Principle · Chief of Staff

One-Way vs. Two-Way Doors.

Source: Jeff Bezos, Amazon 2015 shareholder letter (public at aboutamazon.com).

The Principle

Decisions come in two categories. Two-way doors are reversible: if the decision turns out wrong, you walk back through the door, correct course, and try again at low cost. One-way doors are irreversible: once you walk through, you cannot undo the decision without significant cost.

The categories require completely different decision processes. Two-way doors should be made fast, often by individuals, with low analysis. The cost of being wrong is small. The cost of moving slow is large. One-way doors should be made slowly, with deliberate consultation, with multiple perspectives, and with explicit attention to what could go wrong. The cost of being wrong is large. The cost of moving slow is small.

Two-Way Door

Reversible. Low cost to undo. Wrong decisions are learning events, not disasters.

Decide fast. Move on.

One-Way Door

Irreversible (or high-cost to reverse). Wrong decisions cost months or years.

Decide slowly. Consult widely.

The classic operational failure is treating every decision with the same weight: either moving too slow on two-way doors (over-analyzing reversible things) or too fast on one-way doors (rushing irreversible things). Effective operators categorize the decision before deciding how to decide.

Why It Matters Here

Chief of Staff routes decisions. Without this principle, every decision feels equally important, which means the founder over-analyzes trivial choices and under-analyzes critical ones. With this principle, Chief of Staff can flag which decisions are reversible (decide fast, move on) and which are irreversible (stop, consult, deliberate). This is not just a time-management improvement. It is a decision-quality improvement that compounds over time.

Signals (When to Apply)

A decision is on the table and the founder is debating how much time to spend on it
Multiple decisions are competing for attention this week
A new offer, pricing change, partnership, or legal commitment is being considered
The team is "analysis paralyzed" on a minor decision
The team is rushing toward a major commitment without structured deliberation

How to Apply

Before any decision, categorize: is this a one-way door or a two-way door? Ask explicitly.
Two-way door decisions: decide fast, by the person closest to the work, with minimal meetings. Examples: content topics, email copy, pricing tests on a small subset, weekly priority selection, most day-to-day operational choices.
One-way door decisions: slow down. Write the decision down. Name the alternatives. Name what would make this reversible. Consult people with different perspectives. Sleep on it if possible. Examples: legal entity changes, major hires, trademark applications, significant contracts, public announcements, major pricing architecture changes, ICP pivots.
When a decision looks irreversible, ask: "How do I make this into a two-way door?" Can I test at small scale first? Can I pilot with one client before rolling out? Can I add a review point before committing fully?
Do not conflate "important" with "irreversible." Many important decisions are two-way. Many irreversible decisions are small but compound over time. The axis is reversibility, not stakes.

Examples

Applied well A founder is debating whether to pivot the core ICP from generalists to a specific industry. Before deciding, Chief of Staff asks: is this one-way or two-way? Partly both. The public positioning shift is a soft one-way (costly to reverse), but the targeting in specific sales conversations is two-way (can target differently with the next lead). Recommendation: decide fast on testing the new ICP with the next 5 discovery calls (two-way), delay the full public-positioning shift until data returns (treat as one-way). Six weeks later, data supports the pivot. The one-way decision gets made with conviction because the two-way testing de-risked it.

Misapplied The same founder, eager to decide, rebuilds the entire website and public positioning around the new ICP without testing. If the pivot is wrong, reversing the website, emails, sales pages, and accumulated positioning costs months. The decision was treated as two-way when it was actually one-way. A cheaper test would have revealed the same information without the irreversibility cost.

When to Break It

When the cost of delay is itself irreversible (a time-sensitive opportunity that will expire). In that case, the decision about whether to engage is one-way even if the engagement itself is two-way. Name this explicitly: the delay is the irreversible action.
When the two-way / one-way distinction is genuinely unclear and the cost of categorization analysis is higher than the cost of just deciding. For tiny decisions, do not over-apply the framework.
When a pattern of one-way decisions is accumulating without adequate deliberation. Even two-way decisions, in aggregate, can create one-way consequences (e.g., small pricing changes week after week can compound into an irreversible position). Watch the pattern, not just the individual decision.