# One-Way vs Two-Way Door Decisions: Reversible vs Irreversible

> Two-way door decisions can be undone; one-way doors cannot. Learn Bezos's Type 1 vs Type 2 rule and how to match your decision speed to reversibility.

Canonical: https://mentarium.io/mental-models/reversible-decisions.md

Mental model

6 min read By [Nemo](https://mentarium.io/about) Updated October 30, 2026

> **Reversible vs irreversible decisions** is the idea that how carefully you decide should depend on how easily you can undo the choice. Jeff Bezos called them two-way doors (reversible, decide fast) and one-way doors (irreversible, decide slowly) in his 2015 shareholder letter. Most decisions are two-way doors treated like one-way doors.

- **Category**: Mental model (decision speed and risk)
- **Origin**: Jeff Bezos, 2015 Amazon shareholder letter
- **Best used for**: Deciding how much time and analysis a choice deserves
- **Related**: Regret minimization, opportunity cost, margin of safety

## Ask yourself

- If this turns out wrong, can I undo it, and what would undoing it cost?
- Am I giving a two-way door the analysis a one-way door needs?
- Is there a smaller, reversible version of this decision I can try first?
- What part of this decision is truly permanent?
- What is the cost of deciding slowly here?

## What are reversible and irreversible decisions?

A reversible decision can be undone at low cost if it turns out wrong. An irreversible one cannot. The model says the amount of time, data and deliberation a decision gets should depend mostly on which kind it is.

Jeff Bezos gave the idea its popular image: doors. A two-way door is a decision you can walk back through if you do not like what you find. A one-way door locks behind you. One-way doors deserve slow, careful thought. Two-way doors deserve speed, because the cost of being wrong is small and the cost of being slow adds up.

The common failure is not mixing up which is which in theory. It is treating almost everything as a one-way door in practice, out of caution, and paying for it in lost time and missed chances.

## Where it comes from

The framing comes from Jeff Bezos’s 2015 letter to Amazon shareholders. He described “Type 1” decisions as consequential and irreversible or nearly irreversible, “one-way doors,” which should be made methodically and slowly with careful consultation. “Type 2” decisions, he wrote, are changeable and reversible, “two-way doors,” and can and should be made quickly by individuals or small groups with high judgment.

His central warning was about scale: as organizations grow, they tend to use the heavy Type 1 process for most decisions, including Type 2 ones. The result is slowness, unthoughtful risk aversion and less invention. In his 2016 letter he added a related rule of thumb: most decisions should probably be made with somewhere around 70% of the information you wish you had.

The underlying idea is older. Economists and decision theorists have long noted that irreversibility raises the value of waiting for information, and that when you can reverse cheaply, acting and learning beats analyzing.

## How to use reversible vs irreversible decisions

1. **Ask the door question.** If this turns out wrong, can I undo it? What would it cost in money, time, relationships and reputation? Be honest: many decisions that feel permanent are not.
2. **Score it.** Rate Reversibility from 1 (easy to undo) to 3 (permanent). The [Decision Weight Score](https://mentarium.io/tools/decision-weight-score) combines this with Impact and Uncertainty to tell you how much process the decision deserves.
3. **For two-way doors, set a short deadline.** Decide today or this week, with the information you have. Write down what would make you reverse it.
4. **For one-way doors, slow down on purpose.** Gather input, run a pre-mortem, look for the [opportunity cost](https://mentarium.io/mental-models/opportunity-cost), and sleep on it.
5. **Look for a two-way door inside the one-way door.** Can you test a reversible version first? A pilot, a rental, a trial period, a part-time start. Often the permanent decision can wait until a cheap experiment has answered the key question.

 The Mentarium Decision Journal

Write the decision down. Review it later. Get calibrated.

An Obsidian vault built on the Decision Loop: a guided template, a review queue for decisions that are due, and a calibration dashboard that shows how often you're right at 70%, 80%, 90%.

 [See Decision Journal · $29](https://mentarium.io/shop/decision-journal)

## 3 real-world examples

### Business

Say a startup spends three weeks debating which analytics tool to buy. The tool has monthly billing and can be swapped later: a two-way door getting one-way-door treatment. Meanwhile, the same team signs a three-year office lease in an afternoon because a broker said it would go fast. That lease is close to a one-way door. Reversing the attention, a day on the tool and three weeks on the lease, would have been better. (A hypothetical example.)

### Personal life

Quitting a job feels like a one-way door, and sometimes it is. But it often has a reversible version: ask for unpaid leave, take a sabbatical, or start the new path on evenings first. Moving to a new city can start as a three-month rental. If you combine this with the [regret minimization framework](https://mentarium.io/mental-models/regret-minimization-framework), you can test the choice you would regret not making, without betting everything on the first try.

### Investing

Buying a liquid index fund is largely a two-way door: you can sell tomorrow (with some cost and tax). Investing in a private company, a property or anything with a lock-up period is much closer to a one-way door. The same amount of money deserves very different levels of diligence depending on how easily you can get out. A [margin of safety](https://mentarium.io/mental-models/margin-of-safety) matters most where the exit is narrow.

## Common mistakes

- **Treating every decision as a one-way door.** This is the most common mistake and the one Bezos warned about. Caution feels safe, but delay has a cost too.
- **Missing the hidden one-way part.** A product change might be reversible in code but not in customer trust. Check every kind of cost, not only the obvious one.
- **Confusing reversible with free.** Two-way doors still cost something to walk back through. Fast is not the same as careless.
- **Never actually reversing.** People make a “reversible” choice and then stick with it out of [status quo bias](https://mentarium.io/biases/status-quo-bias). Set a date to review it, or the door is two-way only in theory.
- **Rushing a one-way door because a deadline was invented.** Artificial urgency (“this offer expires today”) is a classic way to push people through a one-way door.

## Two-way doors vs one-way doors

|  | Two-way door (Type 2) | One-way door (Type 1) |
|---|---|---|
| Reversibility | Easy or cheap to undo | Impossible or very costly to undo |
| Right speed | Fast, often same day | Slow, deliberate |
| Who decides | An individual or small team | Wider consultation |
| Information needed | Around 70% is usually enough | As much as is reasonably available |
| Main risk | Overthinking, slowness | Rushing, irreversible loss |
| Examples | Trying a tool, a pricing test, a new routine | Selling a company, a long lease, a major surgery |
| DWS Reversibility score | 1 | 3 |

## The Mentarium take

Reversibility is the first of the three inputs to the Decision Weight Score (DWS) for a reason: it is the quickest question you can ask and the one that sorts the most decisions. A choice that scores 1 on Reversibility cannot get a high DWS no matter how big it feels, which is the point. It deserves a quick decision, a written note, and a review date.

The habit we recommend is to log your two-way doors, not just your big decisions. A short entry in a [decision journal](https://mentarium.io/decision-making/decision-journal) (what you chose, what would make you reverse it, when you will check) turns fast decisions into fast learning. Over a few months you see which reversible bets paid off, and you get better at telling the real one-way doors from the ones that only feel like it. As part of a Model Stack for quitting a job, pair it with opportunity cost and regret minimization, and with the [10/10/10 rule](https://mentarium.io/mental-models/ten-ten-ten-rule) when the emotions are loud.

One more habit worth building: when someone brings you a decision, ask “one-way or two-way?” before you ask anything else. In teams, the answer changes who needs to be in the room and how long the discussion should run. A two-way door can be decided by the person closest to the work, today. A one-way door should be escalated, documented and slept on. Saying the label out loud also makes it easier to push back when a small, reversible choice starts collecting meetings it does not need. Over a quarter, that one question can give a team back more time than any productivity tool.

**Counters** [Status Quo Bias](https://mentarium.io/biases/status-quo-bias)

## FAQ

### What is a two-way door decision?

A two-way door decision is one you can reverse at low cost if it turns out wrong, like trying a new tool or changing a meeting schedule. Bezos argued these should be made quickly, by individuals or small groups, without heavy process.

### What is a one-way door decision?

A one-way door decision is irreversible or nearly so: once you walk through, you cannot get back to where you were. Selling a company, signing a long lease or having a child are examples. These deserve slow, careful deliberation.

### What are Type 1 and Type 2 decisions?

They are Bezos's names from the same 2015 letter. Type 1 decisions are the one-way doors: consequential and irreversible. Type 2 decisions are the two-way doors: changeable and reversible. His warning was that big organizations tend to treat Type 2 decisions with Type 1 process.

### How much information do you need for a two-way door decision?

Bezos wrote in his 2016 letter that most decisions should probably be made with around 70% of the information you wish you had. For reversible decisions, waiting for more certainty usually costs more than being wrong and correcting.

### How do I turn a one-way door into a two-way door?

Look for a smaller test: rent before you buy, run a pilot before a full launch, take a sabbatical before quitting. You often cannot make the full decision reversible, but you can make a reversible first step.

## Related models & biases

[Mental model Regret Minimization Framework The regret minimization framework is a decision method in which you imagine yourself near the end of your life, around age 80, and ask which option you would be least likely to regret. Associated with Jeff Bezos, it helps with big, personal choices where short-term fear hides long-term priorities.](https://mentarium.io/mental-models/regret-minimization-framework) [Mental model Opportunity Cost Opportunity cost is the value of the best alternative you give up when you make a choice. Every use of money, time or attention rules out other uses, so the real cost of anything is not just what you pay, but what that money, time or attention could have done in its next-best use.](https://mentarium.io/mental-models/opportunity-cost) [Mental model Margin of Safety A margin of safety is the buffer between what you expect to need and what you actually prepare for, so that a plan still works if your estimates are wrong. Investors buy well below estimated value; engineers build stronger than the expected load; anyone can apply it to time, money and health.](https://mentarium.io/mental-models/margin-of-safety) [Mental model The 10/10/10 Rule The 10/10/10 rule is a decision method from Suzy Welch: before choosing, ask how you will feel about each option 10 minutes from now, 10 months from now and 10 years from now. It separates short-term emotion from long-term consequence, so the feeling of the moment does not decide alone.](https://mentarium.io/mental-models/ten-ten-ten-rule) [Cognitive bias Status Quo Bias Status quo bias is the tendency to prefer the current state of affairs simply because it is current. People stick with defaults, existing plans and old habits more often than the merits justify, because change feels like a risk and staying put feels like no decision at all, even when it is one.](https://mentarium.io/biases/status-quo-bias)

 Mentarium Decision Coach · free

Use this with your AI

Connect the free Decision Coach to Claude or ChatGPT and it walks you through the Decision Loop, picks the right Model Stack and runs the Bias Radar on your reasoning.

 [Connect it free](https://mentarium.io/coach)

 Mentarium Decision Coach · free

Use this with your AI

 [Connect it free](https://mentarium.io/coach)