A leader should decide about as fast as the decision is cheap to undo, and no slower than waiting can afford. Many of the calls you make in a week can be reversed quickly if they turn out wrong, and those deserve a fast answer with the information you have, plus the willingness to correct course once you see how it’s going. A few can’t be undone, and those deserve slow, careful work. The decisions that hurt are in between: ones you can reverse on paper, like a hire or a price change, but that cost real time and money when you do. That middle band is where two of my own fast calls went wrong, and it’s where I’d slow down long enough to get some data.
If your company decides too slowly, most of this is for you. If you decide fast by nature, as I do, a good part of what follows is about the brakes.
Where the one-way and two-way door idea comes from
The clearest version of the reversibility test is Jeff Bezos’s, from his 2015 letter to Amazon shareholders. He split decisions into two types. Some are “consequential and irreversible or nearly irreversible,” which he called one-way doors, and he said they should be made “methodically, carefully, slowly, with great deliberation and consultation.” Most decisions, he wrote, are two-way doors: “If you’ve made a suboptimal Type 2 decision, you don’t have to live with the consequences for that long. You can reopen the door and go back through.” His worry was that as companies grow, they start running every decision through the heavy process meant for the rare one-way door, and everything slows down.
A year later, in his 2016 letter, he added a rule of thumb for pace: “most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90%, in most cases, you’re probably being slow.” He paired it with the part that makes it work: “you need to be good at quickly recognizing and correcting bad decisions.”
I’d treat the 70% as Bezos’s rule of thumb, not a threshold. Nobody can measure what 70% of the information looks like, and he was writing about how to keep a very large company moving fast. What matters is the question underneath it: if this turns out to be wrong, what does it cost to undo, and how soon will I know?
The middle is where fast calls get expensive
Bezos’s two types are a good start, but the decisions that have cost me sat between them. I’d call those medium: you can reverse them, but reversing them costs months, money, a client’s patience or a person’s livelihood. Two of my own fast calls landed there.
A price raise made on instinct
At the agency I founded and still lead, feedback from prospective clients was that we were expensive, but not the most expensive they’d seen. I took that as room to move, and we raised some of our pricing, incrementally.
I think it priced us out of a lot of later deals we didn’t need to lose. Looking back, I should have been a little less instinctual about that raise and a little more data-driven. Feedback like that, however flattering, isn’t enough evidence on its own for a change that touches every proposal after it. A price is easy to change back, but the deals lost while you find out don’t return. Losing more proposals wouldn’t by itself prove a raise was wrong, because the margin on the work you do win can outweigh what you lose. The trouble with an instinctive raise is that you haven’t set yourself up to tell which happened.
If you’re weighing a price change, the slower version doesn’t have to take long. Decide before you raise it how you’ll judge the result: win rate, average deal size and the margin on the work you win, compared across similar opportunities, plus what the ones you lose say about why. Then set a date to review it and decide whether to keep, adjust or reverse the change. You can also shrink the bet by trying the new price on a subset of comparable proposals before you change all of them.
A hire made because we needed someone now
The second was a hire. We had an immediate need, so we filled the position, and we hired in a less-than-ideal situation that I regretted later. The urgency turned a medium decision into a fast one. I would rather have gotten creative about how to get the work done in the meantime and waited for the right candidate.
That’s where I land on whether hiring is a one-way or a two-way door. It’s reversible, but I’d call it a medium: not an irreversible thing, and much more costly in time and money than a two-way door. The cost of undoing a hire includes the time spent on the wrong person and what it does to the people working next to them. When the pressure to fill a seat is what’s setting the pace, that’s the moment to look for a stopgap (a contractor, a temporary split of the work, a deadline you can move) rather than letting the pressure make the decision. That’s not a rule to wait forever. In a different situation, leaving a critical role empty can cost more than an imperfect hire, so weigh both risks instead of only one.
How to tell which kind of decision you’re facing
Before deciding how long to take, sort the decision. These questions do most of the work:
- What does it cost to undo? Count money, time, people and reputation, not only whether reversal is technically possible.
- How soon would you know it was wrong? A decision you can read in two weeks can go faster than one that takes a year to show anything.
- Is the information you’re waiting for actually coming? If a week’s wait gets you a real answer, wait. If you’re waiting to feel sure, the extra time won’t help.
- Is there data you could get cheaply first? This is the question I skipped on pricing. On a medium decision, I think a little evidence is worth the delay it costs.
- What does waiting cost? The opportunity you might miss, the work held up behind the decision and the cost of leaving things as they are.
- Who owns the call? If every reversible decision has to reach the CEO, the company stays slow however good its process is. Name an owner, set the limits of their authority and say when a decision comes back to leadership.
There’s often a third option besides deciding now or waiting for more information: make the decision smaller. Test a price on a few proposals instead of all of them, cover the work with short-term help while you look for the right person, or pilot a new system with one team before rolling it out. Shrinking the downside can turn a medium decision into a two-way door.
Sorted that way, decisions tend to fall into three groups. The examples are illustrations; your own lists will differ.
| Type | Examples | Pace |
|---|---|---|
| Two-way door | Trying a new tool, testing a campaign, changing a meeting format, adjusting an internal process | Decide fast, with what you have, and correct as you learn |
| Medium | A hire, a price change, taking on a large client, a new service line | Name the evidence that would change your mind, set a decision date, and shrink the bet where you can |
| One-way door | A long lease you can’t exit, a merger, a multi-year contract with heavy penalties | Take the time, bring in the people with expertise, and decide deliberately |
Two habits help with all three. The first is to decide when you’ll decide, before the analysis starts. A date set up front stops a two-way door from drifting into weeks of discussion, and it forces a medium decision to name what information is worth waiting for. The second is to agree what a good-enough outcome looks like before anyone argues for an option, so the debate is about the criteria and not whoever argues hardest.
Fast deciders need brakes
If you decide fast, the risk is different, and I know it from the inside. I’ve described the way Maria Warner, my business partner, balances me like this: “I move fast and make quick decisions. She keeps me grounded and makes sure I’m not pushing us too far into risk.” Having someone whose job, formal or not, is to challenge your assumptions and spot the risks you’d skip on the medium and one-way decisions is worth more than any rule about percentages. Sometimes that person speeds you up, because once the concerns are answered you can commit.
The other brake is a limit on how much you have going at once. Each fast decision can be fine on its own and still add up to too much. I used to jump from one new initiative to the next and end up with a lot of half-finished things and very few finished ones. What fixed it was treating new initiatives as bets, weighing what each one would do for the company’s economics against what it would cost, and capping how many run at the same time. What worked for us was about three at once, which left room for a new idea without the pile-up I used to create.
Fast correction has a limit too. Correcting course quickly is what makes fast decisions safe, but changing direction before a decision has had time to show you anything is just a different kind of thrash. Give each call long enough to produce evidence you can read, then keep going, change it or stop, unless something new shows a risk serious enough to act on straight away. It helps to write down what you decided, what would make you change your mind and when you’ll look again, so the review happens on a set date.
Most of this a leadership team can put into practice on its own. Where outside help earns its place is on the medium and one-way decisions where the downside, the complexity or a split leadership team justifies a second view, and working through those with owners and leadership teams is what I do as a business advisor. If you have one of those in front of you, start a conversation on Work With Me.
