Guard overtime management: caps, approvals, callouts

Bill a post at $25 and pay the guard $18, and you have $7 an hour for burden, overhead and profit. Push that guard past 40 hours and you pay $27 to collect $25. Overtime inverts the margin, so it needs a policy.

What does the law actually require?

Start here, because most scheduling mistakes are mistakes about the rule rather than about the roster.

The Department of Labor states that covered employees must receive overtime for hours worked over 40 in a workweek at not less than time and one-half their regular rate. Three details in that fact sheet cost operators money regularly.

• The workweek is fixed. DOL defines it as "a fixed and regularly recurring period of 168 hours, seven consecutive 24-hour periods." It does not have to start Monday, but once you set it, that is the boundary every hour gets counted against.
• You cannot average. "Averaging of hours over two or more weeks is not permitted." A 48-hour week followed by a 32-hour week is eight hours of overtime, not a wash.
• Weekends and holidays are not overtime by themselves. The FLSA "does not require overtime pay for work on Saturdays, Sundays, holidays, or regular days of rest, as such." Only the hour count triggers it federally.

Why does overtime happen in your operation?

The cause determines the fix, and four causes cover almost all of it. Only one of them is solved by better scheduling software.

Write down which of the four your last month of overtime came from. If you cannot tell, that is cause three, and it is the one to fix first.

How do you see overtime before it happens?

You cannot manage a number you only see after payroll closes. Whatever the scheduler is looking at when they assign a shift has to answer three questions on the same screen.

• Hours to date, this workweek, per guard. Not last week. Not the pay period. The 168-hour window the assignment falls into.
• Overtime by site. Which contract is generating it. That number belongs in the renewal conversation, and it is the evidence behind repricing a post that has been losing money quietly.
• The trend. Whether the overtime line is climbing, and since when. A single bad week is weather. Six climbing weeks is a staffing decision you have not made yet.

Most operations running on a spreadsheet have the first number, somewhere, one tab away. That distance is the whole problem. The scheduler under pressure at 5pm does not go looking.

How deep does the bench need to be?

Deep enough that a callout does not automatically become a premium hour. There is no published ratio worth quoting, so derive your own: take your callout rate and your turnover, both of which you can measure from your own records, and hold enough trained, badged guards per regularly scheduled position to absorb them.

The ratio also has to be site-specific. One armed post with three qualified people on the whole roster is a coverage crisis waiting for one of them to get the flu, no matter how healthy the company-wide number looks.

Two things widen the bench without hiring. Cross-train guards across sites so more people can cover an opening, which requires post orders good enough that an unfamiliar guard can work the site. And recruit before you are short. Hiring while already running overtime means paying the premium the whole time you interview. There is more on that pipeline in hiring guards who show up and on the market conditions in the staffing shortage.

Do shorter shifts actually help?

Federally, yes, for a specific reason: eights divide into 40 and twelves do not.

Take a guard sitting at 32 hours on Thursday. An eight-hour shift finishes the week at exactly 40. A 12-hour shift on the same day is four hours of premium pay you did not budget. Repeat that across a roster and it is the difference between a profitable contract and a break-even one.

Flexible start and end times do the same work at smaller scale. A guard at 38 hours can take a six-hour shift. Insisting every shift is exactly eight forces two hours of premium to solve a six-hour problem.

Twelve-hour shifts still win on some sites: fewer handoffs, better continuity, and a lot of guards prefer three long days. Price them knowing what they cost, and check your state's daily rule before you commit.

Who approves overtime, and when?

Before it happens. An approval workflow that runs after the fact is a reporting workflow.

The reason field is the part people skip and the part that pays. Three months of it will tell you whether you have an attendance problem, a headcount problem or one client who keeps asking for extra coverage on Friday afternoon and never gets billed for it.

What do the patterns usually turn out to be?

Three findings come up again and again once the reasons are logged.

Callouts driving most of it. That is not a scheduling problem. Guards call out of posts they dislike, schedules they cannot plan around, and jobs they are leaving anyway. Fixing it lives in reducing callouts and keeping the people you have.

The same names every week. Some guards want overtime and will quietly stay unavailable for the shifts that would spread it around. Others are simply the only ones who answer the phone. Both are worth knowing, and only one is a problem.

One site, every week. That post is understaffed or underpriced, and usually both. Fix the roster or fix the contract, but stop paying the difference.

Where does software actually help?

It helps at the moment of assignment, and mostly nowhere else. A tool that emails an overtime report on Monday has told you about money you already spent.

What is worth having: running weekly hours on the schedule while you are building it, an alert at the threshold rather than after it, open shifts a guard can claim so a hole gets filled without a phone tree, and no-show alerts so a callout surfaces while there is still time to cover it another way. Timesheets that export to CSV close the loop with payroll.

How much you save depends on how much preventable overtime you have now. An operation scheduling on paper usually finds a lot. One with a disciplined manual process finds less, and should be honest about that before buying anything. Cutting overtime costs covers the money side, and schedule design covers the roster patterns themselves.