The three numbers to read on a hotel daily labor report
· labor cost, scheduling, overtime
A daily labor report is mostly a record of decisions you can no longer change. Yesterday’s hours are worked and owed; yesterday’s cost is spent. Three numbers on it are still live — the punches that aren’t clean yet, the week-to-date hours of anyone near an overtime threshold, and the gap between what you scheduled and what you actually paid for. Read those three, in that order, and the rest of the report is reference material you can skip on a normal morning.
Why the biggest number on the page is the wrong place to start
Most daily labor reports lead with a total: yesterday cost $1,340, or labor ran 29% of revenue. It is the largest number on the page and the least useful one at 8am. It is already spent — nothing you do today moves it — and a single day’s total is noisy enough to be close to unreadable. At a 38-room property, one unscheduled deep clean and a late departure can swing a day’s hours by 10% with nothing wrong underneath.
The monthly percentage is a real number with a real job — what a good labor cost percentage looks like is worth reporting to an owner or a lender. It is just a monthly number. Read daily, it produces a dozen false alarms a month and one real one you will have stopped believing by the time it arrives.
Number one: how many punches aren’t clean yet
This will feel like the wrong place to start, because it is a count of paperwork rather than a number in dollars. It is first because it is the error bar on everything else. A missing clock-out sits in the report as either zero hours or sixteen. An edited punch nobody has checked is a number the report treats as fact. If four of yesterday’s punches are unresolved, the day’s hours total is not yet a number — it is a draft, and the two numbers below inherit its uncertainty.
“Exception” covers more than the obvious one:
- A missing clock-out. The shift has no end, so it has no duration.
- An edited punch. Someone changed a time. Not necessarily wrong — just not yet checked by anyone.
- An implausibly long shift. Thirteen hours at a property that never schedules past 9pm usually means somebody forgot to clock out.
- A clock-in well before the scheduled start. Early drift is the quietest way a labor line grows, because each instance is six minutes and no single one is worth a conversation.
Resolve them the same morning, while the person who worked the shift is reachable and still remembers. A week later it is a reconstruction, and reconstructions have a direction: they trend toward whatever the employee recalls, which is rarely fewer hours. Zero by the end of the morning is the target. A property that runs four or five every day has a training or hardware problem underneath, not a payroll one.
Number two: week-to-date hours for anyone near the line
This is the only number on the report where reading it today changes what you pay this week.
Overtime is not decided at the end of the week. It is decided on the last day you could still have moved a shift. Federal law (29 U.S.C. §207(a), the FLSA) requires overtime past 40 hours in a workweek, and several states add a daily threshold on top; the details differ by state and they change, so check your own state’s rule rather than assuming the federal one is all that applies. The arithmetic below works the same way against whatever threshold you operate under.
A worked example on an invented 38-room independent, workweek Monday to Sunday, read on Thursday morning — so the report covers through Wednesday:
| Room attendant | Mon | Tue | Wed | Week to date | Still scheduled (Thu + Fri) | Projected week |
|---|---|---|---|---|---|---|
| Attendant 1 | 8.5 | 8.25 | 9.0 | 25.75 | 8 + 8 | 41.75 |
| Attendant 2 | 7.5 | 8.0 | 7.75 | 23.25 | 8 + 8 | 39.25 |
Attendant 1 is 1.75 hours into overtime that has not happened yet. At $16 an hour, the extra cost of those hours over straight time is the half-time premium — $8 × 1.75 — so $14.
Fourteen dollars is not a crisis, and pretending otherwise is how operators learn to ignore a report. The pattern is the point. If two of eight hourly staff drift two hours past the line most weeks, that is $8 × 2 hours × 2 people = $32 a week — roughly $1,660 a year at one property, paid for hours nobody decided to buy.
On Thursday morning the fix is cheap: trim two hours off Friday, or move Friday’s overlap to the attendant sitting at 23.25. On Sunday night there is no fix at all, only a number. Sometimes the right call is to let it run — the rooms need cleaning, the arrivals are booked, nobody else is free. A report does not make that decision; it makes sure you made it on purpose.
One caveat: the projection assumes the scheduled hours are the hours that get worked. That assumption is exactly what numbers one and three exist to test.
Assembling these by hand each morning — export the punches, sum the week per person, line it up against the schedule — is why most people do it for two weeks and then stop. ProfitClock emails the report instead: yesterday’s hours, what they cost, and what needs attention, per property. Start a free trial if you want to read one against your own week.
Number three: scheduled hours against actual hours
Third, because it shapes tomorrow rather than today. Say the property scheduled 46.0 hours yesterday and paid for 51.5. That is +5.5 hours, about +12%, or roughly $88 at a $16 blended rate. The sign matters less than the shape, because two different problems produce the same +5.5:
- Concentrated — nearly all of it in one department on one day. That is usually demand. Fourteen departures against a schedule built for nine is five extra full cleans, and a checkout takes materially longer than a stayover, so the overage is arithmetic rather than slack.
- Spread thin — twenty to forty minutes on almost every shift. That is drift: early clock-ins, late clock-outs, a break that quietly stopped being taken. It never shows up large enough on any single day to be worth raising, and over a quarter it is the more expensive of the two.
They need different responses. Concentrated variance is a forecasting fix — schedule against the departure list instead of against last Thursday. Spread variance is a policy and supervision fix, and no report solves it on its own.
A single day tells you almost nothing here. Variance is only readable as a running number: the same weekday against the last four, or a rolling seven-day total. One Thursday at +12% is weather. Four Thursdays at +12% is a schedule that is wrong on Thursdays.
What a daily labor report cannot tell you
Worth being plain about, because the report looks equally confident about all of it.
Whether the standard was right. A property that takes 45 minutes on every checkout looks perfectly consistent. Consistency is not correctness — the report measures actual against the schedule you wrote, so a generous schedule gets confirmed every day for years.
Anything, reliably, at a very small property. With three hourly staff, one dentist appointment is a 15% variance. Below roughly five or six hourly employees, read weekly totals and skip the daily variance line; the exception count and the overtime pace still work at any size.
Revenue. None of the three numbers is a profitability number. Labor per occupied room is the bridge to that question and needs occupancy alongside hours — MPOR is the metric for it, and it answers a different question from this one.
A disputed punch. It can flag one, which is the useful half.
The five-minute version
- Open exceptions first. Resolve or assign every one before reading anything else. Until that count is zero, the rest of the report is an estimate.
- Scan week-to-date hours for anyone within about four hours of a threshold. Decide deliberately: trim, or pay it.
- Read yesterday’s variance against the last four of the same weekday, never against zero.
- Close the report.
Everything else on it — the totals, the percentages, the department breakdowns — is a monthly conversation that happens to be printed every day.
Start seeing what labor actually costs
Set up your property, put staff on the clock, and read tomorrow morning’s labor number.
14-day free trial · No credit card · Cancel anytime