A paper timesheet costs you nothing to buy and about $4,000 to trust. That's a rough figure for a mid-size crew — the gap between what a job actually cost in labor and what you thought it cost, spread across a season of guessing. Paper doesn't lie, but it doesn't tell you anything useful until the payroll clerk keys it in two weeks after the job closed.

So the question isn't "paper vs. app." It's: what does job costing actually need to be worth the friction? Let's be honest about all three options, including where the app people oversell.

What paper is genuinely good at

Paper works offline, never crashes, and every worker on earth already knows how to use it. On a site with no signal and a crew that hates their phones, a clipboard beats software every single day. Don't let anyone tell you otherwise.

The cost shows up later. Hours get rounded up. Lunch breaks get forgotten. Nobody writes down which cost code the work belonged to, so all eight hours land in one bucket and your labor-per-phase data is fiction. You find out you underbid framing three jobs after you stopped being able to fix it.

The problem with paper isn't accuracy today. It's that you can't ask it a question. You can't slice it by phase, by crew, by cost code. It's a record, not a tool.— GC running four crews

What a dedicated time app fixes — and what it doesn't

A purpose-built time app captures hours against cost codes in real time, geofences clock-ins so ghost hours drop, and hands payroll a clean export. If your labor is your biggest cost line — and for most contractors it is — that visibility pays for the subscription fast.

The honest downsides: adoption is a real fight. Crews resent being tracked, older hands fumble the app, and a dead battery means a missing entry. And a standalone time tracker is one more login, one more monthly fee, one more thing that doesn't talk to your estimating or your invoicing. You solved timesheets and created an integration problem.

The third path: build tracking into the tools you already run

The all-in-one pitch is that one platform does time, estimates, invoices, and scheduling so nothing needs to sync. That's real value — until you realize the time module is mediocre because it was bolted onto an accounting tool, and now you're locked into a bundle where three features are great and two are dead weight.

Best-of-breed goes the other way: the best time app, the best estimator, wired together. More power, more logins, more fragility when one vendor changes an API. Neither is wrong. It depends on whether you'd rather manage one so-so system or five sharp ones.

Our bias, stated plainly: most of your stack should be tailored to how *you* run — in-house or close to it, shaped around your codes and your crews, not a generic template. The one place we'd never build in-house is where money changes hands. Client deposits and progress payments belong with a neutral third party, not sitting in your operating account creating a liability you have to explain if the job goes sideways.

The verdict

Paper if you're a one-truck operation and your memory is your ledger. A time app the moment labor cost blindness starts eating your margin — which is sooner than you think. Build the rest of your stack around your actual workflow, and keep escrow as the single outside piece you don't try to reinvent.

If you're rethinking how the money side sits inside all of this, that's the piece we handle.