Remote Monitoring as a Service

Why this matters

Once your shop installs leak sensors, freeze alerts and connected equipment, selling monitoring looks like free recurring revenue: the devices already send alerts, so you just add your number to the list. The trap is in one word. Customers hear "we monitor" and believe "we respond." When a pipe bursts in a vacant house at 2 a.m. and your phone got the alert and nobody acted, the question at the claim is what you promised, not what you meant. This card covers what is worth watching, how to keep alerts from burying the one that matters, and how to write the response commitment so it says what your shop will actually do.

Monitoring and responding are two different products

Monitoring means alerts from the customer's devices reach your shop. Response means someone at your shop does a defined thing, within a defined time, during defined hours, when a defined alert arrives. A shop can sell the first without the second, but it has to say so in writing, and the customer has to keep receiving the alerts themselves.

The liability lives in the gap between the two. If the agreement, the website and the tech's sales pitch all say "monitoring" and nothing says what happens next, a customer and later their insurer will read it as the broadest response they can argue for. A one-line description of what you do, and what you do not do, is worth more than any monitoring feature.

What is worth watching

Not every alert a device can send deserves a person's attention. Sort each alert type by what it predicts and how fast that loss arrives:

The pattern in the right-hand column is the point. Only the top rows warrant waking someone. Everything else is either a delayed alert (it becomes worth acting on only if it persists) or a work queue the office clears in daylight.

Alert fatigue is the real failure mode

Alert fatigue is what happens when a person receives so many alerts that most turn out to be nothing, and they stop reading them closely. It is not a character flaw, it is arithmetic. The alert that matters arrives looking exactly like the forty that did not.

The cure is not fewer devices. It is two filters applied before an alert reaches a person:

• Persistence. A device that drops offline for four minutes while the router restarts is noise. The same device offline for 24 hours is a blind spot. Hold offline and power alerts until they persist past a set window. A common starting point: power loss held for 4 hours in mild weather and 1 hour when the outdoor forecast is below freezing; device offline held for 24 hours on non-critical sensors and 2 hours on a leak sensor or shutoff. Tune both to what your customers' homes can tolerate.
• Confirmation. A single temperature reading below threshold can be a sensor beside an open door. Require two readings at least 30 minutes apart, both below threshold. A common indoor freeze threshold for a vacant home is 45 F, which leaves hours of margin before pipes in exterior walls are at risk; some shops use 50 F for homes with plumbing in unheated crawlspaces.

Water-detected alerts get neither filter. A leak sensor that reports water is either right, in which case minutes matter, or wrong, in which case a call costs a few minutes. That asymmetry is why it goes straight to a person.

A worked month of alerts

A shop monitors 60 homes, most with three or four leak sensors, a shutoff valve and a temperature sensor. Over one winter month, the raw alert feed delivered:

• Device offline: 120
• Low battery: 45
• Temperature below 45 F (single reading): 12
• Power loss at the home: 8
• Water detected: 3

That is 188 raw alerts in a month, about 6 a day, landing on whoever's phone receives them.

Run the filters. Offline alerts that persisted past their window: 6 of 120. Temperature alerts confirmed by a second reading 30 minutes later: 2 of 12. Power losses lasting past the threshold: 1 of 8. Water detected: all 3 pass, unfiltered. Low batteries leave the live feed entirely and go to a weekly list.

What reaches a person as "act now" is 6 + 2 + 1 + 3 = 12 alerts in the month, against 188 raw, about 6 percent of the raw count. Twelve alerts a month is roughly three a week, which one on-call person can treat seriously every time. One hundred eighty-eight is a stream they will learn to swipe away.

What the filter cost. The 114 offline alerts dropped by the persistence window are not free. During each one the shop was blind to that home for up to its window. On a leak sensor that window is 2 hours, which is why the window is short there and long on a battery-powered temperature sensor in a heated home.

What would change the numbers. A cold snap moves the temperature row first: in a week below freezing, confirmed low-temperature alerts can jump several-fold, and the "act now" load with it. Staff for the worst week of winter, not the average month. A shop with many vacant homes (second homes, homes between owners) should expect the power and temperature rows to carry more weight, since nobody is home to notice the furnace stopped.

Writing the response commitment

A response commitment answers four questions in plain words, and an agreement that leaves any of them blank has left it for someone else to fill in later:

1. Which alerts. Name them: water detected, shutoff closed, confirmed low temperature, power loss past the threshold. Everything else is "reviewed during business hours."
2. Which hours. Either around the clock, or stated business hours with the plain statement that alerts outside those hours are reviewed the next business morning.
3. What the response is. Calling the customer and their listed contacts, in order, until someone answers. Dispatching a tech only with the customer's prior written permission and a way into the house. Closing the water remotely only if the customer authorized remote control in writing.
4. How fast. A common starting point for a shop offering overnight response: acknowledge a critical alert within 15 minutes, escalate to a second person if the first has not acknowledged within 15 more, first call to the customer within 30 minutes of the alert. For business-hours-only service: critical alerts reviewed within 1 hour of opening.

Never promise an outcome. "We will prevent water damage" is a promise no shop can keep, since a pipe can burst between a sensor and the valve or a valve can close and the water already on the floor still does damage. "We will call you and your listed contacts within 30 minutes of a water alert during covered hours" is a promise you can keep and prove you kept.

The customer keeps their own alerts. Write in that the customer remains subscribed to every alert from their devices. Your service is a second set of eyes, not the only set.

Where the law and insurance come in

This is not legal advice. Whether a limitation-of-liability clause in your monitoring agreement holds up is a question of state contract law, and states differ widely on how far a business can limit liability for its own negligence; some refuse to enforce a clause that tries to excuse gross negligence entirely, and what counts as "gross" is itself argued case by case. Do not copy an alarm company's agreement and assume its clauses travel to your state.

Two professionals to see before selling the service, and what to bring:

• Your attorney, with the draft agreement, your website copy, and the sales script your techs use. The website and the script count as part of what was promised, so they get reviewed together.
• Your insurance agent, asking whether your general liability policy covers a claim that arises from monitoring and response, or whether it treats monitoring as a professional service that needs separate errors-and-omissions cover. Many trade policies were written for work done at the site, not for a promise to watch a phone.

Pricing it so the response is sustainable

Price the response, not the monitoring. The software costs you little; a person available to answer a 2 a.m. alert within 15 minutes costs you an on-call rotation. Work backwards: the on-call pay for covering nights and weekends across a month, plus the expected overnight dispatches, divided by the homes covered, is the floor for an around-the-clock tier. If 60 homes cannot carry that floor, sell business-hours review only and say so clearly, rather than offering overnight response the shop cannot actually staff.

How to verify your service does what it says

• Fire a test alert monthly from one real device in a staff member's home or the shop, at an off-hours time, and time it from alert to first customer call. Record the minutes against the commitment.
• Audit the routing quarterly. Read the list of phones and emails each customer's alerts go to and confirm every one belongs to someone still on staff and still on the rotation. See related: The Alert Went to an Employee Who Left.
• Count the ratio. If more than about 1 in 10 of the "act now" alerts reaching a person turn out to be false alarms on the monthly review, the filters need adjusting before the on-call person starts ignoring them.

References

• See related: The Alert Went to an Employee Who Left
• See related: Setting Up Freeze and Power-Loss Alerts for a Vacant Home
• See related: Leak Sensors and Automatic Water Shutoffs
• See related: Battery-Powered Sensors and the Maintenance They Create
• Your state's contract law on limitation-of-liability clauses, reviewed with your attorney