Source: The illusion of certainty, by Rory Sutherland
at stake is the difference between deterministic and probabilistic improvement. If you engage engineers, you don’t know what you are going to get. You may be unlucky and get nothing. Or their solution may be so outlandish that it is hard to compare with other competing solutions. On average, though, what you get will be more valuable than the gains produced by some tedious restructuring enshrined in a fat PowerPoint deck.
But in business, let alone in government, it is only in crises that people find a budget for probabilistic interventions of this kind (in peacetime, nobody would have given Barnes Wallis the time of day). The reason is that both bureaucrats and business people are heavily attracted to the illusion of certainty. Standard cost-cutting ‘efficiencies’ can usually be ‘proven’ to work in advance; more interesting lines of enquiry come with career-threatening unknowability.
One problem with this pretense of certainty is that cost-savings are more easily quantified than potential gains
for a long time, the ratio between ‘explore’ and ‘exploit’ has been badly out of whack.
use [an] ‘evidence-based’ data-model up to a point, but correct for the fact that it is incomplete, temporary and weighted to the past. Institutionalised humans obtain a false sense of certainty by assuming … that what is optimal in a one-off transaction in a certain present is also optimal at scale, in an uncertain, long-term future.