What Your Building Data
Actually Tells You

Data is only useful if you know how to read it — and equally important, if you know what it can't tell you. Building monitoring gives you a continuous record of your building's behavior. What you do with that record depends on asking the right questions of it.

This is where a lot of monitoring projects either deliver real value or quietly disappoint. The hardware works, the data flows, but nobody's quite sure what they're looking at — or they expect the data to answer questions it was never designed to answer. Understanding the honest capabilities and real limits of building data sets you up to use it well.

What data does well

It shows you what happened, and when. This sounds basic, but it's genuinely powerful. A timestamped time series tells you exactly when consumption spiked, when a temperature dropped, when a pump stopped cycling. You're no longer working from memory or guesswork — you have a record.

It establishes what normal looks like. Once you have a few weeks of data, patterns emerge. You can see what typical Monday morning startup looks like, what overnight consumption should be, what temperature a tank holds during steady-state operation. That baseline is what makes anomalies visible — because you now have something to compare against.

It surfaces problems early. Many building faults don't announce themselves dramatically. They show up as gradual drift — a supply temperature that's slowly declining, a pump runtime that's slowly increasing, an overnight baseline that's quietly higher than it used to be. Data catches this long before it becomes a service call or a failure.

It supports verification. When you make a change — a retrofit, a setpoint adjustment, a new piece of equipment — data lets you verify whether it actually worked. Before-and-after comparison isn't possible without a reliable before. This is the foundation of measurement and verification, and it's what makes a savings claim credible rather than estimated.

It creates an audit trail. For compliance, reporting, and accountability, timestamped data is far more defensible than manual logs or estimates. Whether you're demonstrating regulatory compliance, supporting a grant application, or resolving a tenant dispute, a continuous data record is evidence.

What data can't do on its own

It can't tell you why. Data shows you what happened. It rarely explains why. A temperature spike at 3am is visible in the data — but whether it was caused by a faulty valve, a control system error, a contractor working overnight, or something else entirely requires investigation. Data is the starting point for diagnosis, not the diagnosis itself.

It can't see what you didn't instrument. A monitoring system measures exactly what its sensors measure, and nothing else. If you have temperature sensors on the supply line but not the return, you're seeing half the picture of heat transfer. If you're monitoring electricity but not water, a leak won't show up. The completeness of your data is bounded by the completeness of your sensor placement — which is why thinking through what you actually need to measure before you deploy matters.

It can't act. Monitoring systems observe and record. They don't close valves, adjust setpoints, or dispatch technicians. Data can tell you that something needs attention; acting on that information is still a human responsibility. This is obvious in principle but easy to overlook in practice — especially when data volumes are high and nobody has a clear process for reviewing it regularly.

It's only as good as the sensor doing the measuring. A temperature sensor installed in a dead leg of pipe isn't measuring what you think it's measuring. A flow sensor sized for the wrong flow range will produce readings that look plausible but aren't accurate. Good data starts with correct sensor selection and correct installation — and no amount of analysis downstream fixes a problem at the source.

At a glance

The practical implication

None of the limits above are reasons not to monitor. They're reasons to go in with clear expectations — and to pair good monitoring hardware with a process for actually using what it produces.

The most effective monitoring projects share a few things: they were designed around specific questions, the sensors were placed to answer those questions accurately, and someone reviews the data regularly enough to act on what it shows. The hardware is the easy part. The discipline of using it well is what separates buildings that improve from buildings that just have more data.

If you're not sure whether your monitoring plan is set up to answer the questions you actually have, that's worth working through before you deploy. Getting the design right at the start costs far less than retrofitting it later. The best place to begin is with a clear goal — and that's exactly what the next article covers: How to Define a Monitoring Goal Before You Buy Anything →