Reversible vs Irreversible Decisions, With a Sorting Test

By·Founder, StandInLast updated: 4 min read
reversible vs irreversible decisionstwo-way door decisionsdecision classificationdecision documentationdecision rigor

The short version

  • A reversible decision can be undone cheaply and quickly. An irreversible one cannot be undone, or only at great cost.
  • Make reversible decisions fast, close to the work, with a one-line record.
  • Make irreversible decisions slowly, with the right decider and a full written record.
  • The most common mistake is treating every decision as irreversible, which slows the whole team down.

Jump to the full answer

A reversible decision is one you can undo quickly and cheaply if it turns out wrong. An irreversible decision is one you cannot undo, or can undo only at a high cost. The difference tells you how fast to decide, who should decide, and how much to write down.

Jeff Bezos popularized the idea in his 2015 letter to Amazon shareholders, as one-way and two-way doors. A two-way door can be walked back through. A one-way door cannot. Most decisions are two-way doors, and teams that treat them like one-way doors move slowly for no reason.

A four-question sorting test

Ask these four questions about any decision. If the answer to all four is "low" or "no," the decision is reversible.

QuestionReversible ifIrreversible if
What would it cost to undo next month?A few hours of workWeeks of work, money, or a migration
Does it change something customers or partners rely on?No, or only internallyYes: a public API, a price, a contract
Does it remove options later?NoYes: deleting data, ending a vendor, a public commitment
How many teams would have to change course?OneSeveral

Some decisions are partly reversible. Choosing a database is expensive to undo, but the schema inside it is cheap to change. Split those decisions into parts and treat each part on its own terms.

Who should make each kind

Reversible decisions belong with the people closest to the work. Waiting for a senior sign-off on a choice that can be undone next week costs more than the occasional wrong call. Give the team a standing rule: if it is reversible and inside your area, decide it.

Irreversible decisions need the person who owns the outcome, input from the people affected, and time to think. They are the decisions that justify an RFC, a review window and a written record. If your team is unsure who owns a given kind of decision, the decision authority map is the place to settle it.

How to document each kind

The documentation should match the stakes.

For a reversible decision, one line in the decision log is enough: what was decided, by whom, and a link to where it happened. The goal is that someone can find it later, not that it is defended in full.

An irreversible decision gets a full record. Write down the context, the options considered, the reasons, the consequences, and who has the authority to reopen it. In engineering, that record is often an ADR. For business decisions, it is a team decision record. Include one more field that reversible decisions do not need: what would have to be true for the team to revisit it.

The misclassification trap

Teams get this wrong in both directions, and each costs something different.

Treating reversible decisions as irreversible is the more common mistake. Every choice goes to a meeting. Engineers wait days for approval on things they could undo in an hour. The senior people who sign off become a bottleneck, and the team learns to stop deciding.

Treating irreversible decisions as reversible is rarer but more expensive. A team ships a public API change as a quick experiment, customers build on it, and now it cannot be taken back. Or someone deletes old data to save storage, and the data turns out to be needed for an audit.

A useful habit: when someone proposes a decision, ask out loud which kind it is. Saying it makes the team match the effort to the stakes.

What this looks like when someone is away

The reversible and irreversible split also tells a stand-in what they may decide while the owner is on leave. A stand-in can make reversible decisions in the owner's area. Irreversible ones wait for the owner unless waiting would cost more than deciding. Write that rule down before the owner leaves, and the backup does not have to guess.

For how this fits with the rest of a team's decision practice, see the guide to a system of record for decisions.

Where StandIn fits

StandIn answers your teammates' questions from the work you have already done, in your words, with a source under every answer. If you are away and someone asks whether a decision can be undone, your StandIn answers from what you wrote down, or says it does not know and points to who does. It never guesses, and it takes no actions on your behalf. See how StandIn works.

Written by

Founder, StandIn

Parker Miller founded StandIn. He writes about coverage, handoffs, and what it takes for software to answer for a team from what people actually wrote down.

More about StandIn

On your behalf. Never as you.

Every decision your team makes becomes something the next person can ask about, with the person who made it named as the source.

Keep reading

You might also like