Here’s a decision-making problem that shows up in almost every part of adult life, and is almost never taught explicitly.
You face two decisions this week. The first is whether to order the chicken or the fish for dinner. The second is whether to take a job offer that would move you and your family to another city for the next five years. How should you approach each?
A surprising number of people spend roughly the same amount of mental energy on both. They deliberate over the menu with the same gravity they bring to the job offer. They ask friends for advice about both. They google reviews for both. They feel equally anxious when the decision is finally upon them. The result is a kind of cognitive misallocation that is both exhausting and, on the job offer especially, often worse for the final choice.
The principle that separates these two decisions is easy to name but harder to apply: some decisions are reversible, and some are not. Treating them the same is one of the most common and costly patterns in how humans spend their thinking time.
The shareholder letter that made the distinction famous
In 1997, in his first letter to Amazon shareholders, Jeff Bezos introduced a distinction that would become central to how Amazon’s internal decision culture worked. He called them Type 1 and Type 2 decisions. Type 1 decisions are one-way doors — you walk through, and there’s no easy way back. Once you’ve acquired the company, sold the division, publicly announced the strategy, hired the senior executive, these are hard to undo. Type 2 decisions are two-way doors. You walk through, and if it turns out to be the wrong room, you walk back. Try the new meeting format; if it doesn’t work, revert. Experiment with the pricing; if customers hate it, change back.
Bezos’s observation, grown out of watching Amazon scale rapidly, was that most large organisations systematically apply Type 1 processes to Type 2 decisions. They treat reversible choices with the same caution, the same committees, the same analysis, the same sign-off chains, as they would irreversible ones. The result is an organisation that moves slowly even when nothing about the decision required slowness. It’s the enterprise equivalent of agonising over the chicken or the fish.
Applied outside organisations — to personal life, to careers, to relationships, to daily decisions — the same mismatch shows up. Much of adult paralysis is Type 1 thinking about Type 2 decisions.
The older tradition behind the distinction
Bezos’s framing was sharp and memorable, but the underlying insight has deep roots in decision theory.
The Nobel laureate Herbert Simon introduced the concept of bounded rationality in the 1950s. Simon’s argument, revolutionary at the time, was that humans don’t actually maximise expected utility across all possible options, because that would require infinite time and computation. We satisfice — search until we find an option that’s good enough, and commit. For most decisions, satisficing is the rational move, because the cost of continued searching (time, attention, opportunity foregone) usually exceeds any marginal improvement in the outcome. Simon’s framework implicitly assumed something important: most decisions are reversible or low-stakes enough that good-enough is better than exhaustive.
A separate tradition comes from the writer and risk theorist Nassim Taleb, whose book Antifragile introduced the concept of preserving optionality — making decisions in ways that keep future options open rather than foreclosing them. Taleb’s observation was that irreversible decisions, taken under uncertainty, can produce catastrophic outcomes that no amount of subsequent skill can undo. The strategy, therefore, is not to avoid all risk — that produces stagnation — but to distinguish sharply between risks that preserve optionality (you can recover) and risks that destroy it (you can’t).
A third comes from the military, where the concept of commander’s intent has long been trained explicitly. The principle: decide fast when mistakes can be fixed; decide slowly and carefully when they can’t. Junior officers are taught that hesitating over reversible tactical choices costs more than wrong tactical choices, because the wrong tactical choice can be corrected but the hesitation cannot be recovered. A wrong choice followed by rapid correction outperforms a perfect choice made too late.
These three traditions — economic, financial, and military — converge on the same practical principle from different directions. Reversibility is a first-order variable in how a decision should be made. Decisions with it get handled fast. Decisions without it get handled slowly.
How to tell which is which
The problem, of course, is that the distinction isn’t always obvious. Some decisions look irreversible but aren’t. Some look reversible but aren’t.
A job change, for example, is usually reversible. Most people who take a new job and hate it can find a different one within six months or a year. The downside is real but recoverable. A job change to another country with children enrolled in school, a partner whose work has to follow, and a house purchase bundled in — that version is much closer to irreversible, because unwinding it costs more than the job change itself.
A conversation with someone you’ve hurt is reversible if both of you are willing to keep talking; it’s irreversible if the other person stops taking your calls. A career pivot is reversible in your twenties, partially reversible in your thirties, mostly irreversible by your fifties — not because the world stops you, but because the cumulative investment in your current field becomes too costly to walk away from.
A useful heuristic: ask what the cost of reversing would be. If it’s days or weeks, the decision is clearly reversible; don’t overthink it. If it’s months, it’s substantially reversible; spend proportionate time. If it’s years, approach carefully. If it’s decades or never, this is a Type 1 decision, and the careful deliberation other people call excessive is probably warranted.
A second heuristic: ask who you’d be asking this of in a month. If the answer is nobody, I’ll have forgotten, it’s a Type 2 decision. If the answer is the same person I’m arguing with now, it’s Type 1.
The counter-view worth hearing
Not everyone thinks the reversibility framework is quite as clean as Bezos made it sound.
The brothers Chip and Dan Heath, in their book Decisive, raise an objection worth taking seriously. Their argument is that an excessive focus on reversibility can produce its own pathology: people who, because they’ve internalised “always move fast on reversible decisions”, never fully commit to anything. They skim along the surface of many choices, each made quickly, none lived with long enough to produce real learning. The person who switches jobs every eighteen months because the switch is theoretically reversible is technically right about the reversibility and wrong about what a career actually requires.
There’s also a more subtle point from the psychologist Barry Schwartz and his work on the paradox of choice. Schwartz found that people who keep decisions mentally reversible — who treat every commitment as provisional, revisable, under continuous review — are consistently less satisfied with what they’ve chosen than people who close off the option of revision once they’ve decided. The sandwich you could still return doesn’t taste quite as good as the sandwich you’ve decided to eat.
So the honest picture is this: use the reversible/irreversible distinction to allocate your deliberation time. But once you’ve made a reversible decision, consider living with it as if it weren’t. The reversibility is there as a safety net, not as a perpetual invitation to reopen the question.
The quieter version of this skill
The reversible/irreversible distinction isn’t only relevant for big life moves. It’s perhaps even more useful in ordinary life, where people waste enormous amounts of mental energy on the question of where to eat, what to wear, which tap to install, whose invitation to accept, what to say in the email.
Most of these are Type 2 decisions pretending to be Type 1. Nothing about the restaurant review, the email wording, the Saturday plans is going to cost you anything meaningful if you get it wrong. You will not be asked, on your deathbed, whether the restaurant on Friday was as good as the one two streets over. Treating these questions with the full weight of deliberation is a mild form of self-imposed tax on your life.
Conversely, some Type 1 decisions are quietly being made without the attention they deserve. Moving to a city because a partner already lives there. Accepting that a friendship is over because a single conversation went badly. Taking a first job that will end up defining the industry you spend twenty years in. These deserve more time than most people give them, because the reversibility isn’t really there.
The question that remains
The skill of life, in one sense, is knowing when to move fast and when to move slowly. The reversibility framework is the most reliable tool for making that distinction. It won’t save you from every bad decision. It will, at minimum, stop you from spending your decision-making energy in the wrong places.
The question to sit with, before the next decision that has you stuck, is a quiet one:
If you got this wrong, what would it cost to put right? For most of what you agonise over, the answer is: not much. For a few decisions, the answer is: more than you can pay. The difference between these two is almost everything.
Key research referenced: Jeff Bezos’s 1997 Amazon shareholder letter; Herbert Simon on bounded rationality and satisficing (1950s-onwards); Nassim Taleb, Antifragile (2012); the military concept of commander’s intent; Chip and Dan Heath, Decisive (2013); Barry Schwartz, The Paradox of Choice (2004).