Two symptoms, one disease. Symptom one: payday involves archaeology, someone reconstructing who shot what from calendars and chat threads, and the crew double-checking the result because last month it was wrong. Symptom two: the owner knows revenue but discovers margin annually, at tax time, as a surprise. Both trace back to a question almost no service business answers explicitly: at what moment is the money on a job actually earned?
Money moves through a job in stages: booked, scheduled, performed, delivered, paid. Crew pay and margin both need to attach to exactly one of those moments, chosen deliberately. Attach pay at booking and every cancellation strands money you already owe, or claw-backs you have to explain. Attach it at payment and your crew's income depends on your collections speed, which they will correctly resent. For work with a deliverable, the honest anchor is delivery: the job is done when the client has the work in hand, so that is when the crew's cut becomes real and when the job's economics are worth reading. Your business might reasonably pick a different moment. What it cannot do is leave the question unanswered, because then every part of the money layer answers it differently, and that disagreement is the archaeology.
The spreadsheet approach recomputes pay from scratch every period, which is why every period is an argument. The system approach writes an earning record at the chosen moment: this person, this job, this amount, earned now. Payday becomes reading a ledger instead of reconstructing history. Two details keep the ledger honest. Each earning gets exactly one identity, one record per person per deliverable, so a job that gets reprocessed or retried can never mint the same pay twice. And corrections are new entries, not edits, so the ledger explains itself a year later, which is exactly what a spreadsheet full of overwritten cells cannot do.
Per-job margin is revenue minus labor minus direct costs, which is arithmetic a napkin can do. The reason businesses do not have it is not the math, it is that the inputs leak. Labor cost gets recorded when a job is created, and then the job mutates: a reshoot gets added, a crew member gets swapped, a service gets upgraded, and the mutation path forgets to carry the cost along. We have hunted exactly these leaks in our own platform, and the lesson generalizes: every code path and every human process that changes what work happens on a job must also update what that work costs, or your margin report is fiction that gets more fictional as your operation gets busier.
Once earnings are records and costs stop leaking, the useful dashboard is small: margin per job as it lands, labor as a share of revenue over time, and unpaid earned amounts, what you owe the crew right now. That last number is the one spreadsheet operations never see, and it is the difference between payroll as a known liability and payday as a monthly surprise. None of this requires accounting software heroics. It requires answering the one question, and then refusing to let any part of the system answer it differently.
The anchor question still applies to margin: hours worked need to land on jobs at a defined moment for per-job economics to exist. Timesheets that never meet jobs are how hourly shops end up margin-blind.
Yes, and in real time. Transparency converts payday from an argument into a formality, and crew who can see earnings accrue per job tend to care more about the jobs.
The record-keeping is lighter than the archaeology it replaces. Small operations feel wrong-payday pain worse, because everyone knows everyone, and the owner is the one doing the archaeology at midnight.
Everything we write about, we build. Thirty minutes maps it to your business.