Bottom line
A business decision is reversible when you can undo it at a cost and within a timeframe you would accept, and irreversible when undoing it is impossible or so expensive that it effectively commits the company. Most real decisions in a growing B2B company sit somewhere in between. The useful question is not "Can we undo this?" but "What would undoing it cost, who would bear that cost, and how long would it take?"
The idea is familiar, but the binary is too simple.
In Amazon's 2015 letter to shareholders, Jeff Bezos described consequential, nearly irreversible "one-way doors" that should be made "methodically, carefully, slowly," and changeable "two-way doors" that "can and should be made quickly by high judgment individuals or small groups." That shorthand is useful, yet few decisions in a 50- or 300-person company are purely one or the other. A price increase can be removed from the price list, but not from contracts already signed at the new rate or from customers' memory of it. A senior hire can be reversed on paper, but it can cost months, team trust and severance. Treat reversibility as a spectrum and examine each part of the decision separately.
Score the cost of reversal across five dimensions.
Rate each as low, medium or high, and name who would bear the cost. Money: sunk costs plus exit costs such as notice periods, termination fees and write-offs. Contracts: leases, multi-year customer or supplier agreements and employment terms; check the actual documents, not your memory of them. People: hires, reorganizations, layoffs and promises made to employees. Customers and reputation: public announcements, price commitments and discontinued products. Time and options: market windows, team capacity and alternatives that close while you pursue this path. A single high rating is enough to treat that part of the decision as a one-way door.
Match the process to the score.
For a two-way door, use one owner, a deadline measured in days and a short note recording what was decided, which signal would reverse it and who checks. Do not escalate it to the board or schedule a two-day workshop. For a one-way door, write down the options including the status quo, list the assumptions, run a pre-mortem or a structured challenge, get legal, tax or financial review where relevant, and name the approver. In practice, much of the damage comes from the mismatch: weeks of debate over a cheap test, while a lease or a market exit is approved in the last ten minutes of a meeting.
More on Decision Science — every week.
Join 2,400+ leaders getting the decision intelligence briefing.
Redesign the decision to make it more reversible.
Before accepting a one-way door, ask whether it can be split. Common moves include a pilot in one region or segment, staged commitments with explicit decision gates, a shorter initial contract term or a break clause, an interim or fractional role before a permanent hire, and announcing a change after it has been tested rather than before. Each option has a price: pilots can produce weak signals, and flexible terms often cost more. A smaller first step is still often the cheapest way to obtain the evidence the larger commitment needs.
Hypothetical example: entering Switzerland.
A 120-person German B2B software company is considering the Swiss market. The full version means founding a Swiss entity, hiring a country manager, signing a three-year office lease and announcing the launch publicly. Scored honestly, money is medium, while contracts and people are high. The redesigned version serves a handful of Swiss customers from Germany for two quarters, uses a sales contractor or partner, and defines in advance which signals would justify an entity: qualified pipeline, sales-cycle length and any local requirements discovered along the way. The entity decision remains a one-way door, but it is now taken with evidence. This is an illustration, not legal or tax advice; cross-border setup, employment and VAT questions need qualified advisors.
Write the reversal trigger down before you start.
A two-way door is only reversible if someone actually walks back through it. Record the signal, the date by which you will check it, who has the authority to reverse, and the planned way back. Without that, "we can always undo it" quietly becomes a sunk-cost commitment. For a one-way door, record why you accepted the irreversibility and which assumptions you were least sure about, so a later review can separate bad luck from bad judgment.
Use AI to challenge the classification, not to make the call.
AI advisors are useful for surfacing hidden lock-ins, questioning a confident "we can always undo it," and proposing ways to stage a commitment. They cannot verify your contracts or know how much disruption your team can absorb. If a one-way-door decision deserves a structured debate, NeuroAgent's council can challenge it against your stated constraints, and a Decision Sprint turns that into a decision-ready brief, while the decision and its approval stay with you.
Research context
In a McKinsey survey, 72% of senior executives said bad strategic decisions were as frequent as good ones or were the prevailing norm.
McKinsey & Company: Untangling your organization's decision making (2017)This independently published statistic provides context for the topic; it does not substantiate every claim in this essay.
More on Decision Science, weekly.
One essay. Every week. No noise.