Running the Office

Payroll Week Shouldn’t Start With Three Phone Calls

By Travis Evans, Founder of TradeGOAT · 20+ years in the trades

On most of the jobs I have run, labor is the biggest number on the sheet. It is also the number that gets written down last, and usually by somebody who was not there. I have twenty years in the elevator trade and four running my own company, and I still catch myself doing payroll on Friday afternoon from a handful of sources that do not agree with each other.

You probably know the sources. A paper sheet in the truck that got rained on. A guy who texts his hours Sunday night, three days after the fact. A guy who does not text them at all and expects you to know. The lead who wrote everybody’s hours on the back of a work order and then took that work order to the next job.

So the week gets reconstructed instead of recorded. That is a different activity, and it costs more than the hour it takes.

Reconstructed hours are estimates wearing a payroll number

When somebody tells you on Friday what they worked on Tuesday, you are not getting a record. You are getting a recollection, and recollections round. They round up on a long day and they round to a clean number on a normal one, and almost nobody is doing it on purpose.

Two things go wrong from there. The first is payroll, which is the one everybody thinks of. The second is quieter and matters more: those same hours are the labor cost on the job. If the hours are soft, then the labor line on that job is soft, and so is whatever you thought you made on it. You will carry that number into your next bid on the same kind of work.

The other cost is the trip. Somebody in the office spends Friday afternoon walking around asking three people what they did on Tuesday, which is time nobody planned for and nobody bills.

Let the record get made where the work is

The fix is not a better spreadsheet on Friday. It is that the hours get written down at the moment they happen, by the person they belong to, from the phone already in his pocket.

In TradeGOAT the crew clocks in and out from the phone at the job. Switching from one job to the next is a clock-switch rather than a clock-out and a new clock-in, because that is how the day actually goes. Then the week comes back to the office as a timecard with a row per person, hours and days on it, and it is already there Friday morning instead of being assembled Friday afternoon.

The part that saves the most time is not the punching. It is that the week arrives already sorted into the rows that are fine and the rows that need you.

The rows that need a look are marked

Every row on the week either has a reason to stop on it or it does not, and the screen says which. The ones it marks are the ordinary field realities:

• Auto-closed. Somebody left the clock running, and it got closed out at the hour your company set instead of billing a nineteen-hour Tuesday.
• Missed clock. A day on the schedule with no punch against it at all.
• Short. The week came in under what that person was scheduled for, with the gap named in hours so you know whether it is worth a call.
• OT. The overtime split out on its own, on your workweek and your threshold, not a default somebody else picked.
• Off-site. The punch did not land where the job is.

None of those is an accusation and none of them changes anybody’s hours. They are the office’s to-do list for the week, and the only thing the tool is claiming is that it noticed. What a flag means is your call. You know the difference between a guy who forgot to clock out and a guy who worked until nine, and the software does not.

Then there is one button for everything else. Approve all unflagged takes the clean rows in a single move, so the only rows left in front of you are the ones you were going to look at anyway.

Two small things we argued about

The first one is the sentence next to that button when there is nothing for it to do. It used to say that flagged rows needed a look, every time, which is fine right up until the reason is something else. An office whose whole crew was still clocked in got sent hunting for flagged rows that did not exist. Naming the wrong cause is worse than naming none, so now it says the true one: someone is still on the clock, or flagged rows need a look, or the rest are already approved, in whatever combination is actually the case.

The second is what happens after you approve. If an entry inside an approved week gets edited later, the week does not stay quietly approved with different numbers under it. It goes back to Reopened and waits for you to approve it again. An approval should mean you looked at these hours, not you looked at some hours once.

The hours land on the job, too, while there is still job left

Payroll is only half of what the hours are for. The other half is the job, and that is where most of the money is decided.

On a job’s scorecard, the hours you quoted sit next to the hours the crew has actually clocked against it. Not after it closes. While it is running, when a two-day gap is still something you can do something about.

The clock keeps its own score there. Clocked time does not go rewrite the labor cost you priced the job with, and it does not decide anything about the job. It sits next to your number so you can see the two of them together. What the gap means is yours to decide. It could be a bad estimate, a hard install, or a change nobody wrote up, and you are the one who knows what happened on that floor.

What this actually buys you

A week you can approve in a few minutes instead of an afternoon. Labor numbers that came from the day they describe. And a job cost that is built while the job is running, by the people running it, instead of assembled afterward from what everybody can still remember.

This is one of the parts of TradeGOAT that came straight out of my own shop, because I was the guy making the three phone calls. If you run into the version of it I have not hit yet, tell me and it will shape what gets built next.