The Art of Making Better Decisions Faster

Quick Answer: making better decisions faster: Speed and quality are not opposites — they come from the same three changes. Classify decisions as reversible or irreversible, so small ones stop consuming deliberation. Give each type a threshold, so you know in advance what evidence is enough. And run a pre-mortem on the important ones, so the failure case gets examined while it is still hypothetical. The result is not faster thinking; it is less thinking on the decisions that do not need it.

What this guide covers

• The reversible/irreversible split and how to classify a decision
• Setting thresholds for each class of decision
• Running a pre-mortem in twenty minutes
• Time-boxing without making snap judgements
• How to decide with incomplete information
• The discipline of deciding when the evidence says no

Speed and quality are not opposites

The assumption that faster decisions are worse decisions is the main thing keeping organisations slow. It is not true. The difference between a fast good decision and a slow bad one is not the thinking speed — it is the structure around the thinking.

Most slow decisions are slow because of three avoidable failures: the decision was never classified, so nobody knows how much deliberation it deserves; there was no threshold, so “enough information” is undefined and therefore never reached; and nobody wrote down the failure case, so the risks sat inside the argument where optimism could keep them comfortable.

Fix those three things and decision speed improves without decision quality falling. The mechanics below are deliberately plain, because a framework nobody can apply under pressure does not help.

Step 1: classify the decision

Two questions, and the answers determine everything else.

Is it reversible? Meaning: if it turns out badly, can it be undone within a short period at low cost? Renaming a product, changing a meeting time, adjusting a campaign’s creative — reversible. Hiring the first key person, entering a market, changing the pricing structure, taking on debt — not reversible.

Is it consequential? Meaning: does the outcome materially affect the business over months? A reversible decision can still be consequential, which is why the two questions are asked separately.

From these, three practical classes:

Type 1 — reversible and inconsequential. Decide in minutes. Do not schedule a meeting. The most common failure here is treating a naming decision like a hiring decision, which is a large part of why organisations feel slow.

Type 2 — reversible but consequential. Decide quickly, but with a real discussion and a written rationale. Price experiments, a new channel, a supplier change. These need evidence and a date for review, not a committee.

Type 3 — irreversible or highly consequential. Take the time. Gather evidence, run a pre-mortem, consult people who have seen the failure mode before, and then commit fully and stop reopening it.

The classification itself is worth writing down, because the moment it is written, most arguments about how much time to spend end.

Step 2: set the threshold in advance

A threshold is the evidence you will accept as sufficient, decided before you look. Without one, the amount of research needed to feel comfortable is undefined, so the decision never happens.

The form is simple and specific: “I will decide when I have X, Y and Z”. For a hiring decision, that might be three reference calls, a work sample, and a clear answer on the one skill the role actually depends on. For a channel decision, it might be 200 qualified enquiries at a defined conversion rate.

Two benefits. The first is speed, because the decision has an exit condition. The second is better, and less obvious: a threshold defined in advance is a commitment that you will not lower it when the answer is inconvenient. It removes the infinite possibility of finding one more piece of information, which is how decisions become procrastination while feeling like diligence.

The threshold should also name what will make you stop. “We will decide by Friday either way” is part of the threshold, because the most common cause of a decision not happening is that nobody ever agreed it would happen on a particular day.

Step 3: run a pre-mortem

The most reliable decision tool available, and the least used.

The normal review asks “what could go wrong with this plan?”. The problem is that the framing invites optimism, because asking about a plan you have already mentally committed to is uncomfortable. The pre-mortem removes the commitment: it asks, in the future tense, “It is twelve months from now and this decision was a disaster. Write the story of how it failed.”

Written and specific, this produces materially better risk analysis, because the failure case is something a group can imagine concretely whereas abstract risks stay abstract. And there is a second benefit, which is often larger: the discussion moves to what would have to be true for the plan to succeed, which produces the evidence to collect and the thresholds to set.

Twenty minutes, five people, written output, and the rule that every scenario must have a plausible mechanism rather than a generic catastrophe. “Competitors copy us” is not a scenario; “they copy our pricing in month four and our customers have a six-week switching window” is.

Step 4: time-box, without snap-judging

A deadline does not make a decision better. What it does is prevent the delay that hides the decision from you.

The mechanism worth knowing: when a decision is postponed repeatedly, the reasons for postponement stop being about evidence and start being about not wanting to face a particular answer. A deadline forces the trade-off into the open.

Two disciplines make time-boxing work. First, before the deadline, write down the two or three outcomes you would accept — that prevents the deadline from producing a decision by default rather than by choice. Second, if the deadline arrives with the evidence still short, the correct output is often a specific experiment rather than a decision: “we do not know enough to choose, so here is the smallest test that would tell us, and it runs for three weeks”.

That third option is routinely skipped, and skipping it is why organisations either guess or delay.

Deciding with incomplete information

Sometimes there is not enough information, and waiting for more is not a strategy. Four moves that work:

Decide on the irreversible parts and stay reversible on the rest. Most decisions can be decomposed. Commit to the market, defer the product; commit to the customer, defer the pricing model. Teams that wait for total certainty usually wait forever.

Look for the dominant option. If one choice is clearly better than the alternatives even under pessimistic assumptions, the remaining uncertainty does not change the decision. Note this is a different question from “which has the highest expected value” and is much more robust to incomplete information.

Ask what would have to be true for the opposite to be right. This converts a vague worry into a specific testable claim, often revealing that the concern is not actually a concern.

Set a review date rather than a permanent decision. “We will run this for a quarter and review on the 15th” is a legitimate decision, not a failure to decide. The important part is committing to the review in advance and not letting it quietly become permanent.

Deciding when the evidence says no

The hardest decisions are the ones where you have done the work and the answer is unfavourable. The obstacles are emotional rather than analytical: sunk cost, the discomfort of disappointing someone, and the tendency to keep generating options until one looks acceptable.

Two disciplines help. State the conclusion first, in one sentence, before discussing it. “This is not working and we should stop.” Leading with the conclusion changes the meeting from a search for support into a discussion of next steps.

Name the sunk cost explicitly, out loud. “We have already spent four months on this” is a fact about the past and has no bearing on the decision about the future. Saying so in words removes its pull.

And the general principle, which is also the habit of good leaders: a decision you have not publicly defended is barely a decision. Write down what you decided, why, what would change your mind, and when you will look again. That last item — what would change your mind — is what distinguishes a decision from a hope.

Related reading

• The Strategy Mistakes That Kill Growth
• The Power of Strategic Thinking
• How Successful Entrepreneurs Think Differently

Frequently asked questions

How can we make decisions faster without making them worse?

Separate the decisions before optimising. Reversible, inconsequential decisions should be made in minutes with no meeting. Reversible but consequential ones need a real discussion and a written rationale, with a review date. Irreversible or highly consequential ones deserve genuine time, a pre-mortem and full commitment afterwards. The slowness most teams experience comes from treating everything as the third category. Deciding the class first is what produces the speed.

What is a pre-mortem and how do we run one?

A twenty-minute exercise in which the group assumes the decision was a disaster twelve months later and writes the story of how it failed. It works better than asking “what could go wrong” because it removes the psychological commitment to the plan and forces the failure case to be concrete rather than abstract. Require every scenario to have a plausible mechanism, and end by asking what would have to be true for the plan to succeed — that produces the evidence to collect and the thresholds to set.

How do we decide when we do not have enough information?

Decompose the decision and commit only to the irreversible parts, keeping the rest reversible. Look for a dominant option — a choice that remains best even under pessimistic assumptions — which is far more robust than a precise expected-value calculation with incomplete data. Convert vague concerns into testable claims by asking what would have to be true for the opposite to be right. And where the evidence genuinely will not arrive, choose the smallest experiment that would produce it rather than guessing.

How do we stop postponing difficult decisions?

Set a threshold with a date attached before you begin: the evidence that will be sufficient, and the day the decision happens either way. Before the deadline, write down the outcomes you would accept, so the decision is a choice rather than a default. And if the deadline arrives with the evidence still short, make “run this test for three weeks” an explicit third option rather than pretending the evidence is sufficient.

How do we decide to stop a project we have invested in?

Say the sunk cost out loud — the months already spent are a fact about the past with no bearing on the decision about the future. State the conclusion in one sentence before the discussion, so the meeting becomes about next steps rather than about locating support. Then apply the threshold you would have applied at the start: on the original evidence, would you start this today? If not, the question is only whether the new information since then changes that, and it usually does not.

What does it mean to think in probabilities rather than certainties?

Holding a belief as a probability rather than a fact changes how you act on it. A 70 percent belief justifies a small, cheap test; a certainty justifies a large commitment. It also changes how you discuss it — “I think this is more likely than not, and here is what would change my mind” is a much more useful contribution than either confidence or hedging. The practical discipline is to attach a confidence level and a disconfirming condition to every significant view, which is what makes it possible to update without arguing about identity.

Written by SkilBrill Training Institute

November 5, 2026 · 11 min read