Decision Velocity: When 70% of the Information Is Enough

In a stable market, waiting for one more report is usually cheap. In a volatile one, the information you are waiting for may arrive after the window to act has closed. Delay is itself a decision, and often an expensive one.

That does not mean faster is always better. The goal is to match the speed of a decision to its stakes and to how much time you really have. Two very different organizations, Amazon and the US Marine Corps, have written down unusually clear guidance on this, and much of it carries over to a small or midsize business.

This guide covers when to stop gathering information, how reversibility changes the answer, how to keep disagreement from stalling a team, and how to push decisions closer to the people who see what is happening.

Speed is part of decision quality

In his 2016 letter to Amazon shareholders, Jeff Bezos drew a distinction worth remembering: "Day 2 companies make high-quality decisions, but they make high-quality decisions slowly." He closed that part of the letter by asking whether readers had "settled only for decision quality" or were "mindful of decision velocity too."

The Marine Corps doctrine publication Warfighting (MCDP 1, 1997) makes a similar point: "Time is a critical factor in effective decisionmaking—often the most important factor." It adds that a key part of decision-making is realizing how much decision time is available and making the most of it.

The same doctrine is careful about the other side. In situations where time is not a limiting factor, such as deliberate planning, it says we "should not rush our decisions unnecessarily." Speed matters when the clock is real, not as a habit.

The 70 percent rule of thumb

Bezos offered a memorable heuristic: "most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90%, in most cases, you're probably being slow."

Treat that number as one executive's rule of thumb, not a measured threshold. The more important part is what he paired it with: you need to be good at quickly recognizing and correcting bad decisions. As he put it, "being wrong may be less costly than you think, whereas being slow is going to be expensive for sure."

Warfighting puts the same idea in terms of courage. Leaders must be willing to decide under uncertainty "when the natural inclination would be to postpone the decision pending more complete information." It continues: "We do not want to make rash decisions, but we must not squander opportunities while trying to gain more information." Because there is no perfect solution, the doctrine says, the task is to select a promising course of action with an acceptable degree of risk.

Sort by reversibility first

The 70 percent rule applies most cleanly to decisions you can undo. In the same 2016 letter, Bezos advised: "never use a one-size-fits-all decision-making process. Many decisions are reversible, two-way doors. Those decisions can use a light-weight process."

The UK Treasury's Green Book (2026), the government's guidance on appraising public projects, makes a related point from a planning angle: decision trees can help clarify alternatives where decisions are "either irrevocable or expensive to reverse." Hard-to-reverse choices earn more analysis.

A useful habit is to ask not only whether a decision is reversible, but whether you can make it more reversible. A pilot instead of a full rollout, a one-year contract instead of five, or a lease with an exit clause can turn a slow, high-stakes decision into a faster, lower-stakes one. For a deeper treatment, see our guide Real-Options Thinking: Small Bets That Keep Your Choices Open.

Ask what the next piece of information is worth

When a decision stalls, list the open questions and test each one:

1. Would the answer plausibly change what we choose?
2. How long will it take to get?
3. What does waiting cost in that time, in money, customers or options that expire?

If no remaining answer would change the choice, you already have enough information. If one would, set a firm date for getting it and decide then, with or without it.

For example, imagine a 12-person landscaping company deciding whether to buy a second truck before spring. The owner wants a firmer read on demand. But the only answer that would change the decision is whether two large commercial contracts renew, and those renewals are due in three weeks. The sensible move is to decide on the day the renewals come in, not to wait for a "clearer picture" that may never arrive.

Disagree and commit, and escalate early

Disagreement is one of the most common reasons decisions slow down. Bezos recommended the phrase "disagree and commit": a person states their view candidly, then commits fully to the direction the team takes. He was explicit that this applies to leaders too: "If you're the boss, you should do this too."

He drew a line between ordinary disagreement and real misalignment, where teams have different objectives. For those, he advised recognizing the issue early and escalating it immediately, because otherwise "the default dispute resolution mechanism for this scenario is exhaustion." Whoever has more stamina wins, which is slow and says nothing about who was right.

Push decisions down with clear intent

The fastest decisions are often made by the people closest to the problem. Warfighting describes "mission tactics": assigning a mission without specifying how it must be accomplished. Subordinates "inform the commander of what they have done, but they do not wait for permission."

What keeps that from turning into chaos is intent. The doctrine separates the task, what is to be done, from the intent, which explains why. "While a situation may change, making the task obsolete, the intent is more lasting and continues to guide our actions."

In a business, that might sound like: "Keep our top 20 customers supplied this month, even if it costs margin." A warehouse lead who understands that intent can reroute a shipment at 6 p.m. without waiting for a meeting.

Key takeaways

• Waiting is a decision with costs; weigh them explicitly.
• Bezos's 70 percent rule is a heuristic, and it works best when you are good at spotting and correcting mistakes.
• Use a lightweight process for reversible decisions, and look for ways to make big decisions more reversible.
• Stop gathering information once no remaining answer would change your choice.
• Use "disagree and commit" for ordinary disagreement, and escalate true misalignment early.
• Clear intent lets people closest to the problem act quickly and consistently.

Sources

• Jeffrey P. Bezos, 2016 Letter to Amazon Shareholders, Amazon (2017)
• US Marine Corps, MCDP 1 Warfighting (June 20, 1997)
• HM Treasury, The Green Book: Appraisal and Evaluation in Central Government (2026)