Classify an engineering decision by how hard it would be to reverse in practice—and what happens if it is wrong. A high-consequence choice with no credible, affordable rollback is Type 1 and merits deliberate consultation. A choice with a workable correction path is Type 2 and can usually be made more quickly, with an owner and a clear rollback signal.
What Type 1 and Type 2 mean
Jeff Bezos introduced the distinction in his 2015 Amazon shareholder letter, using the image of one-way and two-way doors. The metaphor is about whether a team can return to its prior state, not whether a decision sounds important or technical.
- Type 1: A consequential decision that is irreversible or nearly irreversible. Bezos recommends making these decisions methodically, carefully, slowly, and with deliberation and consultation.
- Type 2: A decision that can be changed or reversed. Bezos says these can be made quickly by high-judgment individuals or small groups.
In his 2016 shareholder letter, Bezos reiterated that teams should not use one decision process for every choice and should be able to correct bad decisions promptly. These letters state a management principle; they do not establish a formal engineering standard or prove that faster decisions produce better engineering outcomes.
How to classify a specific engineering decision
Use these questions to assess practical reversibility. They are an engineering application of the letters’ distinction, not a checklist supplied or validated by Bezos.
#1 Best Overall
- Define the decision and its scope. State exactly what will change, which systems or teams are involved, and who may be affected. A broad “change the architecture” proposal is harder to assess than a specific choice about one service or interface.
- Describe the real rollback. Identify the steps needed to restore the previous state. Account for time, data recovery, compatibility, coordination across teams or services, and customer impact. A theoretical option to revert code is not enough if dependent systems or users have already adapted.
- Assess the downside of being wrong. Consider the likely consequences, including blast radius and any safety or regulatory exposure. Also consider the cost of delay: a choice can be difficult to reverse without every extra review step being worthwhile.
- Look for a smaller commitment. Ask whether a prototype, staged rollout, feature flag, or limited experiment can reduce the cost of changing course. Treat it as a genuine reversible path only if rollback is feasible and the consequences during the experiment are acceptable.
- Match the process to the risk. Use broader consultation and deliberate review when consequences are high and reversal is difficult. For a genuinely reversible choice, a responsible person or small group can usually decide with a lighter process.
- For a Type 2 choice, name the correction trigger. Specify what signal would prompt rollback or adjustment, who monitors it, and who has authority to act.
- Reassess when conditions change. New dependencies, external commitments, or emerging risks can turn an initially reversible choice into a harder-to-reverse one.
Why labels alone are not enough
Classify the actual commitment, not the name of the technology or decision. An API change may be easy to revert in source control but costly after external users have built against it. A database migration may have a theoretical reversal path that becomes risky once new writes have occurred. Conversely, a major architecture decision may be made easier to change through incremental rollout or a clean compatibility boundary.
These examples illustrate how to apply the reversibility principle; they are not examples or findings reported in the shareholder letters. A useful assessment asks not simply “Can we undo this?” but “How long would undoing it take, who would have to act, and what harm could occur before the prior state is restored?”
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How much review is proportionate?
Bezos argues that applying a heavyweight Type 1 process to reversible decisions can make an organization slow and inhibit experimentation. That is his management argument, not independent causal evidence about engineering teams. The practical aim is proportionality: reserve time-consuming consultation for choices whose consequences and rollback difficulty justify it, while ensuring faster choices still have an accountable owner and a correction path.
The letters provide no scoring formula, numeric thresholds, exhaustive catalog of engineering decisions, or engineering-specific proof of effectiveness. Treat Type 1 and Type 2 as a qualitative way to choose a decision process, not as a compliance label or a substitute for technical, safety, or regulatory review.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchQuick Recap
Rank #4
Rank #3
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.
