The False Alarm Follow-Up SOP

Purpose

A false dispatch costs the customer a fine in many jurisdictions, costs police time, and costs the dealer a little credibility each time, because a city that sees the same accounts over and over starts treating the whole dealer's signals with less urgency. Most false alarms have a findable cause: a user who armed with a door open, a motion detector looking at a heat source, a contact that drifted. The cause is easiest to find the morning after, while the customer still remembers who was home and what they heard. This procedure makes every dispatched alarm someone's ticket by 9:00 the next morning and closes it only when the cause is named and dealt with.

Scope

Covers every alarm the station dispatched police, fire or a guard to on a dealer's account, from the daily pull to a 30-day check that the cause is gone: classifying, the customer call, fixing or training, recording, flagging repeat accounts, and supporting a permit appeal with records.

Does not cover the repair visit itself, which runs under 'The Alarm Service Call SOP'. A fire dispatch from a combination panel is reviewed by a licensed tech under 'The Fire-Side Work and License Routing SOP'. If the customer believes there was a real intrusion, the event is not a false alarm and leaves this procedure at step 3.

Roles and responsibilities

Procedure

1. Pull every dispatched alarm by 09:00. The office runs the station's dispatch report for the previous 24 hours (72 hours on a Monday) and opens a ticket for each event. Acceptance: the number of tickets equals the number of dispatches on the report, written on the daily log. Wrong looks like a report showing alarms canceled by passcode mixed in with dispatched ones; the stop rule is only events with a dispatch get a ticket here, and canceled alarms are reviewed weekly as a trend. Hazard: none, desk work.

2. Read the history around each event and classify it provisionally. For each ticket the office reads the 30 minutes before and after the alarm (see related: Reading Event Codes on a Signal History): the arming event and by which user, any entry/exit zone opened, which zone alarmed first, whether a cancel was attempted. Provisional classes: user error, device, environment, unknown. The evidence each one leaves differs. User error usually shows a user event close to the alarm: an arming with an entry/exit zone still open, an entry zone opened and no disarm inside the delay, a wrong-code attempt, a cancel that came too late. Device usually shows a trouble or a restore pattern on the same zone in the days before, or an alarm with no restore. Environment usually shows an interior or perimeter zone alarming alone with no door event, often at a repeatable time or in weather. Acceptance: one class and the evidence for it written on the ticket. Wrong looks like "user error" written with no user event in the history; the stop rule is the class is unknown until the call. Hazard: none, desk work.

3. Call the customer before 11:00. Use the script below and record the answers in their words. Acceptance: who was home, what they heard or saw, and anything new at the house, written on the ticket. Wrong looks like a customer who says nobody was home and something may have been disturbed; the stop rule is the event leaves this SOP, the customer is told to contact police directly if they believe someone entered, and the office does not reset or test anything until the customer says the police are done.

OFFICE: Morning, this is the office at the alarm
        company. The system went into alarm at your
        house yesterday at 2:12 in the afternoon and
        the police were sent. Is everyone all right?
CUSTOMER: Fine, nobody was home. Police said all
        doors were locked.
OFFICE: Good. It came from the family room motion,
        with no door opened first. Has anything
        changed in that room in the last few days?
CUSTOMER: My daughter's birthday was Sunday. There
        are still balloons floating around in there.
OFFICE: That fits. A balloon drifting on warm air
        can trip a motion. Can you take them out of
        that room, or let them go down, before you
        arm today? And we'll check that sensor on a
        visit this week. Thursday morning work?
CUSTOMER: Thursday's fine.

Hazard: none, a phone call.

1. Act on the cause. User error: retrain by phone the same day, or book a visit for anyone who cannot arm and disarm reliably (see related: How to Train a Customer on Their Alarm System). Device or unknown: book a service visit within 3 business days. Environment: remove the cause with the customer if they can, and book a visit to check the device's aim or location. Acceptance: one action per ticket with a date. Wrong looks like an environment cause closed with "customer will watch it"; the stop rule is the device is still checked on a visit, because an explained alarm can also be a detector aimed where it never should have been. Hazard: the visit carries its own hazards under the service call SOP; this step books it.

2. Record the cause on the account. Close the ticket with one of the four classes, the specific cause in one sentence, and the action. Acceptance: class, cause, action and date on the account record, so the next false alarm is read against this one. Wrong looks like the class left as unknown after a visit found the cause; the stop rule is the ticket is reopened and corrected. Hazard: none, desk work.

3. Flag repeat accounts. An account with 2 dispatched false alarms in 90 days (a shop default) is flagged for the owner's weekly review, whatever the causes. Acceptance: the flag on the account with both ticket numbers. Wrong looks like two "environment" alarms closed separately and never seen together; the stop rule is the owner reviews them as one problem and proposes a fix to the customer, such as a different detector, cross-zoning, or the false-alarm reduction features a panel offers (see related: False Alarm Reduction Features in the SIA CP-01 Standard). Hazard: none, desk work.

4. Support a permit appeal with records. Many jurisdictions fine after a set number of false alarms in a permit year and some allow an appeal or a waiver, for example when a device failure was repaired; the count, the fine and the appeal process vary by city and county (see related: Alarm Permits, False Alarm Fines and the Dealer's Part). Where the customer appeals, the office provides the ticket, the cause and the repair record on request. Acceptance: the records sent within 2 business days of the request. Wrong looks like the office telling the customer the fine will be waived; the stop rule is the office supplies records and never predicts the outcome. Hazard: none, desk work.

5. Check the account at 30 days. Read the account history 30 days after the action. Acceptance: no further alarm from the same zone or cause. Wrong looks like a repeat from the same zone; the stop rule is a tech visit even if the cause was "user error". Hazard: none, desk work.

The record this produces

The daily log of dispatch counts and ticket numbers, each ticket's class, evidence, the customer's words, the cause and action, the repeat flag, any permit records sent, and the 30-day result. The owner reads the flags weekly; the office reads a ticket's history the next time the same account alarms; the customer receives the repair record when appealing a fine.

A worked pass

Step 1, Tuesday. The report shows 3 dispatches for Monday; 3 tickets opened.

Step 2 for the first ticket. Family room motion (zone 4) alarmed at 14:12, armed away at 08:05 by user 2, no entry/exit zone opened before zone 4. The office wrote "user error, someone returned" as the class.

Step 2 missed its acceptance. "User error" went on the ticket with no user event near the alarm, which is the wrong-looks-like; the class should have stayed unknown until the call. The customer, in the call above, says nobody was home: no user event, no door, so it cannot be user error. Stop rule taken: reclassified as environment, pending the visit, with the balloons as the likely cause.

Step 4. Visit booked for Thursday. The tech finds the motion's coverage, per the maker's pattern, includes the airspace above a floor register; the balloons are gone and the zone is quiet, but the detector is re-aimed so the register's warm-air column is outside its pattern (see related: How to Place and Aim a Motion Detector).

Steps 5 and 6. Recorded as environment, balloons drifting on register air, detector re-aimed. The account had one earlier dispatch 61 days before, a door left ajar, so this is the second in 90 days and the account is flagged; the owner calls the customer about adding a second-zone confirmation on that motion.

The other two tickets. The second was a front door (entry/exit) opened at 18:40 with no disarm inside the 30-second entry delay and a correct code entered 52 seconds after the door opened; the customer confirmed a sitter who arrived with full hands. Class user error, retrained by phone that day with the sitter on the line, and the delay was left alone because the customer did not want it longer. The third was a basement window zone that had logged 4 troubles in the previous week before alarming at 03:15; class device, visit booked for Wednesday.

Step 8. At 30 days, no further alarms on zone 4.

References

• See related: Alarm Permits, False Alarm Fines and the Dealer's Part
• See related: False Alarm Reduction Features in the SIA CP-01 Standard
• See related: Reading Event Codes on a Signal History
• See related: How to Train a Customer on Their Alarm System