How to scale from 10 to 100 guards without adding admin overhead

At ten guards, the owner knows everyone. They know who is reliable, who is covering which site, and who to call when someone does not show. The schedule fits on one screen. Post orders live in the owner's head and get transmitted by conversation.

This is not a failure of professionalism. It is genuinely efficient at that size. Formal process would be slower than just knowing. The trap is that it works well enough that nobody replaces it until it has already broken.

The underlying reason is that manual coordination grows with the number of relationships between people and sites, not with the number of guards. Doubling the roster does considerably more than double the coordination load, which is why growth feels manageable and then abruptly does not.

This is first because everything else depends on it. While site knowledge lives in people's heads, every new guard needs an experienced guard to train them, every absence is a crisis, and supervisors have no standard to supervise against.

• Patrol routes, checkpoints, and how often rounds run
• Access rules: who is allowed in, and how visitors and contractors are handled
• Emergency procedures and the escalation chain, with names and numbers
• What goes in the DAR, and what triggers an incident report

The test is whether a guard who has never worked the site can read the post orders and work it correctly. Until that is true, you cannot staff flexibly, and every schedule change is constrained by who happens to know the site.

At ten guards a supervisor can check on everyone. At sixty they cannot, and the usual response, checking a sample, means most problems are still found by clients rather than by you.

Exception-based supervision inverts it: assume shifts are running normally and spend supervisor attention only where the data says otherwise.

• A guard who has not clocked in ten minutes after shift start
• A shift with no DAR submitted by the expected time
• A patrol round missed or run outside its window
• An incident report filed at a severity that needs a supervisor now

This is the change that most directly decouples headcount from admin cost, because supervisor time stops scaling with the number of guards and starts scaling with the number of problems.

Scheduling by text has no single source of truth. Two guards can believe they have the same shift, or neither can. Availability lives in a conversation history nobody can search. Every change requires the scheduler personally.

Publishing the schedule where guards can see their own shifts removes an entire category of coordination work, because most of the messages a scheduler answers are guards asking what they already could have looked up.

Paper DARs impose a fixed admin cost per shift, per day, forever: collect, chase, sort, retype, file. That cost scales linearly with the roster and is invisible because it is spread across people doing other jobs.

Capturing reports digitally at the moment they happen removes the handling entirely rather than making it faster. It also removes the failure mode where the report simply never arrives, which is the one that costs contracts.

1. 1Write post orders for every site. Nothing else works reliably without them.
2. 2Move clock-in and reporting off paper, so you have data to supervise with.
3. 3Switch supervision to exceptions once that data exists.
4. 4Publish the schedule so guards self-serve.
5. 5Add a supervisor layer. By then you have written standards and real data for them to work from.