How to reduce overtime costs at a guard company

Overtime is the biggest controllable cost in most contract security operations. You pay time and a half, you bill the client the regular rate, and the difference comes out of margin. That is how a company with healthy revenue loses money all year.

What does an overtime hour cost you?

Do the arithmetic once, on paper, before you touch the schedule. Take a guard near the national median guard wage of $18.29 an hour (BLS, May 2025). The overtime hour costs you about $27.44 in wages alone, before payroll burden. Bill it at a $28 standard rate and you have cleared 56 cents on an hour that still owes you taxes, insurance and overhead. You worked it at a loss.

Run that line for your own wage and your own bill rate. The result decides everything after it. If the gap is negative, overtime control is worth more to you this year than any rate increase you are likely to win, and it does not need the client's signature.

Why does overtime get out of control?

Four causes, and they feed each other.

Nobody can see the running total. A scheduler who learns about hour 43 when payroll closes cannot prevent it. The decision that created it happened four days earlier, on a phone call, with no number in front of anyone.

Callouts force emergency coverage. The post has to be filled tonight, so whoever answers the phone gets it and their current hours never come up. The guard already on site stays late, or someone scheduled off comes in at 38 hours and goes home at 46. Cutting the callout rate is overtime work, even though it files under HR.

The roster is structurally short. If you cannot cover the contract at straight time, no scheduling trick will fix it. That is a hiring problem wearing a scheduling costume, and it belongs with the staffing shortage.

Scope crept and the price did not. A client asks for four more hours on Saturdays. A supervisor says yes. Nobody reprices the contract, and the hours land on someone already at 40.

Where do the hours actually come from?

Before you fix anything, find out which bucket your overtime sits in. The chart below is a worked example rather than measured data, but it shows the shape operations tend to find the first time they categorize a month of it. Scheduling gaps and callout scrambles dominate. Admin hours are a rounding error that somehow absorbs all the attention.

Run the same breakdown on your own last quarter, sorted by site as well as by cause. One awkward contract is usually carrying most of it.

Scheduling changes that move the number

Check the hours before you fill the shift. This is the whole game, and most operations still cannot do it. A scheduler assigning Thursday night needs to know that this guard is at 36 hours and that one is at 22. Color coding, a threshold alert, a sorted list: the mechanism barely matters. What matters is that the number is on screen at the moment of the decision rather than in a report on Monday. Write the threshold down as a rule anyone can follow, which is what an overtime approval policy is for.

Shorten the shift blocks. Twelve-hour shifts give a guard three before overtime and leave you nothing to flex with. Eight-hour blocks give you five, and they let you cover a four-hour gap without pushing anyone over. The trade is more handoffs, which is a real cost, so settle your handoff procedure before you make the switch.

Build a float pool. Keep a bench of part-timers who actively want extra hours and have said which posts they can work. Coverage that would have cost time and a half comes off the bench at straight time instead. That only works if their availability is recorded rather than remembered, which is the difference between a pool and a list of phone numbers. Scheduling for lower overtime goes further into the mechanics.

What the scheduling software has to do

All of this is maintainable by hand on a roster of twelve and impossible by hand on a roster of two hundred. Four capabilities carry it.

TeamMap covers part of that: scheduling records clock-in and clock-out with a GPS stamp, so hours come from what the guard actually worked rather than a payroll import, and open shifts can be posted for guards to claim from the phone. Whatever tool you use, the test is the same. Does the number appear before the assignment, or after it?

How do you cut the callouts that cause it?

Emergency coverage is the most expensive coverage you will ever buy. It is also the most preventable.

Post uncovered shifts where guards can see and claim them from a phone. A good share of the time an open shift gets taken by somebody under 40 before a supervisor starts ringing round. Make early notice worth something, too. A guard who tells you on Tuesday that Saturday is a problem costs you nothing. The same guard calling at 5pm Saturday costs you eight hours at time and a half.

The rest of it is retention. People who like the post, the schedule and the supervisor turn up for them. Turnover work and overtime work are the same work.

Pricing the overtime you cannot avoid

Some overtime is real. A guard holds over because the relief's car died on the interstate. A client wants a body on site at 2am on Christmas. No schedule prevents either.

The mistake is absorbing it quietly. Price the historical rate into the contract from the start, bill premium hours at a rate that reflects what they cost you, and log the cause every time so the client conversation has evidence behind it. Setting your bill rate covers where the multiplier comes from.

What to measure

Overtime as a percentage of total hours, monthly, by site. That one number tells you whether anything you changed worked.

A rate that climbs while you are actively managing it is not a scheduling failure. It says the roster is too small for the contracts you sold, or that one site is priced below what it costs to staff. Both are fixable. Neither gets fixed by another round of calling people in.

Estimate what it is worth to you

Put your own guard count, average wage and current overtime rate into the calculator. The figures it returns are an estimate built from your inputs, not a promise, but they size the prize well enough to decide where to start.