Security guard retention: the five levers
Security guard retention comes down to five levers: pay structure, schedule, supervisor, promotion rungs and the first ninety days. Pay alone will not do it.

A guard resigns on a Tuesday. Three weeks later someone is standing at that post who has never seen the site, and the shifts in between got covered at time and a half. Every security guard retention lever below is cheaper than that.
Five levers decide retention: the check, the schedule, the supervisor, the rungs above the post, and the first ninety days. Pay is necessary and not sufficient. Companies that fix only the wage watch turnover barely move, because the guard who left at month four left over a schedule change.
Where does retention actually get decided?
Not in an HR program. In five ordinary operational places, most of which cost less to fix than to ignore.
Before you work through them, know which one is yours. Measuring your own turnover and splitting it by tenure tells you whether you have a hiring problem, a schedule problem or a pay problem, and those get different fixes.
What do you have to pay to keep a guard?
Enough that leaving for a dollar more is not obviously worth it. That is the bar, and it is lower than people assume.
Start from the market rather than from last year's budget. The Bureau of Labor Statistics puts median security guard pay at $18.29 an hour as of May 2025, and its state and metro wage table for security guards will tell you where your market sits against that. Then check live job postings, because the survey lags and the postings do not.
Pay at or slightly above market. The premium does not need to be large. What it needs to be is visible.
Then build some structure into it, so that staying produces something:
- Tenure steps. A defined increase at six months and at a year. Predictable, small, and it gives a wavering guard a reason to get to the next one.
- Skill pay. Additional cents an hour for CPR, fire watch, a specialty credential. This is the cheapest raise you will ever give because it buys you flexibility as well as loyalty.
- Site differentials. Pay more for the hard post rather than rotating people through it and hoping. The difficult site is difficult whoever you send.
- Real differences between performers. A guard who never calls out and one who calls out monthly earning the same rate is a message, and everyone on the roster reads it.
Benefits do work that wages do not. Even basic health cover separates you from the competitors offering none, and paid time off says something about how you see the job. There is more on the trade-offs in what a guard benefits package should contain, and on where the money comes from in how you set the bill rate.
What does a schedule people can live with look like?
This is the lever most companies underuse, and it is nearly free.
Post two weeks out. Guards with a second job, a class or a child in daycare are making commitments against your schedule, and a Thursday-for-Saturday post makes that impossible. Then hold the line: every last-minute change spends trust you need for the change you genuinely cannot avoid.
The rest of it is fairness that people can see:
Night work deserves specific attention. NIOSH's work-hours training, written for nurses but drawn from general shift-work research, concludes that shifts of eight hours or less are better tolerated during the evening and night than longer ones, consecutive night shifts on a rotating pattern should be kept to a minimum, and workers adjust better when they have some ability to control their schedules. If you are running twelve-hour overnights back to back, that is a retention decision as much as a coverage one.
Forced overtime is the fastest way to lose a good guard, and it usually comes from the same root as your turnover. Better scheduling and overtime control are the same project as retention, not a competing one.
Why does the supervisor decide more than the wage?
Because the supervisor is the company, as far as a guard on a lone post is concerned. They are the only person who calls, the only one who visits, and the one who decides whether a complaint gets heard or filed.
Most security supervisors were promoted for being reliable guards. Nobody taught them to manage people, and then we are surprised when they manage badly.
- Train supervisors on the people half of the job, not just post coverage and paperwork.
- Put turnover on their scorecard. What gets measured by supervisor gets managed by supervisor.
- Require site visits on a cadence, including overnight ones. A supervisor who has never seen the 3am shift is guessing about it.
- Back the guard when a client is unreasonable. Every guard on the account watches how that goes, and it sets what they believe about working for you.
- Act on the supervisor whose team keeps emptying. You already know which one it is.
What recognition looks like when it is not a plaque
Specific and quick. "Good work" in a group message is noise. "Ramirez caught the propped door at the loading dock on Tuesday, which is why we did not have a break-in" is recognition, because it names the person, the act and the consequence.
Service anniversaries matter more here than in most industries, precisely because longevity is rare. A guard who has held a post for four years in a business with this much churn has done something unusual, and saying so out loud costs nothing.
What does not work: recognition that arrives only in the quarter you noticed turnover was up. People can tell.
What can a guard be promoted into?
Flat organizations still have rungs if you build them. The point is not the title. It is that year three looks different from year one.
Field training officer is the easiest one to create and the most useful, because your best guard training your newest hire fixes two problems at once. Site lead gives someone ownership of a post without a full supervisor role. Beyond that: specialist tracks in investigations, a vertical you work heavily, or a training role.
Pair each rung with training that has value outside your company. Guards notice when an employer invests in credentials they can keep. A training program with a syllabus beats ad-hoc shadowing, and keeping the certifications current is the part that quietly lapses.
Publish the criteria. An advancement path nobody can describe is not a path.
What happens in the first ninety days?
If your turnover is concentrated early, and for most companies it is, this section is the whole job.
- Between offer and first shift, stay in contact. A candidate with three weeks of silence is still interviewing elsewhere. A call at day two and day seven costs nothing.
- Do not put a new guard on a post cold. Someone walks the site with them, points at the things the post orders describe, and answers the question they are too new to ask.
- Get the first paycheck exactly right. Nothing else you do in the first month matters as much. Check it by hand if you have to.
- Call at day 30 and day 60. Not a performance review. One question: is this what you expected.
- Watch the callouts. A new guard's first unexplained callout is usually the beginning of the resignation, not an isolated event. Acting on it early is how you keep them.
None of this is complicated. It is just work that nobody owns unless you assign it, and it competes badly against whatever is on fire this morning. Give it to a person and a date, the same way you would with a client escalation, and the numbers move.
Key Takeaways
- Pay at or slightly above market, then add tenure steps, skill pay and site differentials.
- Post schedules two weeks out and distribute the undesirable shifts visibly. Nearly free, badly underused.
- Keep evening and night shifts to eight hours where you can, and minimize consecutive nights.
- Train supervisors to manage people and put turnover on their scorecard.
- Most losses land in the first ninety days: contact before the start, a walked site, a correct first paycheck.
Continue Reading

How to measure guard turnover in your own company
No agency publishes a guard turnover rate. Calculate your own, split it by tenure to find the ninety-day cliff, and cost one departure from your own ledger.

Security guard staffing shortage: what works
The guard shortage is churn, not demand. Where the pay gap comes from, when guards actually quit, and the staffing fixes that hold a roster together.

Customer service for security companies that win
Customer service is what security clients renew on. Teach officers to greet, know the building, enforce rules by explaining them, and measure how it lands.