The word is perimeter security, and it splits cleanly in two. Detection is knowing that something is happening: the layer that classifies an event and raises it. Delay is the physical barrier making the thing take longer to do: the layer that converts an intention into elapsed time.
Vendors sell detection as a product, tuned and benchmarked at the point of classification. Buyers inherit the barrier as a condition of whatever hardware happens to be standing at the time. The two halves are paid for by different parties, on different schedules, and only one of them is ever bought as a line of security spend.
Where a False Alarm Is Actually Priced, and Who Pays for It
Programs built to reduce false alarms describe the problem in terms of what went wrong at the alarm. The Douglas County Sheriff’s Office runs one such program and names the causes: user error, installation or service error, and equipment failure. Each named cause belongs to the system that gets serviced on a schedule. The barrier itself is not part of that schedule. That is one program’s published framing, representative of how the problem gets described where a list exists.
Two city fee schedules put a number on the consequence. A fee schedule records an occurrence and a date against a registered system; it does not record how long the response took or what condition the perimeter was in when the alarm fired.
The City of Atlanta‘s false-alarm program treats a first false alarm as a warning and attaches a $50 fine to the second, charged to the homeowner. Cincinnati’s program begins later: fees start at the third false alarm on a registered system. These remain two municipal schedules, and neither should be read as a national rate or a typical figure.
What Atlanta’s schedule prices is a dispatch, charged per occurrence to the alarm owner. A fine is priced against consuming a public resource; it says nothing about what that resource costs to run. Once that substitution is separated out, the money in an alarm goes toward the response side; what analytics improves lives on the classification side.
The Alert Is Only Worth the Time It Buys
If the only charge on record falls on the response side, then what a true alert is worth is whatever the response can do with the time it has. That gap is the interval: the span between a breach beginning and a response beginning to act. The alarm’s own records do not measure that gap. A barrier is a delay device before it is anything else, and the interval is a property of its present condition. The specification sheet dates from the install.
Detection software treats the perimeter as static, fixed at whatever specification it was given at install. Maintenance state changes the delay without changing the alert: a barrier that has been cut, bridged, overgrown or left unpowered raises the identical event and grants less time to respond. The same classification is worth different amounts on different days, invisible to whatever does the classifying — a classifier’s output stops at the event; how long the response has left sits outside anything it measures.
Veritech Security, in its own published material, recommends vegetation and sightline upkeep along a perimeter as a measure against false alarms.
Maintenance state is an input to both sides of the equation at once: it raises the count of spurious events on one side, and it lowers the worth of the genuine one on the other. It is also the one input neither the cause lists nor the fee schedules touch. Which turns the whole question into one of scale.
An Industry Without a Denominator
An industry-wide false-alarm rate would need a denominator: every alarm event, real and spurious, across every registered system in every city that keeps one, counted somewhere in one common record. City schedules and vendor material between them fall short of that denominator, and no yearly price for false alarms appears in either.
What is published instead is the same fee-schedule pair from Atlanta and Cincinnati, alongside reduction percentages that analytics vendors publish about their own products. Those percentages are vendor self-reports about classification performance at the sensor; they say nothing about an industry-wide rate.
Without a rate and without a cost base, a perimeter cannot yet be priced in actuarial terms. What remains is a question of instruments: what actually gets documented about a barrier’s condition, by whom, and on what occasion.
What a Certificate Actually Transfers
A compliance certificate is an attestation: a named party states that a specified installation met a defined standard on a given date, and signs that statement. Its weight comes from the signer’s identity and the date — what makes it usable later, to a purchaser at transfer. Structurally it is a dated, signed statement of condition. The date on it marks when it was issued; the condition of the barrier after that date is outside what the document attests.
South African law makes that signature compulsory. Since October 1, 2012 the Electrical Machinery Regulations, under the Occupational Health and Safety Act 85 of 1993, have required a valid certificate of compliance for electric fence systems, and the requirement bites hardest at the transfer of a property carrying a live system. Electric Fencing Repairs, a South African site that routes property owners’ inquiries to independent installers, issues no certificate of its own — the installer who inspects the system is the party that signs one, and the party that prices the work.
Certification tends to be filed as a line item on an install spec. As an object it is something else: a named signature and a date against the barrier’s condition — the only record in the stack that ties a name and a date to it. The half of the word that gets watched reports continuously; the half that decides the interval is documented only when an event forces it.
Budgets Are Being Written for the Cheaper Half
The budget now splits three ways. The response side carries the only charges in evidence — set per occurrence, by a city, and in Atlanta’s schedule charged to the homeowner. The detection side is where spending is measurable and improvement demonstrable, which is why it is the easier side to fund. The barrier’s condition decides the interval and appears nowhere in the fee-schedule record.
An analytics upgrade lowers the number of events reaching a human. The time a response has left comes from somewhere else entirely — the barrier. Both halves need a line item, and only one currently gets one.
The same asymmetry sits under any stack where a maintained physical barrier is watched by software — gates, doors, turnstiles, bollards — and it widens as the detection layer gets cheaper. The barrier decides the interval whether or not anyone budgets for it.
