Fast-growing companies have a habit of slowing down as they scale, and one specific cause is treating every decision as if it were momentous. Jeff Bezos gave the antidote a memorable frame in his shareholder letters: some decisions are one-way doors you can't walk back through, and most are two-way doors you can. Confuse the two, and you either move recklessly on the dangerous ones or agonize over the harmless ones.
Key Takeaways
- Bezos splits decisions into Type 1 (one-way doors, irreversible) and Type 2 (two-way doors, reversible) (2015 shareholder letter).
- Type 2 decisions should be made fast, by people close to the work, and reversed if wrong.
- Applying heavyweight Type 1 process to Type 2 decisions is a major cause of organizational slowness (2016 letter).
- Aim to decide with about 70% of the information you wish you had.
The Two Kinds of Door
Bezos's framing, laid out in the 2015 and 2016 Amazon shareholder letters, is simple (2015 letter). A one-way door is a consequential, irreversible or nearly-irreversible decision: walk through, dislike what you find, and you can't get back. Those deserve slow, careful, heavily-consulted deliberation. A two-way door is reversible: if the choice turns out badly, you reopen the door and walk back, having lost only a little time. Most decisions, he argues, are two-way doors, and they should be made quickly by high-judgment individuals or small groups, not run through a committee.
The trap he warns about is directional: as organizations grow, they tend to apply the careful, one-way-door process to everything, including the reversible decisions. The result is slowness, excessive risk-aversion, and too little experimentation (2016 letter). Big companies rarely die from moving too fast on reversible calls; they ossify from moving too slowly on all of them.
| One-way door (Type 1) | Two-way door (Type 2) |
|---|---|
| Irreversible | Reversible |
| Decide slowly, carefully | Decide fast |
| Consult widely | Empower individuals |
| Rare | Most decisions |
Why the Distinction Speeds You Up
The point is to spend your deliberation where it matters rather than spreading it thin. Careful decision-making is expensive, in time, meetings, and momentum, so lavishing it on reversible choices is pure waste, and worse, it trains an organization to fear decisions. By explicitly labeling a decision as a two-way door, a leader gives the team permission to move: try it, watch what happens, and undo it if it's wrong. Bezos pairs this with the 70% rule, decide when you have about 70% of the information you wish you had, because waiting for 90% means you're almost always too slow (2016 letter).
A Concrete Version
An engineering team spends three weeks and four meetings debating which internal task-tracking tool to adopt, treating it like a permanent architectural commitment. It's a two-way door: if the tool is wrong, they can switch in a day, having lost nothing but a little setup. Meanwhile the actual one-way door that quarter, the shape of their core data model, which will be brutally expensive to change once customer data lives in it, gets decided almost casually in a single meeting. They inverted the effort: agonized over the reversible, rushed the irreversible. Naming which door each decision is would have flipped the time allocation correctly.
The Honest Counterpoint
The framework's failure mode is misclassification, and it's easy to talk yourself into "this is a two-way door" for a decision that quietly isn't. Some choices look reversible but have lock-in that only appears later: a database or vendor you can technically leave but practically can't once you've built around it, a public API others now depend on (Hyrum's Law again), a cultural norm that hardens once set. And "move fast on two-way doors" can be abused to skip due diligence on genuinely consequential calls. The discipline is honest classification: ask specifically what it would cost to reverse this a year out, rather than only today, and respect the answer.
What This Means for Teams
The two-way-door lens is one of the most valuable things a leader can teach a team, because it attacks decision paralysis at the root. Push reversible decisions down to the people closest to the work and tell them explicitly to move fast and be ready to undo, and reserve slow, careful process for the genuine one-way doors. That judgment about which decisions are reversible, and the confidence to move quickly on most of them, is a mark of experienced engineers and leaders, and it pairs with the error-budget mindset of making risk explicit and shared. See available engineers.
Frequently Asked Questions
What are one-way and two-way door decisions?
Bezos's framework: one-way doors (Type 1) are irreversible and deserve slow, careful deliberation; two-way doors (Type 2) are reversible and should be made quickly, since a wrong call costs little and can be undone.
Why does this make companies faster?
Because most decisions are reversible, and applying heavyweight, careful process to reversible decisions wastes time and breeds risk-aversion. Labeling a decision a two-way door gives the team permission to move and undo if needed.
What is the 70% rule?
Bezos's guideline to decide with about 70% of the information you wish you had. Waiting for 90% usually means you're too slow, and for reversible decisions the cost of being wrong is low anyway.
What's the risk of the framework?
Misclassification. Some decisions look reversible but have hidden lock-in (databases, vendors, public APIs, cultural norms). Ask honestly what reversing it would cost a year from now, rather than only today.
The Bottom Line
Most decisions are two-way doors you can walk back through, so make them fast and undo them if they're wrong, and reserve slow, careful deliberation for the rare one-way doors you truly can't reverse. Companies get slow by treating everything as irreversible. Classify honestly, push reversible calls down and out, and decide at 70% information rather than waiting for certainty that never comes.
Roberto Espinoza is CEO of Ruzora, which helps US startups hire pre-vetted senior LATAM engineers, with a vetted shortlist in 72 hours. See available engineers.
