The analysis that will not converge
The market-entry case has been rebuilt three times. The optimistic scenario shows a business worth ten times the pessimistic one, and every rework moves the midpoint without narrowing the spread. The team is asked for "a firmer view" and produces the same fog with more decimal places. Meanwhile a competitor has announced something ambiguous, the board wants a recommendation, and the strategy director privately suspects that a fourth round of analysis will discover exactly what the third did.
Why organisations handle fog so badly
Corporate machinery is built for a world where analysis reduces uncertainty, so when uncertainty is irreducible, the machinery keeps running anyway. Planning templates demand single-point forecasts, so someone invents one. Confidence is rewarded in the room, so ranges get compressed into commitments as they travel upward. A number that survives three committees acquires an authority that has nothing to do with how it was produced. Presenting honest width ("somewhere between break-even and transformative") feels like weakness in front of a board trained on precision, so the width gets hidden, which is how organisations end up certain about things nobody ever actually knew.
The characteristic failure modes
- False precision: a guess dressed in decimal places, which then anchors every downstream decision.
- Paralysis dressed as rigour: analysis rounds that everyone senses will not converge, commissioned because commissioning them is safer than deciding.
- Narrative capture: when evidence cannot settle the question, the best storyteller wins, and the story's quality is unrelated to the outcome's.
- Full-scale commitment where a staged entry was available, because the plan was built to justify the investment rather than to learn.
- Bets sized by the upside case rather than by what the organisation survives if the downside lands.
Deciding inside the fog, not after it
The first discipline is sorting the uncertainty: separate what is knowable with work (customer willingness to pay, regulatory direction of travel, unit economics at pilot scale) from what is genuinely unknowable this decade, and stop paying for analysis of the second category. Then change the shape of the commitment rather than the quality of the forecast. Buy options: pilots, staged entry, reversible structures, partnerships that cap the downside, even where the optioned route costs more per unit than committing outright. Replace point forecasts with ranges and tripwires, agreed in advance: if the leading indicator has not crossed this line by this date, we exit, double down or renegotiate. Size the total exposure so that being wrong is survivable. And write down today's reasoning, because in three years the outcome will be visible and the fog will not be, and the organisation should judge the decision it actually faced.
What operators who have been in fog can tell you
People who have committed under comparable uncertainty carry something no model does: knowledge of which early signals turned out to matter, which confident assumptions collapsed first, and what they would stage differently now. Independent perspectives from selected senior operators cannot remove the uncertainty. What they change is the shape of the bet: which tripwires to set, where the survivable-loss line really sits, and which parts of the internal case are precision without knowledge.
