When the clock is the binding constraint, cut analysis before you cut evidence

Speed is a constraint, not a style

Most advice about deciding assumes the deliberation is free and only the outcome is at stake. Real decisions have a deadline attached, and past a certain point the cost of continuing to think exceeds whatever accuracy it buys. The interesting question isn’t whether to decide quickly but which components of a good decision degrade gracefully under pressure and which collapse.

They do not degrade equally. The parts that depend on gathering information hold up reasonably well if you narrow what you are gathering. The parts that depend on comparing many options against many criteria fall apart almost immediately, because that work scales badly and is the first thing a hurried mind abandons without noticing.

What actually happens to judgement when time is short

Under time pressure, people tend to search less, attend to fewer attributes, weight negative information more heavily, and rely more on whichever cue is most salient. This is a reasonably consistent picture across the literature on decision-making under deadline, and it’s not entirely a bug: attending to fewer things is the correct response when there is not time to attend to everything.

The failure is in which things get dropped. The attributes that survive are the ones that are easy to observe and easy to compare, not the ones that matter most, because ease of comparison determines what gets used when attention is scarce. A candidate’s presentation survives the cull; their record on a specific hard task does not, since checking that takes a phone call.

The triage order

A workable priority under pressure runs roughly like this. Keep the question — the single largest waste in a fast decision is answering a subtly different question from the one that needs answering. Keep one or two pieces of decisive evidence, chosen before you start looking, on the grounds that they would move the answer. Keep the check for irreversibility, because that’s the only property that determines how bad a mistake can be.

Then drop, in this order: exhaustive option generation, precise weighting of criteria, second opinions on details, and the write-up. These are the things you can add afterwards if the decision holds, and the last of them isn’t a decision input at all.

What should never be dropped is the explicit statement of what you are assuming. It takes one sentence, and it is the only artefact that lets you revisit the decision intelligently when the pressure lifts.

Fast is not always worse

There are conditions under which quick judgements are close to as good as slow ones. When the environment is regular, when the person has substantial experience with prompt feedback, and when the cues that predict the outcome are few and reliable, rapid pattern recognition performs well. Studies of expert judgement in fireground command and similar settings describe practitioners generating one workable option and testing it mentally rather than comparing several — and doing so effectively.

The conditions are the whole claim, though. Take away the regularity or the feedback and the same rapid process produces confident errors, because the pattern matching still runs and no longer corresponds to anything. Popular versions of this research that recommend trusting instinct in general leave out the part that made it work.

The decisions that should not be made fast

Some situations are dominated by the cost of a mistake rather than the cost of delay, and they can usually be identified in advance: the choice is hard to reverse, it commits resources you can’t re-earn quickly, or it creates obligations to other people. In those cases the correct move under time pressure is frequently to buy time — a partial commitment, a smaller version, an explicit extension — rather than to decide well quickly.

This is worth separating from the urge to delay out of discomfort. The test is whether waiting brings information. If nothing new will be known tomorrow, tomorrow’s decision is today’s decision with more anxiety attached, and the deadline is doing you a favour.

Deciding in advance is the real trick

Most of what makes a fast decision good is work done before the clock started: a known threshold, a rule about what would count as sufficient, a shortlist prepared while there was time. People who appear to decide quickly and well have usually pre-decided the general case and are recognising an instance of it.

That is the transferable lesson. You cannot think faster on demand, but you can move the thinking earlier, and any category of decision that recurs is a candidate. Write the rule while the pressure is off; under pressure, apply it and note the exception rather than rebuilding the reasoning from nothing.