Ask a facilities director with 25 locations how their HVAC is doing right now, and the honest answer is: nobody knows. They know which stores called this week. They know which units got serviced last month. But the current condition of the other hundred-plus rooftop units is a mystery that resolves itself one phone call at a time.

This is not a staffing problem or a diligence problem. It is a structural one. A rooftop unit's condition is invisible from inside the building, most units fail gradually over weeks, and no organization can put eyes on hundreds of rooftops on any useful schedule. So the standard operating model for multi-site HVAC is: wait for the complaint.

The costs of that model are well documented. Roughly 70 percent of restaurant repair and maintenance spend happens reactively, and emergency repairs cost two to five times the same work done on a planned basis. Every one of those emergencies spent weeks as a detectable problem first.

Here are the actual options for fixing the visibility gap, with honest tradeoffs.

Option 1: More frequent preventive maintenance visits

The traditional answer. Quarterly or semi-annual PM visits per location, a technician on every roof a few times a year.

What it costs: Commercial PM contracts vary widely, but multiply any per-visit cost by locations and visits per year and the number gets large quickly. It also consumes technician hours during a national shortage of roughly 110,000 HVAC technicians, with about 30 percent of the current workforce over 55 and heading toward retirement. Service rates are projected to climb steeply over the next five years as that shortage bites.

The real limitation: A PM visit is a snapshot. A unit that passes inspection in March can begin failing in April and nobody will know until the store calls in June. PM visits are necessary, they are just not visibility. They tell you the unit's condition on 2 to 4 days per year and nothing about the other 361.

Option 2: Energy management and control platforms

Several established platforms serve the restaurant space with controllers, meters, and sensors at each location, providing scheduling, setpoint control, and energy management across the fleet. Many multi-site operators already run one, and for good reason: centralized control of schedules and setpoints delivers real savings, and the hardware generates a continuous stream of operating data from every unit.

What it costs: These platforms bundle hardware, software, and services, typically with multi-year commitments. The investment makes sense for operators who want the control and automation capabilities, which is why they are widely deployed.

The real limitation: Control is not the same as diagnostics. These platforms are built to manage how units are told to run: schedules, setpoints, demand response. Assessing how each unit is actually performing, whether a compressor is degrading, whether cooling capacity is sliding week over week, whether a unit is drifting toward failure, is a different analytical job, and it is not the core focus of a control platform. The encouraging part: if you already have one of these systems, you also already have the data stream that deeper equipment analytics need. That investment becomes the foundation for option 4 rather than a sunk cost.

Option 3: Thermostat portals you already have

If your locations run a modern energy management system or connected thermostats, you may already have a web portal showing current temperatures and setpoints across sites.

What it costs: Usually nothing extra, which is why it is worth mentioning.

The real limitation: Portals show current state, not health. A space temperature of 74 degrees looks fine on a dashboard even when the unit holding that temperature is running twice the hours it should, losing refrigerant, or one failed component away from a hot dining room. The portal answers "what is the temperature," not "is this unit okay," and by the time those two questions have different answers on the screen, the failure has already happened. Someone would also have to look at hundreds of zones every day, which nobody does.

Option 4: Analytics on the data you already generate

Here is what most multi-site operators do not realize: if your locations already have an energy management system or connected thermostats, those systems are already producing the data needed to assess unit health. Runtime, staging, supply and space temperatures, setpoints, and schedules are flowing every minute. The equipment records its own decline in detail. It just is not being analyzed.

Software-only monitoring connects to those existing systems through their interfaces, collects the readings continuously, and evaluates every unit's behavior against its own history and against comparable units: is the temperature drop across the coil declining, is the unit cycling abnormally, is runtime out of line with the weather, is performance degrading week over week.

What it costs: Because there is no hardware to buy, install, or maintain, this layer is priced per unit as software, and deployment across a fleet takes days rather than an installation project across every site. It adds to whatever systems are already in place rather than replacing them. Research from Lawrence Berkeley National Laboratory on fault detection deployments found software-only approaches to be the fastest-growing segment of the category, largely on those economics.

The real limitation: It requires that your sites have connected systems generating data. Locations running standalone manual thermostats have nothing to analyze, and those sites need connectivity first. Analytics also observe and diagnose, they do not fix. You still need service providers, you just direct them toward known problems on planned schedules instead of dispatching them toward emergencies.

What visibility actually changes

The point of fleet-wide visibility is not dashboards. It is that three operational conversations change:

Dispatch changes. Instead of "store 214 called, send someone today," it becomes "unit 3 at store 214 has been declining for two weeks, add it to Thursday's route." Same repair, planned rates instead of emergency rates, and the store never got hot.

Budgeting changes. Repair-or-replace decisions and capital planning stop running on technician anecdotes and start running on which units are actually declining fastest across the fleet.

Accountability changes. After a repair, the data shows whether performance actually recovered. After a coil cleaning, it shows whether efficiency improved. Vendors do better work when the results are measured.

Where to start

Whether you operate three locations or three hundred, if you have any form of connected thermostats or energy management system in place, the practical first step costs nothing: find out what data those systems can expose. Most operators discover they have been generating years of unit-level performance data that has never been looked at. The visibility problem, it turns out, was never a data problem. It was an analysis problem, and analysis is cheap now.

Elite Energy Management is an intelligence layer on top of the energy management systems multi-site operators already have. Every rooftop unit, every location, scored nightly, with no new hardware. See your fleet the way we see it.