The defining problem of sitework is not difficulty. It is breadth. Clearing, mass grading, storm, sanitary, water, subgrade, base, curb, flatwork, and landscape can all be live in the same month, on the same site, with crews moving between them. Each has its own pay items, its own production rate, and its own way of going wrong.
The day, as it actually runs
Three crews on site. One is finishing storm on the north end. One is proof-rolling subgrade in the parking field. One is chasing curb ahead of the concrete sub who shows up Thursday whether you are ready or not. The dozer operator spent the morning on mass grading and the afternoon shaping a pond. Two loads of base showed up and one got rejected. The GC's super wants to know if you will be ready for asphalt on the 14th.
All of that has to end up correctly allocated to a dozen different cost codes. In practice, it usually does not.
On sitework, the job total almost always looks fine. It is the item-level detail that is bleeding, and it hides inside the total until the last month.
What hurts most
- Cross-coded hours. A crew that touched four items in a day gets coded to one. Every one of those four items is now mis-costed, and the errors partly cancel, which is worse than if they did not.
- Small items nobody watches. Individually trivial pay items add up to real money and get no attention until they collectively overrun.
- Progress questions with no fast answer. The GC asks whether you will hit a date and the honest answer requires three phone calls.
- Compliance spread across trades. Compaction on utility trench, proof rolls on subgrade, density on base, each with a different inspector and a different record.
- The one bad item hides. A job running 4 percent over overall might be three items running 40 percent over and everything else fine, but you cannot see that from a job-level number.
How the platform fits a sitework operation
The core move is that hours and quantities carry item-level context at the moment they are captured. Time is recorded with GPS-tagged clock in and out against the project and the work being performed, so a crew that moves between activities can be reflected accurately instead of averaged. Installed quantities are logged against their pay items as work goes in. Compaction and pressure tests are captured in structured form with photos and signatures at the location. Material deliveries are logged with photos when they arrive, and equipment assignments are tracked against the fleet.
With that context intact, the office gets earned value by item rather than only by job, which is the resolution sitework actually needs. The Axsus Intelligence OS layer reads the same verified record and surfaces which items are diverging from plan, which is exactly the question a job total cannot answer.
What changes in the first month
- Problem items surface individually. Not in the last month, when the only remaining option is to eat it.
- Schedule answers get faster. When the GC asks about the 14th, the answer comes from a record rather than a round of calls.
- Crew moves get cheaper. Reallocating a crew mid-week stops corrupting the cost data, because the hours follow the work.
Where to start
Sitework contractors usually get the fastest return from time tracking, because cross-coded hours are the largest single source of bad item-level data. Pair it with quantity tracking and the earned value picture becomes trustworthy at the item level. Daily reports then assemble themselves from what was already captured.