Why the Hardest Part of Early Warning Is Getting Anyone to Listen

John Carmean

In early 2022, conditions in U.S. natural gas markets were already deteriorating. Climate patterns were shifting. Storage levels were diverging from seasonal norms. Supply-side indicators were moving in the wrong direction. By spring, the picture was getting worse across multiple fronts simultaneously.

Natural gas prices didn’t peak until June. Months of deteriorating conditions before the market moved.

Months is an eternity in energy trading. It is long enough to restructure procurement contracts. Long enough to adjust retail pricing before margins compress. Long enough to prepare fuel switching logistics across a generation fleet. Long enough to brief the board and document that leadership saw the risk and acted.

But months is also long enough for every person in the decision chain to talk themselves out of acting.

This is the real problem with early warning. The technology works. The data is available. What fails is the institutional willingness to act on a signal when the sky still looks clear. Blue sky paralysis.

There is a well-documented pattern in organizational behavior. When a warning arrives too early, it feels speculative. When it arrives too late, it feels obvious. The window where a warning feels both credible and actionable is remarkably narrow. Most organizations close that window themselves by requiring consensus before action.

The 2022 gas crisis hit large energy companies hard. Fuel costs spiked billions of dollars in a single year across the industry. Revenue growth was consumed. Margins inverted. The forward curves were already climbing. But acting on early indicators, months before the market consensus, requires someone in the room willing to move on a signal that the rest of the market hasn’t validated yet.

That is a leadership problem, not a technology problem.

The organizations that benefit from early warning are not the ones with the best models. They are the ones with the shortest distance between signal and decision. Where the trading desk can act on an alert without waiting for three committees to agree that the alert is real. Where the CFO can authorize defensive positioning based on documented, timestamped intelligence rather than gut instinct.

Specificity matters here. False positives destroy credibility. If a warning system cries wolf, the organization learns to ignore it. That is why validation rigor is not an academic exercise. It is the foundation of institutional trust. An early warning instrument that has been validated against real events, across unrelated domains, with zero false positives is not a suggestion. It is a reading. Whether anyone acts on it is a separate question.

I have learned that selling early warning is not about convincing people that the technology works. It is about finding the person in the organization who already knows the risk exists and is looking for the instrument that makes it visible to everyone else.

That person is usually not the CEO. It is the risk manager who cannot get the meeting. The trader who sees the position building but cannot quantify the exposure broadly enough to justify action. The operations lead who has been asking for systematic monitoring for years.

Early warning does not fail because the signal is weak. It fails because the signal arrives before the pain. And humans are not wired to act on pain they haven’t felt yet.

The organizations that solve this problem will not just avoid the next crisis. They will be the only ones in the room who can prove they saw it coming.