On a typical construction job, roughly a third of the work week — about 14+ hours — goes to work that isn’t building anything: hunting for the current drawing, resolving a conflict between what the field has and what the office sent, redoing something that got done wrong because two people had different information. That’s per PlanGrid and FMI’s “Construction Disconnected” study, a 2018 benchmark that’s aged into a floor, not a ceiling, given how much project volume and subcontractor coordination has grown since. If your crews are burning that time, the fix isn’t a new system everyone has to learn — it’s making sure the information that already exists is in one place instead of scattered across texts, calls, and whoever remembers.
Why this leak hides inside “busy”
A crew that’s hunting for information looks identical to a crew that’s working. Nobody clocks “spent 40 minutes figuring out which version of the plan is current” as its own line item — it just gets absorbed into the day, and the day looks full. That’s exactly why owners underestimate this leak: there’s no invoice for it, no missed call, no obviously dead deal. Just a job that took longer than it should have, for reasons nobody can point to precisely.
The mechanism is simple. Information about a job — the current drawing set, a client-approved change, a sub’s schedule commitment, an inspection note — gets created once, usually in the field or on a call, and then has to travel to everyone who needs it. When that travel depends on someone remembering to forward a text, relay a conversation, or notice a printed sheet is out of date, some of it doesn’t arrive. The person working from stale information does the wrong thing, competently. That’s rework, and per the same PlanGrid/FMI research, roughly 48% of all rework traces back to exactly this — bad project data and miscommunication, not bad workmanship.
The math on a mid-size crew
Run this against your own crew size and blended labor rate. Here’s the shape of it:
| Input | Example |
|---|---|
| Field + office staff affected | 8 |
| Non-optimal hours per person, per week | 14 |
| Blended labor rate (loaded) | $45/hour |
| Weekly cost of non-optimal work | ~$5,040 |
| Annual cost | ~$262,000 |
Even if only a third of that time is genuinely recoverable — some hunting and rework is simply the nature of physical, multi-party work — that’s still upwards of $85,000 a year in labor going to a problem that has a much cheaper fix than it looks like. And that’s before counting the schedule slip: a job that takes 10% longer because of rework delays the next job from starting, which compounds across a year of bookings.
The fix: one place information lives, not a new system to run
The instinct when a company feels this is “we need better project management software.” Sometimes true — but the software isn’t what’s missing in most shops; it’s that whatever system exists gets bypassed the moment someone’s in a hurry, which is most of the time in the field.
- Information gets captured once, where the work happens. A field update, a client decision, a schedule change — logged from the truck or the site, not reconstructed later from memory in the office.
- Everyone pulls from the same current version. Not “check your texts, check email, check if the plan set in the truck is the latest one.” One current version, one place, and it’s obvious when something changed.
- A person still resolves the actual conflicts. This is where AI earns a place, and a narrow one: surfacing that a field note contradicts the office plan, or that a decision was made twice by two different people — flagging the mismatch before it becomes rework, not deciding which version is right. A PM makes that call. The tool just makes sure the PM sees the conflict before the crew builds the wrong thing.
The gate holds here the same as anywhere else: nothing about client-facing scope, price, or schedule gets decided by software. The system’s whole job is making sure the humans making those decisions are looking at the same facts.
When new software actually is the answer
If your team is already disciplined about logging and checking a shared source of truth, and the 14 hours are still disappearing, the problem probably isn’t workflow — it’s tooling that can’t keep up with your job complexity or sub count, and a real PM platform is worth the license and the training time. Similarly, if your rework is concentrated in one trade or one recurring design conflict rather than spread across general miscommunication, that’s a scope-and-spec problem a workflow fix won’t touch — you need better documents going out, not better tracking of the bad ones. The fix above pays off when the information exists somewhere and simply isn’t reaching the people who need it; it doesn’t manufacture information that was never captured in the first place.
Find out where your week is actually going
The 14 hours are the time leak in a contracting business — the other two are the bids that go quiet after they’re sent (why slow bid follow-up loses jobs you already priced) and the change-order work that gets done but never invoiced (the change-order revenue you never invoice). The 3-minute scorecard puts a rough number on all three for your business — deals, time, and cash — and tells you which one to fix first. See how this shows up across the rest of a GC’s operation on the use cases page. Free, no call, no pitch.