Plant rarely fails instantly. Most equipment degrades measurably for weeks first, which means the difference between an emergency and a planned repair is usually whether anyone was watching.
Every managing agent knows the pattern. The lift fails on a Friday evening. The booster pump stops on the hottest weekend of the year. The communal boiler dies in the first cold snap. In each case the first anybody knows is a resident's phone call, the engineer attends at emergency rates, and the repair takes longer because parts were not anticipated.
Plant failures are rarely instantaneous. Most equipment degrades measurably before it stops, which means the difference between an emergency and a scheduled repair is usually just whether anyone was watching.
Two distinct benefits, and being clear about them prevents overselling.
Knowing sooner. Even with no predictive capability at all, learning that the lift is out at eight in the evening rather than at eight the next morning is worth something. Residents can be informed, an engineer can be booked, and a vulnerable resident on the fifth floor can be checked on.
Knowing earlier. Trend data on temperature, vibration, run hours and power draw frequently shows equipment working progressively harder for weeks before it fails. That is the genuinely valuable part, and it converts an emergency call-out into a planned visit.
What it is not is a substitute for maintenance contracts, statutory inspection or competent engineers. Monitoring tells you something is changing; a qualified person still has to determine what and why.
The single most disruptive failure in a residential building, particularly where residents have limited mobility.
Machine room temperature is the highest-value single measurement. Lift controllers are heat-sensitive and a hot machine room in summer is a common cause of shutdown. A temperature sensor and an alert costs very little and catches a predictable seasonal failure.
Movement detection. A vibration or accelerometer sensor on the car or a door contact in the shaft, reporting whether the lift has moved in the last hour, tells you it is out of service without waiting for a call. Discuss placement with the lift contractor rather than fitting something yourself.
The emergency communication line. Lifts have an emergency phone or intercom that must work. Where the building's telephony has moved to internet-based service, these lines have failed silently in a great many buildings. Confirm yours works and consider a routine automated test — this is a genuine safety issue that monitoring addresses well.
Run hours and door cycles, where the lift contractor can provide them, are the basis of genuine predictive maintenance. Ask; some will share data from their own monitoring.
In taller buildings, booster pumps deliver water to upper floors, and their failure means no water above a certain level.
Power draw on each pump is the most informative measurement. A pump drawing more current for the same duty is wearing. A pump that is cycling far more often than its partner suggests a control or non-return valve problem. And zero draw when there should be flow is an unambiguous failure.
Pressure at the outlet, if a sensor can be fitted, is the outcome measure residents actually experience.
Plant room flooding. Pumps fail wet. A leak sensor on the floor of the pump room is trivially cheap and prevents a failure becoming an electrical incident.
Where a building has a communal boiler or district heat connection, its failure affects everybody simultaneously.
Flow and return temperatures on the primary circuit tell you whether the plant is doing its job, and a falling return temperature frequently indicates a circulation problem before residents notice lukewarm radiators.
Hot water storage temperature is both a service and a safety measurement, since bacterial growth control depends on it. In many jurisdictions logging this is a legal duty for a landlord or managing agent, and automated logging replaces a clipboard with a defensible record.
Pressure on sealed systems, since a slow loss indicates a leak somewhere in the building.
Mechanical extract in bathrooms and kitchens is frequently communal, frequently neglected, and its failure produces damp and mould complaints that are expensive to resolve and easy to attribute to residents' behaviour instead.
Monitoring whether the fan is actually running — through power draw or a simple airflow indication — distinguishes a system fault from a resident who does not open windows. That distinction matters, because remediating damp caused by failed ventilation is the landlord's responsibility and blaming a tenant for it is both wrong and, increasingly, legally risky.
Plant alerts are useless without a defined response, and managing agents typically operate with small teams and out-of-hours contractors.
Route to the on-call arrangement, not to an individual's phone. Whatever answers at three in the morning is what should receive a plant alarm.
Differentiate urgency clearly. A lift out of service overnight is not the same as a plant room flooding. Three tiers — informational, needs attention tomorrow, needs attention now — with different routing.
Give the contractor access to the data. A maintenance engineer who can see the trend before attending arrives with the right part. Several will value this considerably, and some will provide their own monitoring if asked.
Monitor the monitoring. A sensor that has stopped reporting in a plant room is itself a maintenance item. Silence must never be read as normality.
Plant monitoring costs money from the service charge and produces no visible benefit until something does not go wrong, which is a difficult thing to demonstrate.
The arguments that work are specific rather than general. Avoided emergency call-out rates, which are typically two to three times standard rates, and residents understand this immediately. Reduced outage duration, particularly for lifts, which is the failure residents care about most. And compliance evidence for hot water temperatures and emergency lighting, which protects everybody.
Reporting one avoided or shortened incident in an annual statement does more to justify the expenditure than any projection.
Item six is worth doing before item one. A surprising number of maintenance contractors already monitor the plant they service and will share it, which is faster and cheaper than duplicating their sensors.