The Two-Way Door Reality Check
Most decisions that look reversible are one-way doors at your level — pause before you commit
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 3
- Confidence
- 90%
Bezos's 'two-way door' heuristic says reversible decisions should be made fast. Doshi's correction: a decision that is genuinely reversible at Bezos's level is often irreversible for a PM leader, because organizational commitments (shipped features, sales promises, roadmap debt) can't actually be walked back. So deliberately pause and think hard before committing, rather than fast-deciding your way into accumulated work.
Origin
Shreyas Doshi's counter to the widely-cited Jeff Bezos two-way-door / one-way-door decision framing; Doshi notes the same idea was earlier framed as 'reversible and irreversible decisions' by a blog author whose version didn't catch on.
Core principles
- 01A door that is two-way at Bezos's level is often one-way at yours
- 02Fast decisions on shipped features accumulate into roadmap debt you can't unwind
- 03Thinking is cheap — do more thinking, not less
- 04The perceived reversibility of a decision is relative to your organizational power to reverse it
How to run it
- 1
Name the decision as a door
When someone frames a product decision as a quick 'two-way door,' stop and ask whether it is actually reversible for you and your team in practice.
Pro tip Watch for authority bias — 'Bezos said it' makes people skip the analysis.
- 2
Trace the real commitment
Walk the decision forward: the five-to-six weeks to build, the ramp, the QBR where you must defend adoption, the sales pressure, and the 'table stakes' follow-on work you'll sign up for.
Watch out Once someone says 'table stakes' in a QBR, your only move is to commit even more engineers — the door has closed.
- 3
Deliberately pause before committing
Give the decision two minutes, two days, or two weeks of real thinking — clarifying customer motivation, differentiation, and distribution — before green-lighting the build.
Pro tip Explicitly think through customer motivation, differentiation, and a distribution approach for the feature before saying yes.
Watch out Deciding based on 'is Bob the engineer free?' is how most teams actually decide — and how they accumulate debt.
In the wild
A team fast-decides to build a feature as a 'two-way door.' It ships after weeks of work, shows poor adoption at the QBR, gets defended with a favorable anecdote, then draws sales complaints and a 'table stakes' demand — forcing the team to allocate even more engineers to a feature they shouldn't have built.
→ Instead of a reversible experiment, the team signed up for compounding work — the core reason product leaders stay chronically busy.
Common mistakes
Falling for the catchy metaphor
Doshi points out the only reason 'two-way door' beat 'reversible/irreversible decisions' is the catchier metaphor plus authority bias — not superior logic.
Using data and anecdotes selectively
Defending a low-adoption feature with a single customer's 'life-changing' anecdote hides the failure and delays the real reckoning.
Is it for you?
Best for
PM leaders under pressure to make fast feature calls in stakeholder meetings
Not ideal for
Genuinely low-stakes, truly reversible choices where speed matters more than deliberation
From the transcript
“most doors that look like two-way doors are actually one-way doors they are two-way doors at bezos's level but as a PM leader for you…”
“sometimes it is useful to pause for 2 minutes or two days days or two weeks before making that decision”
“thinking is cheap so you should do more thinking not less”
“now you have signed up for even more work for a feature you should not have built in the first place”
From the episode
4 questions Shreyas Doshi wishes he’d asked himself sooner
Former PM leader at Stripe, Twitter, Google