How do I know what to start measuring?

It’s one of the questions we hear most from utility companies. Not “should we collect data?” — most organizations are already convinced of the answer. It’s the next question that stalls progress: where do we start?

Many utility companies recognize the dilemma: either you measure everything you technically can, without a clear sense of why. Or you wait until a strategy is in place and end up doing nothing at all. Neither approach creates value.

There’s a simpler way to think about it.

The fundamental question isn’t “what can we measure?” but “where does a failure cause the most damage?”. Risk should drive where you focus your first efforts, and risk breaks down into two components: the likelihood of a failure occurring, and the consequence if it does.

The consequences of a failure at a substation can range from a brief outage to serious personal injury, environmental impact, or extended disruptions to customers and critical infrastructure. Likelihood can be determined through historical failure data, learnings from peer utilities, inspections, or technical assumptions about age and load.

The combination of the two is called asset risk and gives you a prioritization list. That’s where you start measuring.

A natural starting point is to treat all assets equally — but in practice, the distribution is uneven. A small number of high-risk assets require rigorous and continuous condition monitoring, while the vast majority can be managed with simpler and more cost-effective methods.

Take power transformers as an example. An insulation failure there carries significant fire risk and long lead times for replacement. That justifies detailed gas and oil analysis, even at considerable cost. Compare that to inspecting cable cabinets or secondary substations, where failure consequences are more limited and simpler methods go a long way.

It’s not that some assets don’t matter. It’s about matching the level of effort to what’s actually at stake.

Once you know which assets are priorities, the next question is: are we measuring the right thing? For a measurement to deliver real value, it needs to capture the specific failure mechanism you’re trying to detect.

A good first step is to take stock of the data you already have: inspection records, maintenance history, logged readings, and failure reports. There’s often more to work with in existing systems than organizations have had time to leverage, and it provides a valuable starting point for identifying where the gaps are.

Once that picture is clear, the next step is setting up continuous data collection. That’s what makes analysis dynamic and valuable over time — and it’s when questions around communication, storage, and analysis become critical. The entire chain determines whether your measurement program actually creates value.

A measurement is only justified if its cost — including communication, storage, and analysis — is proportional to the risk it helps minimize. That sounds obvious, but it’s easy to lose sight of when evaluating technical solutions.

Frequency matters too. Does data need to come in every second to be useful, or is monthly sufficient? The answer affects both cost and which solution is appropriate. A high-cost real-time system on a low-risk asset is rarely justifiable.

But there’s another side to it. If the corrective actions are affordable and effective once a failure is detected, even a more expensive measurement solution may be worth the investment. It’s about putting cost in context.

The framework is straightforward, but putting it into practice requires a factual foundation. That’s why the best first step is often to map what you actually know about your assets today: which components carry high asset risk, where is failure history missing, and what data already exists but isn’t being used?

That mapping usually reveals the most obvious gaps quickly. And that’s a reasonable place to start — with a focused scope, a clear goal, and a measurable outcome.

You don’t need a large strategy or a large budget to get started. You need a clear question to answer.

Gomero specializes in data collection and predictive maintenance for electricity grid companies. We help you move from question to insight, and from insight to data-driven decisions. Reach out if you’d like to discuss where your organization stands today and where you want to go.