before anything is visible. Permitting, utility connections and site access are the activities most likely to slip, they are outside the builder's control, and a delay there moves everything downstream by the same amount. A schedule that shows permitting as a fixed two-week bar is describing a hope.

The second is inspection gates. Each rough inspection is a hard dependency: the trades behind it cannot proceed until it passes. A failed inspection costs the re-inspection interval plus the rescheduling of every trade that was waiting, and rescheduling a subcontractor who has moved to another job is not a phone call, it is a queue position.

The third is the finishing tail, for a reason that is about people rather than sequence. Finishing trades are short-duration, and a subcontractor whose two-day job slipped by a week will not sit idle waiting: they take other work and return when they can. That is a rational response to an unreliable schedule, and it is why a job that loses two weeks in the middle can lose six by the end.

What long-lead items do to the plan

Some components take months to arrive and cannot be substituted late without a redesign: windows, exterior doors, mechanical equipment, cabinetry, some tile and stone, and any custom item.

The distinguishing property of a long-lead item is that its order date, rather than its installation date, is the schedule constraint. Windows that install in month five may have to be ordered in month one, which means the window schedule has to be finalised in month one, which means the elevations were fixed before that.

A schedule that shows only installation dates is therefore hiding the decisions that drive it. The version worth asking for shows order dates and the decision deadlines behind them, and the second of those is the list an owner is actually responsible for. Most owner-caused delay on residential projects is a selection made a few weeks after the date on which it was needed.

Weather, and the assumption nobody writes down

Every residential schedule contains an unstated weather assumption, and it is usually the assumption that this year will be normal.

The exposure is concentrated rather than spread evenly. Once a building is weathertight, weather mostly stops mattering. Before that, it governs: excavation in saturated ground, concrete placed in freezing conditions or extreme heat, framing in wind, and roofing in rain are each either impossible or ill-advised, and the trades know it even when the schedule does not.

Which gives a specific question worth asking about any schedule: what does it assume about the season in which the building becomes weathertight, and what happens if that date moves by six weeks. In a cold climate, a job that expects to be closed in by late autumn and is not has bought itself a winter, and a winter is not a delay of a few weeks, it is a delay until the ground thaws.

Builders who work in serious climates plan around this by protecting the front of the job rather than the end of it. The activities worth accelerating are the ones before the building is closed in, because those are the ones with a deadline set by something other than the contract.

Owner-built and the coordination job

The Census Bureau counts owner-built houses as those the owner supervises as general contractor, roughly 50,000 of the 1,005,000 completed in 2025. [1]

Everything in this piece is the work that category describes. The general contractor's job is not the trades, it is the dependency chain: knowing what each trade needs in place before it arrives, having it there, and re-sequencing when something fails. It is a scheduling and procurement role, performed under time pressure, and the fee it commands is the price of somebody doing it every day.

Which suggests the fair test for anybody considering it. Not can I manage subcontractors, but can I keep a dependency chain of a hundred activities correct while employed elsewhere, and can I absorb the cost when I cannot.

What to ask for instead of a bar chart

Four things, and a builder running a real schedule has all of them.

The dependency list: which activities cannot start until which others finish. The inspection points and who calls them. The long-lead items with their order dates and the decision deadlines that precede them. And the activities that are not on the critical path, because those are where slack exists and where a delay is genuinely absorbable.

That last one is the most useful and the least often supplied. Knowing which delays matter is what separates an owner who can make a sensible decision under pressure from one who treats every slipped day as equally alarming. The rest of the sequence, from the permit onward, is set out on the front sheet under planning.

It is worth asking for all four before signing rather than after, for a reason that has nothing to do with distrust. A builder who can produce them is running the job from a plan, and a builder who cannot is running it from memory and from whichever subcontractor called this morning. Both approaches finish houses. Only one of them can tell you in March what is likely to happen in September.