Imagine a folder on a conference table, its cover marked “Decision.” Inside are two possible next steps: change a price for a fortnight, or sell the business unit that has made the company recognizable. Both can arrive in the same meeting, under the same pressure to move. But they do not deserve the same kind of thinking. The useful first question is not simply, What should we do? It is, How carefully should we decide?
Start by asking whether the choice can be undone
Reversibility is the cost of returning to your starting point. If a choice can be changed with little loss of time, money, or reputation, it is relatively reversible. If unwinding it would be difficult, expensive, or impossible, it is relatively irreversible. The distinction is practical, not philosophical: imagine the work of reversing the choice, and name who would bear the cost.
A pricing A/B test that can be rolled back in two weeks leaves room to learn. A 90-day hire trial with an exit clause and a sales-channel pilot that can be paused after three months also preserve an off-ramp. A five-year exclusivity agreement narrows future options. Division-wide layoffs can leave lasting damage to trust, and selling a core business unit may mean there is no way to buy it back. The labels are not automatic. A supposedly temporary pilot can create obligations or expectations that make stopping it harder than it looked at the outset.
So do not sort choices into two neat bins on instinct. Ask what it would take to unwind this particular decision. Count money and time, but also consider relationships, reputation, and the options that would disappear while you waited.
Judge the size of the consequences second
Reversibility is only one side of the question. A reversible decision can still have serious consequences if it goes wrong, and an irreversible choice may have modest effects in a particular setting. The scale matters. Ask how large the harm, expense, or lost opportunity could be, not only how likely you feel it is. There is no consequence threshold in the available guidance, so a precise score would suggest more certainty than the framework supplies.
The source offers a compact rule of thumb, “Reversibility + Consequences = Decision Mode,” and places the two questions in a Reversibility-Consequences Matrix. Its account describes four pathways, but does not spell out their recommendations. That gap is worth keeping visible. The matrix can help a team decide what kind of process to use, but the available description does not support assigning a named procedure or numeric cutoff to each square. The framework is a prompt to examine the choice, not a calculator that can settle it.
Try the questions in order: Can this be undone? What would unwinding cost? How large could the consequences be if I am wrong? Even answering them aloud can change the conversation. “We can always reverse it” becomes less reassuring when someone names the people, money, or trust that reversal would consume.
Use quick trials where feedback is cheap, and slow down where it is not
When a decision is easy to reverse and its consequences are limited, a trial run can be more useful than another round of speculation. Choose a bounded test, decide what you will observe, set a point to review it, then act, observe, adjust, and repeat. A pilot is not a promise to continue. It is a way to gather information while keeping an exit available. The two-week rollback on a pricing test matters because it makes the learning period and the stopping option concrete.
When a decision is hard to reverse, a snap judgment carries a different risk. A structured process gives people time to surface assumptions, examine likely consequences, and ask what they may be overlooking. That process can take longer, and delay has costs too. A slow review can hold back a useful action; a quick test can mislead if its signals are noisy or if the test itself changes the conditions it is meant to measure. The point is not to praise slowness or speed in isolation. It is to match the process to the cost of being wrong.
One warning sign is a leader describing a consequential commitment as “just a pilot” without naming how it could be stopped. Repeatedly misjudging irreversible calls can wear away colleagues’ confidence in that leader’s judgment. Ask who can end the experiment, when they can do so, and what remains after it ends. If those answers are vague, the choice may be less reversible than its label suggests.
Let data inform judgment, not impersonate it
In business, data-led decisions are part of the ordinary record: a dashboard, a forecast, a table of results. But numbers can give a comfortable feeling of accuracy while leaving important parts of the story out. A test may report what happened under its particular conditions. It cannot, by itself, tell you every consequence of scaling the result or what remains unknown.
The authors of In a data-led world, intuition still matters argue that the challenge is judgment, not simply a shortage of information. They make the case for combining good data with good judgment and introduce an approach called Quantitative Intuition. The available excerpt does not explain the method in detail, so there is no basis here for treating it as a step-by-step technique. Its useful caution is more modest: data can inform a choice without replacing the person responsible for interpreting it.
Before treating a figure as the answer, ask whether it is creating false certainty about what you do not know. What was measured? What fell outside the measure? What would change your mind? These questions do not diminish evidence. They keep evidence in its proper place, beside context and accountability.
Keep the framework’s limits in the folder
The decision guidance discussed here focuses on business leaders. It does not establish how the same approach performs in personal decisions, ordinary household choices, or disagreements in which two sides have competing claims. Nor do the sources provide evidence about how often this framework leads to better outcomes. It should be used as a reasoning tool, not presented as a proven guarantee.
The underlying article, “Reversible vs Irreversible Decisions: Use the Matrix to Decide How to Decide”, supports a useful sequence: assess reversibility, consider consequences, then choose the level of deliberation. It does not supply a numerical measure for either dimension or the full recommendations for all four matrix pathways. That is a reason to make your estimates explicit, not a reason to pretend the blanks are filled.
For the next decision on the table, write down what can be undone, what unwinding would cost, and what could happen if the choice is wrong. Then choose a trial or a more structured review accordingly. When the folder comes around again, compare your estimate with what actually happened. The record may show that a choice you called temporary had lasting effects, or that a serious-looking decision had a safe exit. Either way, the next deliberation starts with a better account of the last one.
Sources
- Reversible vs Irreversible Decisions: Use the Matrix to Decide How to Decide
- In a data-led world, intuition still matters
- Roelof Botha on Why Success Is Never the Summit | Stanford Graduate School of Business
- Roelof Botha, MBA ’00, and Huifen Chan, MBA ’00 | Stanford Graduate School of Business
- Workplace Friction: How to Make the Right Things Easier and the Wrong Things Harder | Stanford Graduate School of Business
- EX-99.1
- How Founder Decision-Making Changes as a Company Grows
An exceptional writer at Deeperpectives, she covers public memory, family archives, and the histories held in ordinary objects. Before joining the desk, she cataloged oral histories for a neighborhood museum, where she learned to treat a small detail as a door, not a conclusion.
