Reversible vs Irreversible Decisions: A Practical Framework

The reversible vs irreversible decisions framework is a simple mental model that sorts every choice into two buckets: decisions you can undo if you're wrong, and decisions you can't. Amazon formalized this as Type 1 (irreversible, high-stakes) and Type 2 (reversible, low-stakes) decisions. The practical payoff is clear: you move fast on Type 2 and apply deliberate process only to Type 1.

Why Most Teams Get This Backwards

Most teams apply the same level of scrutiny to every decision. A pricing experiment gets the same four-week approval cycle as a company acquisition. A new onboarding email gets the same committee review as signing a three-year office lease. The result is that teams move slowly on things that deserve speed and occasionally move too fast on things that deserve careful thought.

The cost is real. Slow decisions on reversible items drain momentum. Engineers wait for feedback. Campaigns miss seasonal windows. Salespeople lose deals while waiting for sign-off. Meanwhile, irreversible decisions sometimes slip through because everyone assumes someone else did the analysis.

A structured framework fixes both problems at once.

Amazon's Type 1 and Type 2 Framework

Jeff Bezos introduced this framing in Amazon's roughly 2015 shareholder letter. The core idea is straightforward.

Type 1 decisions are one-way doors. Once you walk through, you can't easily come back. They tend to be high-cost, high-impact, and hard to reverse. Shutting down a product line, committing to a new market, changing your core pricing model, signing a long-term contract: these are Type 1.

Type 2 decisions are two-way doors. You can step back through them if the result isn't what you expected. They are lower-cost, lower-impact, or easy to adjust. Running a new ad campaign, changing a landing page headline, testing a new onboarding flow, adjusting a team's sprint process: these are Type 2.

Bezos's argument was that large organizations default to Type 1 process for everything, which slows them down on the majority of decisions that are actually Type 2. He wanted teams to move fast on Type 2 and reserve heavyweight process for the small percentage of decisions that are genuinely irreversible.

For a small business or startup, this framing is even more valuable. You don't have the organizational slack to run slow processes everywhere. Speed on reversible decisions is a genuine competitive advantage.

The Two-Question Filter

You don't need a complicated scoring rubric. Ask two questions in order.

Question 1: If this goes wrong, can we undo it within 90 days without catastrophic cost?

If yes, it's a Type 2 decision. Approve it fast, set a review date, and move on.

If no, move to question 2.

Question 2: Are the costs of a mistake disproportionate to the upside?

If yes, it's a Type 1 decision. Slow down, gather more information, and apply structured analysis before committing.

If no (the downside is bounded even if it's not reversible), treat it as a Type 2 with slightly more scrutiny.

The 90-day window is not arbitrary. It's short enough to be honest about whether something is truly reversible, and long enough to capture the real cost of unwinding. A lease you can break with 60 days' notice and a $5,000 penalty is probably Type 2. A three-year lease with a personal guarantee is Type 1.

How to Apply the Filter

Step 1: Write the decision down in one sentence

Vague decisions stay vague. "We should probably expand" is not a decision. "We will open a second location in Austin by Q2" is a decision. The more specific you are, the easier it is to evaluate reversibility.

Step 2: List the undo costs

What would it actually cost to reverse this in 90 days? Include money, time, relationship capital, and team morale. If you can't list the undo costs, you don't understand the decision well enough yet.

Step 3: Apply the two questions

Run both questions. If question 1 comes back "yes," stop there and approve it.

Step 4: Match the process to the type

For Type 2: a brief note to stakeholders, a review date, and a go signal. No committee. No lengthy deck.

For Type 1: a structured process. At minimum, a written summary of the decision, the key assumptions, the downside scenario, and who owns the call. Running a pre-mortem before committing to Type 1 decisions is worth the hour it takes.

Step 5: Set a review trigger

Every decision, reversible or not, should have a defined trigger that prompts you to reassess. For Type 2 decisions, set a 30-day check-in. For Type 1, define the early warning signs that would tell you the decision is going wrong before the full cost materializes.

Type 1 vs Type 2: A Quick Reference

Worked Example: A $180,000 Hiring Decision

A 12-person SaaS company is deciding whether to hire a Head of Sales at $120,000 base plus $60,000 in expected on-target earnings. The founder applies the two-question filter.

Question 1: Can we undo this in 90 days without catastrophic cost?

Terminating the hire within 90 days would cost roughly $30,000 in severance, three months of salary already paid, and recruiting fees of about $18,000 (15-20% of base). For a company doing $800,000 in annual revenue, that's a 6% revenue hit and two to three months of lost momentum. That's not catastrophic, but it's not cheap either.

The founder decides it fails question 1. The undo cost is too high.

Question 2: Are the costs of a mistake disproportionate to the upside?

If the hire works, the company's sales capacity doubles and there is a realistic path to $2M ARR within 18 months. If the hire fails, they lose $48,000 in direct costs and three to four months of time. That's painful, but not existential for a company with $200,000 in cash reserves.

The founder decides the costs are not disproportionate to the upside. This is a Type 1-adjacent decision: not easily reversible, but not existential either. The right process is: write a one-page brief on the role, define success metrics at 30, 60, and 90 days, and get a second opinion from an advisor before making the offer.

That's the filter in action. It doesn't tell you whether to hire. It tells you how much process to apply before you decide.

The Most Common Mistake (and How to Avoid It)

The most common mistake is treating a reversible decision as irreversible because it feels unfamiliar.

A marketing manager at a 25-person e-commerce company spent four weeks preparing a presentation to get approval for a $3,000 test campaign on a new ad platform. The test was fully reversible: the spend was capped, the platform contract was month-to-month, and results would be visible within two weeks. It was a textbook Type 2 decision. But because the platform was new, it got treated like a Type 1.

The four-week delay cost the company a seasonal window. The campaign launched after the window closed and underperformed.

The fix is straightforward. Before you start building a deck or scheduling a review meeting, run the two-question filter. If the answer to question 1 is yes, you probably don't need the meeting at all. Set a budget threshold (for example, any reversible test under $5,000 is approved by the team lead with a brief written note) and move on.

This also connects to how you set strategic priorities. Treating every item as urgent and high-stakes is how teams burn out and lose speed on the things that actually matter.

What Makes a Decision Harder to Categorize

Not every decision falls cleanly into one bucket. A few patterns make categorization harder.

Path dependency. A decision that seems reversible can become irreversible if it sets a precedent. Giving one customer a 40% discount is reversible in isolation. If it becomes the expectation for that customer segment, it's much harder to walk back. Second-order thinking helps here: ask what happens if everyone assumes this decision applies to them.

Time compression. Some decisions are reversible in theory but not in practice because the reversal window is very short. Accepting a term sheet is technically reversible, but withdrawing after a week damages investor relationships in ways that matter for future rounds. Apply the 90-day window strictly.

Cumulative effect. A single reversible decision in isolation is fine. Thirty reversible decisions in the same direction, all made quickly without review, can compound into a direction that's hard to unwind. Periodic reviews of your decision log help you see patterns across decisions, not just evaluate them one at a time.

When a decision genuinely straddles the line, default to treating it as Type 1. The cost of applying extra process to a Type 2 decision is low. The cost of applying Type 2 speed to a Type 1 decision can be very high.

Making This Stick on Your Team

The filter only works if people use it consistently. A few practical ways to embed it:

• Add a one-line reversibility note to your decision log or project briefs: "Reversible: yes/no. Undo cost: $X."
• Set dollar and time thresholds that define Type 2 decisions for your team so people don't have to ask every time.
• In team meetings, when someone raises a decision, make "is this reversible?" a standard first question before discussing options.

When you prioritize initiatives, reversibility is a useful lens. Reversible bets can run in parallel with less coordination overhead. Irreversible commitments need more alignment before you start. If you want a fuller framework for weighing options with uncertain outcomes, making decisions under uncertainty covers complementary tools that pair well with this one.

Key Takeaways

• Amazon's Type 1 (irreversible) and Type 2 (reversible) framework matches your decision process to the actual stakes, not to how the decision feels.
• The two-question filter: can you undo it in 90 days without catastrophic cost, and are the costs of a mistake disproportionate to the upside? Together they tell you how much process to apply.
• Most decisions are Type 2. The default bias should be toward moving fast and setting a review date, not building decks and scheduling committees.
• The most expensive mistake is treating a reversible decision as irreversible. It kills speed, drains time, and often misses the window.
• Path dependency and cumulative effect are the two traps that turn a series of reversible decisions into an irreversible direction. Build periodic reviews into your planning process to catch this pattern.
• When a decision straddles the line, default to Type 1. Extra process on a reversible decision costs little. Moving too fast on an irreversible one can cost everything.