The problem of warning has never been a shortage of information. After almost every strategic surprise, the signals turn out to have been present, buried in noise, split across desks, or dismissed because they contradicted what everyone already believed. Saturation has made that old problem worse, not different.
An old problem at new volume
The classic studies of surprise reached an uncomfortable conclusion decades ago. Roberta Wohlstetter's analysis of Pearl Harbor found that the signals of attack were available in advance, the failure was in separating them from the surrounding noise and assembling them into a picture anyone would act on. Cynthia Grabo, who spent a career in warning intelligence, distilled the discipline into a blunt principle: warning is not a product of information volume; it is a product of method indicators defined in advance, watched deliberately, and connected to someone with the authority to act.
Information saturation does not change this logic. It amplifies its cost. The modern environment adds three pressures the classical analysts did not face at today's scale: velocity (the interval between an indicator appearing and the event arriving keeps shrinking), volume (more channels than any team can read), and manufactured signal content engineered to look like an indicator, produced by actors who understand that decision-makers now drink from the same firehose as everyone else.
Why “monitor everything” fails
The intuitive response to saturation is broader collection: more feeds, more dashboards, more alerts. In practice this reliably degrades warning, for three reasons. First, attention is the scarce resource, not information, every additional feed dilutes it. Second, unfocused monitoring produces unfocused alerts, and unfocused alerts train recipients to ignore them; warning fatigue is a collection problem wearing an audience costume. Third, without pre-defined indicators, monitoring becomes retrospective: the feed explains events after they happen, which is journalism, not warning.
A warning function is not a bigger window. It is a smaller, better question, asked continuously.
The structure of a working warning function
Organizations without an intelligence agency behind them can still run genuine strategic warning. The structure has five parts.
- Warning problems. Name, in writing, the specific developments that must not surprise the organization, a supplier region destabilizing, a regulatory posture hardening, a campaign forming against the sector. If it is not named, it is not being watched; it is being hoped about.
- Indicators. For each problem, define observables that would move before the event: leading, not lagging. Good indicators are specific enough to be checkable and diagnostic enough that their movement actually changes the assessment. “Increased tension” is a mood; “permit suspensions in district X” is an indicator.
- Thresholds and responses. Decide in advance what amber and red look like for each indicator set, and critically what the organization does at each level. A threshold without a pre-agreed response is a decoration.
- Cadence and ownership. One named owner, a fixed review rhythm, and a short standard product: what moved, what it means, what we recommend. Warning that arrives as an occasional essay is not warning.
- A customer. The oldest failure in the discipline: warning produced for nobody in particular. Every warning problem needs a decision-maker who has agreed, in advance, that this is their call to make when the threshold trips.
Defending the function against itself
Two internal threats deserve standing attention. The first is confirmation drift: indicator sets quietly reshaped to support the house view. The antidote is periodic red-teaming of the framework itself asking what evidence would indicate we are wrong, and whether anything on the list could detect it. The second is the cry-wolf cycle: thresholds set so sensitively that ambers become routine and reds become negotiable. Calibration reviews comparing what the framework said against what happened are unglamorous and indispensable.
Manufactured signal requires one further habit: provenance before weight. In a saturated environment, who is showing me this, and why now is part of the indicator's value, not a separate question. Content that arrives pre-packaged as urgent warning deserves more skepticism than less.
The honest promise of warning
Strategic warning cannot promise prediction. It promises something more defensible: that the organization decided in advance what would matter, watched it deliberately, noticed movement earlier than an unstructured competitor would, and had already agreed what to do about it. In an environment where everything looks like a signal, that discipline not access to more information is the advantage.