--- title: "Reversible vs irreversible decisions: one-way and two-way doors | Len P. van der Hof" description: Before you ask whether a decision is right, ask whether you can undo it and at what cost. Reversibility, not size, sets how much process a call deserves. image: "https://lenvanderhof.com/media/generated/blog-hero-reversible-vs-irreversible-decisions-v1.a4fa90a89e0f.wide.webp" --- [Founder Performance](https://lenvanderhof.com/en/blog/category/founder-performance/) Research guide # Reversible vs irreversible decisions: classify the door before you decide Bezos split decisions into one-way and two-way doors. Most founder decisions sit somewhere in between, and some doors close while you still think they are open. Len P. van der HofPublished 23 October 20269 min read One door you can walk back through, one you cannot. Ask which before you choose the process. Direct answer Classify a decision by reversibility before you debate it: ask whether you can undo it and what undoing would cost. Calls that are cheap to undo deserve speed and a short feedback loop. Calls you can undo only at real cost deserve a named owner, a night's sleep and a written reversal trigger. Effectively irreversible calls deserve a deadline, a pre-mortem, a five-line journal entry written before you act, a reconsideration trigger and a rested brain. The same split applies to AI agents: irreversible, public, financial or legally binding actions wait for human confirmation. ## Key takeaways - Ask "can I walk back through this door, and at what cost?" before "is this right?". Reversibility, not size, sets how much process a decision deserves. - Reversibility is a spectrum, and it decays: a two-way door can close quietly as data, contracts and habits pile up behind it. - Split the two-way door in half: Type 1 is reversible and low risk, Type 2 is reversible but costly to undo, Type 3 is effectively irreversible and gets the full framework. - Defer Type 3 calls when you are underslept, unfed, or acutely angry, frightened or defensive. Reversible calls may proceed. - For AI agents, cap the action, not the role: irreversible, public, financial and legally binding actions wait for human confirmation every time. **The first question about a decision is not “is this right?” but “can I walk back through this door, and at what cost?”** Reversibility, not size, sets how much process a decision deserves. A call that is cheap to undo earns speed. A call you cannot take back earns a slow, written process and a rested brain. Founders who get this wrong usually get it wrong in both directions at once. Watch a hard week. Three hours go into the copy on a pricing page. Then, late on a Friday after a run of bad nights, the same founder agrees to a senior hire or a core architecture choice just to get it off the desk. In chapter 6 of [*The Psychology of Founder Resilience*](https://lenvanderhof.com/books/psychology-of-founder-resilience/), part of the [RESILIENT framework](https://lenvanderhof.com/frameworks/resilient/), I call this inversion the founder’s most expensive habit: reversible decisions made slowly, and irreversible decisions made fast. One reversible mistake costs a week; one irreversible mistake can cost the company. ## What did Bezos mean by one-way and two-way doors? In his 2015 letter to Amazon shareholders, Jeff Bezos split decisions into two kinds. Some are “consequential and irreversible or nearly irreversible – one-way doors”, and they “must be made methodically, carefully, slowly, with great deliberation and consultation.” He called these Type 1. “But most decisions aren’t like that – they are changeable, reversible – they’re two-way doors.” Those are Type 2, and they “can and should be made quickly by high judgment individuals or small groups.” His warning runs both ways. As organizations get larger, he wrote, there seems to be a tendency to use the heavy Type 1 process on most decisions, including many Type 2 ones. The result is “slowness, unthoughtful risk aversion, failure to experiment sufficiently, and consequently diminished invention.” A footnote covers the opposite error: companies that habitually use the light Type 2 process for Type 1 decisions “go extinct before they get large.” He called that case less interesting and added that there is “undoubtedly some survivorship bias.” A founder can make both mistakes in the same week. The book names the pull behind each: founders agonize over minor reversible choices because the deliberation feels responsible, and rush major irreversible ones to relieve the anxiety of holding ambiguity. Neither is a knowledge problem. In both cases the founder reads the size of the choice, or the discomfort it causes, instead of the door. ## Why is “reversible” not a yes-or-no answer? Two doors are a good start and a poor finish. Most real decisions sit between them. Reversibility is a spectrum. You can leave a vendor contract, roll back a pricing change or unwind a tool migration, but each costs cash, team attention or customer trust. Lump those in with free experiments and you move recklessly. Lump them in with bet-the-company calls and the week stops. Reversibility also changes over time, and the two-door picture hides this. A door that is open on Monday can close without anyone deciding to close it. A hypothetical example: a team moves its customer records into a new CRM as a “trial”. In week one, going back is an export and an apology. By month three, the sales process, the reports and two integrations depend on it, and going back is a project. Nobody made an irreversible decision. The door closed while everyone treated it as open. So the useful question is not only “can I undo this?” but “what will undoing cost in thirty days, and who is watching that cost rise?” ## How do you classify a decision before you debate it? In chapter 6 of the book I split Bezos’s two-way door in half and number the tiers from lightest to heaviest. Note the flip: Bezos’s Type 1 is the irreversible door; here, Type 1 is the lightest. - **Type 1, reversible and low risk:** decide fast. - **Type 2, reversible but costly to undo:** sleep on the call and check your emotional state. - **Type 3, effectively irreversible:** demand a pre-mortem and recorded dissent. The chapter’s decision tree starts with one question, “Reversible?” If the honest answer is “no, or nearly”, the full Type 3 framework applies. Before you commit to a major path, it asks four more: Can this be undone in thirty days? What does undoing it cost? Do I have enough facts? Am I rested and operating in a protected time block? The same chapter gives each tier an owner: Decision classDefault ownerFounder roleRequired guardrailType 1: reversible and low riskFunctional lead or closest informed operatorSet principles, then stay outFast feedback loopType 2: reversible but costly to unwindAccountable exec or founder, depending on scopeConfirm owner, deadline and undo costSleep on it; document the reversal triggerType 3: effectively irreversibleFounder or board-defined accountable ownerRun pre-mortem, journal, dissent and reviewProtected state, qualified input, and a second reviewer where possibleRegulated, legal, HR, security or clinical boundaryQualified professional or authorized ownerEnsure the right expertise is in the roomDo not improvise outside scope Size comes second. The chapter crosses impact with reversibility to get four rules. High impact and easy to undo: test and measure (a pricing-page experiment). High impact and hard to undo: process required (a co-founder equity split). Low impact and easy to undo: delegate or automate (a SaaS tool choice). Low impact and hard to undo: contain, meaning decide, move on and limit exposure (a six-month specialized contractor). Size changes how you handle a door. It does not tell you whether the door is one-way. Then run the pre-sort: write the decision in one sentence, label it Type 1, 2 or 3, and name what the choice would foreclose. If you cannot name the foreclosed option, you are not ready to commit. ## What does a one-way door deserve? Slow does not mean vague. The chapter warns that ambiguity provides excellent cover for avoidance, and that telling the team “we are still thinking about it” usually masks a fear of commitment. A Type 3 call gets four elements: 1. **A deadline**, so drift cannot pass for diligence. 2. **An accountable owner**, so a committee cannot dilute responsibility. 3. **Required inputs**, so premature closure cannot pass for decisiveness. 4. **A reconsideration trigger**: the metric or date that forces a second look. Write it early. After new money, time or ego has gone in, the same sentence is usually a rationalization rather than a guardrail. For major commitments the chapter adds a pre-mortem, a five-line journal entry written before you act, and a three-month review. The five lines: the decision and what it forecloses; what you believed; what you feared; what happened, left blank until the review; one process adjustment. For the meeting itself, see [How to run a pre-mortem in 20 minutes](https://lenvanderhof.com/en/blog/how-to-run-a-pre-mortem/) and [Write the five lines before you act](https://lenvanderhof.com/en/blog/five-line-decision-journal/). ## Who is deciding: the rested you or the depleted you? A process cannot fix the brain running it. The chapter cites research showing that sleep deprivation impairs multiple cognitive domains and emotional regulation, and draws a narrow rule from it: do not schedule irreversible calls inside your depleted windows. Its two-minute daily state check asks three questions: - **Sleep:** did I get at least seven hours? - **Emotion:** am I acutely angry, frightened or defensive? - **Fuel:** have I eaten and hydrated in the last four hours? If sleep or fuel is “no”, or emotion is “yes”, defer all Type 3 decisions until tomorrow. Reversible decisions may proceed. The book states the limit itself: the evidence does not establish a universal sleep threshold, so treat seven hours as a checklist line, not a law. Chapter 2 puts the same rule on the RESILIENT dashboard: if executive function or somatic regulation reads empty, no irreversible decisions get made today. Once a week, take fifteen minutes in a protected block for the chapter’s reversibility audit. What did I decide that was easily reversible, and did I move fast enough? What did I decide that was irreversible, and did I use the full framework? What decision am I avoiding? What will I decide next week, and by when? The third question matters most. Naming avoided decisions keeps the backlog from becoming your company’s de facto strategy. ## Does the same split apply to what an AI agent may do alone? Yes, and there you can enforce it in software. Chapter 5 of [*The Agent Dream Team*](https://lenvanderhof.com/books/the-agent-dream-team/), built on the [ROSTER framework](https://lenvanderhof.com/frameworks/roster/), calls it the irreversibility asymmetry: “Reversible actions tolerate mistakes. Irreversible or partially-reversible actions do not.” An email sent to the wrong person cannot be unsent. Three rules from the book carry the split into an agent workflow: 1. **Gate the dangerous tools.** Chapter 5 lists email send, external calendar writes, file overwrites, payment or billing actions, CRM data changes and publishing. The default for each is human confirmation before execution. Enforce it at the strongest level your stack supports: native (the runtime blocks the tool), workflow (a separate approval step executes the action) or, weakest, procedural (the charter forbids it, but the platform would still allow it). 2. **Do not gate on the model’s confidence.** Chapter 6 cites an ICLR 2024 study in which verbalized confidence clustered between 80% and 100%, while accuracy within each confidence bin was much lower. Gate on what you can check without asking the agent: who sent the event, which folder it landed in, whether the action is reversible, and whether money or a client-visible message is involved. 3. **Cap the action, not the role.** Chapter 8’s Agent Trust Ladder runs from L1 (draft-only) to L5 (full autonomy). Irreversible, public, financial and legally binding actions cap at L3, whatever the score history, and on those actions the founder confirms before execution, every time. A role can run at L4 on routine work and drop to L3 the moment a task touches money or the public. For what should happen when the approver is not there, see [When nobody approves in time, your AI agent must not act](https://lenvanderhof.com/en/blog/human-approval-timeout-test/). For work that should not go to an agent at all, see [When not to use an AI agent](https://lenvanderhof.com/en/blog/when-not-to-use-an-agent/). ## Try this today Fifteen minutes, one page: 1. **List every decision open on your desk this week**, including the ones you are avoiding. 2. **Next to each, write how you would undo it and what undoing would cost** in cash, time and trust, now and in thirty days. If you cannot write how you would undo it, mark it Type 3. 3. **Move one two-way door forward today.** Pick a Type 1 item you have been deliberating over and decide it, or hand it to the person closest to the work. 4. **Give one one-way door its process.** Name one accountable owner, set a review date, book a pre-mortem, and write the first three journal lines before anyone acts. If a line on the page reads “undo cost: unknown”, look at that decision first. Cite this:Reversible vs irreversible decisions: classify the door before you decide.Len P. van der Hof. [https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/](https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/) · Published 23 October 2026. ## Sources 1. [2015 Letter to Shareholders, section "Invention Machine" and footnote 1 (Jeff Bezos, Amazon.com)](https://www.sec.gov/Archives/edgar/data/1018724/000119312516530910/d168744dex991.htm) · Amazon.com, filed with the U.S. SEC as exhibit 99.1 (April 2016) 2. [The Psychology of Founder Resilience](https://lenvanderhof.com/books/psychology-of-founder-resilience/) 3. [The Agent Dream Team](https://lenvanderhof.com/books/the-agent-dream-team/) 4. [RESILIENT framework](https://lenvanderhof.com/frameworks/resilient/) 5. [ROSTER framework](https://lenvanderhof.com/frameworks/roster/) ## Further reading - [How to run a pre-mortem in 20 minutes](https://lenvanderhof.com/en/blog/how-to-run-a-pre-mortem/) - [Write the five lines before you act](https://lenvanderhof.com/en/blog/five-line-decision-journal/) - [When nobody approves in time, your AI agent must not act](https://lenvanderhof.com/en/blog/human-approval-timeout-test/) - [When not to use an AI agent: five signs, and what to use instead](https://lenvanderhof.com/en/blog/when-not-to-use-an-agent/) - [The Psychology of Founder Resilience](https://lenvanderhof.com/books/psychology-of-founder-resilience/) About the author ## [Len P. van der Hof](https://lenvanderhof.com/en/authors/len-p-van-der-hof/) Entrepreneur, AI Innovator and Venture Builder Len P. van der Hof builds practical AI systems, digital ventures and evidence-informed tools for founders. ```json { "@context": "https://schema.org", "@graph": [ { "@type": "Person", "@id": "https://lenvanderhof.com/#person", "name": "Len P. van der Hof", "alternateName": [ "Len van der Hof", "L.P. van der Hof", "Leendert Pieter van der Hof" ], "honorificSuffix": "MSc", "url": "https://lenvanderhof.com/", "image": [ "https://lenvanderhof.com/photos/len-portrait-1.jpg", "https://lenvanderhof.com/photos/len-portrait-2.jpg", "https://lenvanderhof.com/photos/len-portrait-3.jpg", "https://lenvanderhof.com/photos/len-portrait-4.jpg", "https://lenvanderhof.com/photos/len-speaking.jpg", "https://lenvanderhof.com/photos/len-hero.jpg" ], "jobTitle": "Entrepreneur, AI Innovator and Venture Builder", "description": "Len P. van der Hof, MSc, is a Dutch entrepreneur and AI innovator in Zwijndrecht. He builds ReasonKit, MindSesh, Undominated.ai, books under his name, the fiction imprint LPH98.lifestyle, and technology ventures through LPH98.ventures. Eleven titles in Systems for the Strategic Self are available now, in English and Dutch.", "address": { "@type": "PostalAddress", "addressLocality": "Zwijndrecht", "addressCountry": "NL" }, "alumniOf": { "@type": "CollegeOrUniversity", "name": "Rotterdam School of Management, Erasmus University" }, "knowsAbout": [ "Artificial intelligence", "AI agents", "Agentic AI systems", "LLM routing", "SEO", "Generative engine optimization", "Answer engine optimization", "Venture building", "Founder performance", "Founder psychology", "Evidence-based decision-making" ], "sameAs": [ "https://www.linkedin.com/in/lenvanderhof/", "https://x.com/LenvanderHof", "https://www.youtube.com/channel/UCTG20buKqYYbitqqf7l3zJA", "https://www.instagram.com/Lenvanderhof/", "https://www.threads.com/@lenvanderhof", "https://github.com/Lenvanderhof", "https://huggingface.co/LPH98", "https://www.npmjs.com/~lenvanderhof", "https://www.goodreads.com/author/show/70983905.Len_P_van_der_Hof", "https://www.amazon.com/author/lenvanderhof", "https://www.bol.com/nl/nl/b/len-p-van-der-hof-msc/609879394/", "https://bsky.app/profile/lenvanderhof.com", "https://mastodon.social/@Lenvanderhof", "https://crates.io/users/Lenvanderhof", "https://cursor.com/@Lenvanderhof", "https://medium.com/@Lenvanderhof", "https://gitlab.com/Lenvanderhof", "https://hub.docker.com/u/lenvanderhof/", "https://dev.to/lenvanderhof", "https://www.facebook.com/Lenvanderhof", "https://soundcloud.com/Lenvanderhof" ], "affiliation": [ { "@id": "https://lenvanderhof.com/#publisher" }, { "@id": "https://lenvanderhof.com/#mindsesh" }, { "@id": "https://lenvanderhof.com/#lifestyle" } ] }, { "@type": "WebSite", "@id": "https://lenvanderhof.com/#website", "url": "https://lenvanderhof.com/", "name": "Len P. van der Hof", "description": "Len P. van der Hof, MSc, is a Dutch entrepreneur and AI innovator in Zwijndrecht. He builds ReasonKit, MindSesh, Undominated.ai, books under his name, the fiction imprint LPH98.lifestyle, and technology ventures through LPH98.ventures. Eleven titles in Systems for the Strategic Self are available now, in English and Dutch.", "inLanguage": [ "en", "nl" ], "publisher": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "Organization", "@id": "https://lenvanderhof.com/#publisher", "name": "LPH98.ventures", "url": "https://lph98.ventures", "founder": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "Organization", "@id": "https://lenvanderhof.com/#mindsesh", "name": "MindSesh", "url": "https://mindsesh.net", "founder": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "SoftwareApplication", "@id": "https://lenvanderhof.com/#reasonkit", "name": "ReasonKit", "url": "https://reasonkit.sh", "creator": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "SoftwareApplication", "@id": "https://lenvanderhof.com/#undominated", "name": "Undominated.ai", "url": "https://undominated.ai", "creator": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "Organization", "@id": "https://lenvanderhof.com/#lifestyle", "name": "LPH98.lifestyle", "url": "https://lph98.lifestyle", "founder": { "@id": "https://lenvanderhof.com/#person" } }, { "@type": "ImageObject", "@id": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/#primaryimage", "url": "https://lenvanderhof.com/media/generated/blog-hero-reversible-vs-irreversible-decisions-v1.a4fa90a89e0f.wide.webp", "contentUrl": "https://lenvanderhof.com/media/generated/blog-hero-reversible-vs-irreversible-decisions-v1.a4fa90a89e0f.wide.webp", "representativeOfPage": true }, { "@type": "BreadcrumbList", "@id": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/#breadcrumb", "itemListElement": [ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://lenvanderhof.com/" }, { "@type": "ListItem", "position": 2, "name": "Blog", "item": "https://lenvanderhof.com/en/blog/" }, { "@type": "ListItem", "position": 3, "name": "Founder Performance", "item": "https://lenvanderhof.com/en/blog/category/founder-performance/" }, { "@type": "ListItem", "position": 4, "name": "Reversible vs irreversible decisions: classify the door before you decide", "item": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/" } ] }, { "@type": "WebPage", "@id": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/#webpage", "url": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/", "name": "Reversible vs irreversible decisions: classify the door before you decide", "description": "Before you ask whether a decision is right, ask whether you can undo it and at what cost. Reversibility, not size, sets how much process a call deserves.", "isPartOf": { "@id": "https://lenvanderhof.com/#website" }, "primaryImageOfPage": { "@id": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/#primaryimage" }, "breadcrumb": { "@id": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/#breadcrumb" }, "inLanguage": "en-GB" }, { "@type": "BlogPosting", "@id": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/#article", "mainEntityOfPage": { "@id": "https://lenvanderhof.com/en/blog/reversible-vs-irreversible-decisions/#webpage" }, "headline": "Reversible vs irreversible decisions: classify the door before you decide", "description": "Before you ask whether a decision is right, ask whether you can undo it and at what cost. Reversibility, not size, sets how much process a call deserves.", "datePublished": "2026-10-23T07:00:00.000Z", "author": { "@id": "https://lenvanderhof.com/#person" }, "publisher": { "@id": "https://lenvanderhof.com/#person" }, "image": [ "https://lenvanderhof.com/media/generated/blog-hero-reversible-vs-irreversible-decisions-v1.a4fa90a89e0f.square.webp", "https://lenvanderhof.com/media/generated/blog-hero-reversible-vs-irreversible-decisions-v1.a4fa90a89e0f.landscape.webp", "https://lenvanderhof.com/media/generated/blog-hero-reversible-vs-irreversible-decisions-v1.a4fa90a89e0f.wide.webp" ], "articleSection": "Founder Performance", "inLanguage": "en-GB" } ] } ```