From wage to post#
The U.S. Bureau of Labor Statistics puts the median hourly wage for security guards at $18.46 (May 2024). That number is the start of the cost, not the cost. Four multipliers sit between a guard's wage and what a site pays for a staffed post.
1. FTE per post
A post staffed around the clock is hours a year. The common shortcut divides by 2,080 paid hours and gets FTE. But paid hours are not productive hours: vacation, holidays, sick leave and training take 150–280 hours a year out of each guard. With productive hours per FTE:
Our model's median is 4.62 FTE per post — about 10% above the shortcut. Budgets built on 4.2 are short by roughly one guard in ten.
2. Billed rate
where is wage, payroll burden (taxes, insurance, benefits), contractor markup (overhead and margin), the share of hours paid at time-and-a-half, and supervision overhead. Annual post cost is — the FTE factor is already inside the billed hourly rate, because the contractor prices it in.
Model and assumptions#
| Input | Distribution (triangular: low / mode / high) | Source |
|---|---|---|
| Guard wage | $16.00 / $18.46 / $28.00 | Mode = BLS May 2024 median |
| Payroll burden | 1.18 / 1.28 / 1.45 | Assumption |
| Contractor markup | 1.20 / 1.35 / 1.60 | Assumption |
| Productive hours / FTE | 1,800 / 1,890 / 2,000 | Assumption |
| Overtime share | 0% / 5% / 15% | Assumption |
| Supervision | 5% / 8% / 12% | Assumption |
| Sentries per post replaced | 0.50 / 0.67 / 1.00 | Assumption (site-dependent) |
| Sentry-hour price | $8.50 / $9.80 / $11.00 | Mode = Tencer list price |
| Hive (Roost) cost share | 6% / 7% / 10% | Tencer pricing |
| Incremental monitoring per post | $0 / $2 / $6 per hour | Assumption |
We drew 100,000 samples (guard_cost.py, seed 5). These are assumptions, stated so you can replace them with your own; the BLS wage is the only externally sourced number.
Results#
| Quantity | P5 | Median | P95 |
|---|---|---|---|
| Billed rate per post-hour | $33.20 | $41.30 | $53.25 |
| Annual cost per 24/7 post | $290,845 | $361,819 | $466,467 |
| FTE per post | 4.45 | 4.62 | 4.79 |
| Autonomous cost per post replaced | $66,653 | $89,400 | $116,010 |
| Saving per post replaced | 65.2% | 75.5% | 83.3% |
A median 24/7 post costs $361,819 a year; the middle 90% range runs from $290,845 to $466,467. Replacing it with autonomous coverage costs a median $89,400. In 100% of draws the saving exceeded 50%; in none was it negative.
What the saving is most sensitive to#
The largest single swing is monitoring cost, not hardware price. A site that adds dedicated operators for the robots can give back a third of the saving; a site that folds them into an existing operations center keeps nearly all of it. Second is the guard wage — high-wage metros make the case stronger. Third is how many sentries it takes to replace a post, which is a site-survey question (see the viewshed study). The sentry-hour price itself ranks low.
Break-even and small sites#
Two practical questions follow from the model. First, how much monitoring cost would erase the saving? With other inputs at their medians, saving per post is where is the billed guard rate and the hourly autonomous cost. Break-even () happens at a monitoring cost of
— in other words, only if you staffed a dedicated full-time operator for each post replaced. The design lesson is to fold robot monitoring into an existing operations center, where marginal cost per post is a few dollars an hour.
Second, small sites. Our Site Watch minimum is two Kestrels and one Roost. That floor costs about $183,715 a year. For a site replacing a single 24/7 post at the median $361,819, the saving is about 49% — still positive but well below the per-post figure, and within the noise of the wage assumption at low-wage sites. The economics strengthen quickly with two or more posts, which is why single-post sites are usually better served by a fixed tower or trailer plus remote monitoring.
What the model leaves out#
Guards do things robots do not: customer service, first aid, physical intervention, judgment in ambiguous situations. Almost no site should go to zero staff; the realistic design keeps a smaller number of mobile responders. The model also omits transition costs, integration work, contract exit terms, and the value of better detection (which is the point of the exercise but hard to price). For the commercial framing of these trade-offs see security robots vs. security guards.
Frequently asked questions#
How much does a 24/7 security guard post cost per year?
Our Monte Carlo model, anchored on the BLS median wage of $18.46/hour, gives a median of about $362,000 per year for a contracted 24/7 post, with a 90% range of roughly $291,000 to $466,000.
How many guards does a 24/7 post need?
8,760 hours divided by productive hours per guard. With 1,800–2,000 productive hours, that is 4.4–4.9 FTE; the common 4.2 figure ignores leave and training.
What drives security robot ROI?
In our sensitivity analysis: monitoring cost, guard wage level and how many robots it takes to replace a post — more than the robot's hourly price.
References#
- U.S. Bureau of Labor Statistics, Occupational Outlook Handbook, “Security Guards and Gambling Surveillance Officers”: median hourly wage for security guards $18.46 (May 2024). bls.gov
Reproducibility: model code is published at /research-code/ with a README. Results are model outputs under the assumptions stated above, not field trials.