The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →A two-way door decision is relatively easy to reverse; a one-way door decision is consequential and difficult—or costly—to undo. Engineering teams can use the distinction to match decision speed and scrutiny to the real risk: move quickly on bounded choices with a clear rollback path, and slow down when a choice commits customers, data, interfaces, teams, or long-lived infrastructure.
What do the two doors mean?
The metaphor, popularized in Amazon shareholder letters, describes how reversible a decision is—not whether it is important or whether the decision-maker is senior. Andy Jassy defined a two-way door in Amazon’s 2024 letter as a choice where, if it is wrong, “you can walk back through that door, revert to where you were, and there are few (if any) ramifications.” Amazon CEO Andy Jassy’s 2024 Letter to Shareholders
As an Amazon Associate I earn from qualifying purchases.
A one-way door is consequential and irreversible or nearly irreversible. Jeff Bezos’s 2015 shareholder letter says those decisions should be made “methodically, carefully, slowly, with great deliberation and consultation.” Amazon’s 2015 Letter to Shareholders The labels are a decision aid, not fixed categories: assess the actual undo path and consequences rather than deciding by the name of a project or technology.
Amazon materials have used different numbering conventions when associating the doors with Type 1 and Type 2 decisions. It is clearer to use the door names unless referring to a specific source’s convention.
#1 Best Overall
How do you tell whether an engineering choice is reversible?
Ask: If this is wrong, what would it take to get back? A code change may look easy to revert while the effects of its rollout—customer reliance, changed data, dependent services, or a public interface—are not. Look at the consequences of undoing the choice, not just the ease of reverting a commit.
| Consideration | More like a two-way door | More like a one-way door |
|---|---|---|
| Undo path | Can be rolled back or replaced promptly. | Requires a migration, major rework, or cannot restore the previous state. |
| Reach | Limited to a local experiment or bounded feature. | Commits multiple teams, customers, interfaces, or long-lived infrastructure. |
| Cost of error | Failure is contained and can provide useful feedback. | Failure could cause substantial customer, operational, financial, or compatibility consequences. |
| Decision process | A small accountable group can decide, observe, and adjust. | Evidence gathering and consultation with affected stakeholders are warranted. |
| Record | A lightweight note may be enough. | For a significant architecture choice, record the context, decision, and consequences in an ADR. |
These are practical comparison axes, not a formal scoring system. AWS gives an A/B test of a web or mobile feature as an example of a reversible decision and building a fulfillment center or data center as examples of high-capital choices that are hard to reverse. Those examples illustrate the idea; they do not classify every experiment or infrastructure decision automatically. AWS Executive Insights: Elements of Amazon’s Day 1 Culture
How should engineering teams choose and act?
For bounded choices with a clear exit
Keep the decision close to the people doing the work. Make a reasoned choice, define what you will observe, and set a point at which you will keep, adjust, or reverse it. A staged rollout or feature flag can, as an engineering practice, limit exposure and help preserve an exit path; it does not make a decision harmless if the rollout changes data or creates dependencies that cannot easily be undone.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteFor choices with broad or lasting consequences
Identify the assumptions, affected teams and customers, compatibility or data implications, and the cost of changing course. Gather relevant evidence and consult the people who will bear the consequences before committing. Bezos’s 2015 letter also warns against applying a heavyweight process to every decision: that can create slowness and risk aversion, and reduce experimentation. Amazon’s 2015 Letter to Shareholders The point is to scale the process to the choice, not to make every decision slow.
Rank #3
AWS’s “about 70% of the information you wish you had” guidance appears in its undated Day 1 culture article, which cautions against waiting for 90% or more when the delay would be too slow. This is Amazon/AWS management advice, not a measured engineering optimum or a threshold for high-consequence decisions. AWS Executive Insights: Elements of Amazon’s Day 1 Culture
When should an engineering team write an ADR?
An architecture decision record (ADR) is for a significant architecture choice. AWS Prescriptive Guidance describes ADRs as a way to capture the decision’s context, the decision itself, and its consequences. Examples of topics include system structure, non-functional requirements, dependencies, interfaces, and construction techniques. AWS Prescriptive Guidance: Using architectural decision records to streamline decision-making during development
Rank #4
For an accepted ADR, AWS’s process guidance says to preserve the record rather than rewrite history: if the decision changes, create a new ADR that supersedes the earlier one. That keeps the original reasoning and the later change visible. AWS Prescriptive Guidance: Architectural decision record process
- State the context. Describe the problem, constraints, relevant assumptions, and affected systems or teams.
- Record the decision. Say what the team chose and, where useful, the alternatives considered.
- Describe the consequences. Note trade-offs, operational effects, and how the choice can—or cannot—be changed later.
- Supersede, don’t silently replace. When an accepted decision changes, write a new ADR that links to and supersedes the old one.
Where does the metaphor stop being useful?
Reversibility is not binary. A choice can be technically reversible but still expensive in customer trust, data migration, service disruption, or coordination. Conversely, a large commitment may be made less risky by piloting it, narrowing its scope, or preserving an exit path. Those are engineering ways to change the practical cost of reversal, not guarantees that a difficult decision becomes easy to undo.
Jassy’s 2024 letter also notes that the metaphor assumes doors are unlocked and describes invention as opening doors that once seemed bolted shut. That is a leadership metaphor, not a promise that engineering constraints can always be removed. The useful question remains concrete: what can be undone, by whom, how long will it take, and what consequences remain afterward?
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
