Fast or Right? Resolving the Decision Speed-Quality Tradeoff

Speed and quality aren't opposite ends of one dial — they're governed by two separate variables: how reversible the decision is, and how fast the information behind it is decaying. Calibrate for those two factors instead of guessing, and most decisions turn out to deserve far less deliberation than teams instinctively give them.

Treat every decision as either a two-way door (reversible — decide with roughly 70% of the information you wish you had) or a one-way door (hard to reverse — slow down and gather more). Most decisions your team agonizes over are two-way doors being run through one-way-door process.

Why Speed and Quality Aren't Actually Opposites

The tradeoff feels real because most organizations apply one process to every decision, regardless of what's actually at stake. Speed and quality only trade off when you're forced to buy certainty with time you don't have. In reality, a decision's reversibility usually matters more than its size.

A pricing-page headline test, a Slack channel structure, a backlog re-prioritization — these are all cheap to undo. Deciding on them slowly doesn't buy you a better outcome; it just buys you a slower feedback loop, because you learn the truth from the market either way, only later. Waiting for certainty is itself a decision, and it's usually the most expensive one on the table, because it's the only option that guarantees you learn nothing until the delay is already spent.

Contrast that with decisions that are genuinely hard to unwind:

• Committing to a platform migration that touches every downstream team
• Signing a multi-year vendor contract with steep exit costs
• Restructuring reporting lines or eliminating a role
• Setting a pricing model that resets customer expectations publicly

These deserve real deliberation — not because they're "bigger," necessarily, but because a wrong call compounds and can't be cheaply reversed. This distinction, not raw stakes, is the variable that should set your decision speed. It's a theme worth internalizing across every part of the job, and one we return to throughout our complete guide to PM leadership: the leaders who feel most "decisive" usually aren't faster thinkers — they've just gotten disciplined about sorting decisions into the right bucket before they start deliberating.

The Second Variable: Information Decay

Reversibility explains who should own the pace of a decision. Information decay explains why waiting past a certain point stops helping at all. Customer sentiment, competitor pricing, market timing, and team morale are all perishable — the data you gather about them has a shelf life, not a permanent value.

Treat "gather more information" as an investment with diminishing, then negative, returns. Early research resolves real uncertainty cheaply. Past a certain point, the environment you're studying has already shifted, so additional research measures a market that no longer exists by the time you act on it. A decision made confidently on data that's gone stale isn't actually higher quality than one made faster on fresher, slightly less complete data — it just feels safer.

The 70% Rule: How Jeff Bezos Decided When to Stop Gathering Information

Jeff Bezos's 70% rule holds that if you wait until you have 90% of the information you wish you had before deciding, you're moving too slowly. For most decisions, acting on roughly 70% of the information you'd like is the right threshold, because the marginal value of that last 20% is almost always smaller than the cost of the time it takes to gather it.

Bezos has described this logic across Amazon's shareholder letters and in his collected writings, Invent and Wander (2020), tying it directly to his broader "Day 1" philosophy: organizations that insist on near-certainty before moving eventually calcify into "Day 2" companies — process-heavy, slow, and outpaced by competitors willing to act on good-enough information. The rule isn't an argument for recklessness. It's an argument that the cost of the last 20% of certainty is usually higher than the cost of being occasionally wrong.

Three reasons that last 20% costs more than it returns:

1. Information decays. Market conditions, competitor moves, and customer needs shift while you research, so the "complete picture" you're chasing is often stale by the time you finish assembling it.
2. Analysis creates its own risk. Every extra week of research is a week your team isn't building, testing, or learning from something real — and a week competitors might spend shipping instead.
3. Certainty-seeking is often morale-seeking. Teams frequently keep researching not because the data is inconclusive, but because nobody wants to be blamed if the call is wrong — a comfort tax disguised as rigor.

This shows up constantly in discovery work. Teams run one more round of interviews through a Jobs to Be Done lens long after the underlying job is already clear, chasing a certainty that customer research was never going to fully deliver. At some point, more interviews stop reducing uncertainty and start just delaying the decision you already have enough evidence to make.

Reversibility Is the Hidden Variable: One-Way vs. Two-Way Doors

Bezos's companion concept, laid out in his 2015 shareholder letter, is the split between one-way doors and two-way doors. A one-way door is a decision you can't walk back once you've stepped through; a two-way door is one you can reopen and reverse at acceptable cost. Almost every decision-speed problem traces back to misclassifying which kind of door you're standing at.

The table's takeaway is simple: each framework answers a different question, and most teams only ever use one of them — usually none. Stacking all four gives you a full loop: classify the door, decide with 70% of the information, price the delay, and later audit whether the call itself was sound.

A one-line copy fix you spotted while walking a customer journey map is almost always a two-way door: ship it, watch the metric, revert if it doesn't move. A decision to sunset a product line customers depend on is a one-way door, and deserves the slower process. The mistake most teams make isn't picking the wrong process for a hard decision — it's defaulting every decision to the hard-decision process out of habit.

The Cost of Delay: Why Waiting Is Never Free

Delay is never neutral; it always has a price, even when nobody bothers to calculate it. Donald Reinertsen's CD3 metric — Cost of Delay divided by Duration — reframes prioritization around one question: given two initiatives, which one is more expensive to leave sitting in the queue? Reinertsen has argued that most organizations treat queues and delay as effectively free, when in economic terms they never are.

"He who has the fastest feedback loop wins" is a summary often drawn from Reinertsen's flow-based approach to product development — a useful gut-check whenever a decision stalls waiting for one more data point.

Nowhere is the cost of delay more visible, or more consistently ignored, than in the decisions teams avoid longest: managing a low performer out. Leaders wait for "one more quarter of data" long after the team has already reached its own verdict.

The cost of that delay isn't paid by the person avoiding the decision. It's paid by the rest of the team, whose trust in the manager's judgment erodes every week the call doesn't come. Our piece on managing someone out with dignity makes the case that delay itself, not directness, is usually the less humane path.

Daniel Kahneman's work on overconfidence and the planning fallacy, from Thinking, Fast and Slow, points at why this happens: people systematically overrate how much a bit more analysis will improve their judgment, and underrate how much time that analysis will actually consume. The fix isn't more willpower — it's a structural default that doesn't require willpower in the first place, which is the whole point of the audit below.

Retired Army general Stanley McChrystal makes a related argument in Team of Teams: centralizing every decision at the top for the sake of "consistency" often just relocates the cost of delay onto whichever team is waiting on an answer. His fix was to push authorization for reversible, front-line calls down to the people closest to the information, while reserving senior review for decisions with irreversible, organization-wide consequences — the same reversibility split, applied to who holds the authority to decide, not just how fast they use it.

Run a Decision-Latency Audit on Your Team

A decision-latency audit finds the low-stakes, easily-reversible decisions your team took the longest to make, then resets the default speed for that whole category. It's a one-hour exercise, not a quarterly initiative, and it typically surfaces more wasted time than any single process fix a team has tried before.

Run it in six steps:

1. Pull the last month of decisions from meeting notes, Slack threads, ticket comments, or a running decision log — anywhere a call actually got made and recorded.
2. Tag each one on two axes: stakes (low / medium / high) and reversibility (easy / hard to undo). Ignore how long it felt like it should take — just classify it.
3. Log the actual time-to-decide for each: hours, days, or weeks from "this needs a call" to "we did it."
4. Flag every low-stakes, easily-reversible decision that took longer than a day. These are your decision-latency offenders — the ones costing you the most for the least reason.
5. Ask the owner what they were waiting for. Usually it's one of three things: a missing stakeholder sign-off nobody actually required, more data that wouldn't have changed the call, or simple discomfort with being wrong in public.
6. Set a new default per quadrant and publish it to the team, so the next low-stakes call doesn't relitigate the same debate.

Most teams discover the bulk of their wasted decision time sits in the top-left quadrant — low-stakes, reversible calls being run through a process built for the bottom-right one. That's the fastest fix available, because it requires no new framework, only a new default.

Make Speed the Default for Reversible Calls

The single highest-leverage governance change most teams can make is a blanket default: two-way doors get a single owner and a same-day decision window; one-way doors get a named reviewer and a real deliberation window. Writing that split down, rather than leaving it implicit, is what actually changes behavior — an unwritten norm reverts to "ask everyone" the moment anyone feels nervous.

Which leadership style you bring to a given call matters here too. A pacesetting or commanding style suits a two-way door decided in an afternoon; a coaching or democratic style earns its cost on the rare one-way door where buy-in matters as much as the answer. Daniel Goleman's research on six leadership styles and when to use each is really an argument for matching your style to the decision's actual stakes, not applying one default style to everything you touch.

The catch is that "speed" and "rigor" both feel virtuous in the moment, so teams rarely self-correct without a way to check their own judgment against the record. This is exactly what a Decision Journal is built to do: log a decision when you make it, tag it by speed and by stakes, and revisit it later against what actually happened.

That review step is the part most teams skip entirely, and it's the only way to find out whether your instinct to slow down is actually protecting quality or just costing you time for nothing. For a deeper walkthrough of building that habit, see our guide to using a Decision Journal to calibrate your judgment.

Key Takeaways

• Reversibility, not size, should set decision speed — a "big" decision that's cheap to undo deserves less process than a small one that isn't.
• The 70% rule argues for deciding on roughly 70% of the information you wish you had, because the last 20% usually costs more time than it's worth.
• One-way doors deserve deliberation; two-way doors deserve a single owner and a same-day default — misclassifying which door you're at is the root cause of most decision-speed problems.
• Waiting is never free. Reinertsen's CD3 framing — cost of delay divided by duration — makes that cost visible instead of implicit.
• A decision-latency audit (find your slowest low-stakes calls, tag them, set new defaults) usually surfaces more wasted time than any single process change.
• Delayed personnel decisions are a textbook cost-of-delay case: the team usually reaches its verdict before the manager does, and every extra week erodes trust rather than protecting it.
• Reviewing past decisions against outcomes, not just logging them, is what actually calibrates judgment over time, whether that review happens in a notebook or a structured tool like a Decision Journal.