Run the experiment
The short answer
Most big life decisions are treated as leaps when they could be tested. Find the single assumption the whole decision rests on, work out the cheapest way to get real evidence about it, and put a date on that instead of on the decision. You are almost never as stuck as you feel. You are usually just missing one piece of information you have never gone and looked for.
Thinking is not evidence
People circling a decision for a year have usually thought about it several thousand times. That's the problem rather than the solution. More thinking produces more confidence in whatever you already believed, and none of it produces new information.
The question that breaks the loop isn't "what should I do." It's what would I need to know to be sure?
Answer that honestly and you'll usually find one thing. Not ten. One.
Find the load-bearing assumption
Every plan rests on something you're not certain about. Say the number out loud: how confident are you, honestly, that the alternative delivers what you want. Give it a percentage.
Above 80%. You don't have a decision problem, you have a permission problem. Something else is stopping you and it's worth naming.
Between 30 and 60%. This is the useful zone, and it's where the experiment lives. You've found the thing to go and test.
Under 30%. The alternative may be a fantasy rather than a plan. Worth asking what it's standing in for before you spend anything on it.
Low confidence doesn't mean no. It means you don't have the information yet, and missing information has a completely different fix from a hard choice.
What a good experiment looks like
Cheap enough that running it costs you nothing you'd miss. If the test itself is a big decision, it isn't a test.
Aimed at the assumption, not the outcome. You're not trying the new life. You're checking one belief the new life depends on.
Time-boxed, with a date. An open-ended trial becomes another way of not deciding.
Defined before you start. Say what result would change your mind. Otherwise you'll interpret whatever happens as confirmation, which is what your brain does when nobody stops it.
Examples, by what you're actually unsure about
You think you'd like living there. Go for a month and live on your real budget, paying rent, cooking, taking local transport. Not a holiday with a laptop in it.
You think you could earn freelance. Land one paid project at the rate you'd need. One is enough to learn most of it.
You think the work would suit you better. Find three people doing it and ask what their ordinary Wednesday looks like. Not their best day. A normal one.
You think your employer would say no. Ask. A remarkable number of these are never actually asked, and the answer is free.
You think you'd be happier with more time. Take two weeks off and live the shape of the life, rather than resting. Most people run a holiday when they meant to run a trial.
Door, bridge or cliff
Not everything needs a test, and knowing which kind of crossing you're at saves you months.
A door is cheap and you can walk back through it. Asking to go remote. Taking the recruiter call. Open it, today, and stop deliberating.
A bridge is a real change you can cross in pieces. Moving countries. Going freelance. That's where experiments belong.
A cliff is expensive and hard to undo. Quitting with no runway. Selling everything before you've spent a week somewhere.
The expensive mistake is treating a door like a cliff, and most people circling a decision have been doing exactly that for about a year.
When the result is mixed
It often is, and that's not a failed experiment. Ask what specifically was good and what specifically wasn't, because "it was fine" is not data.
Usually one of two things is true. Either you learned the assumption was wrong and the decision just got easier, or you learned it was right and something else is the real obstacle. Both are progress.
And if you genuinely cannot tell after a fair test, that's information too. It means the difference between the two paths is smaller than the story in your head, which means the cost of choosing wrong is also smaller than you think.
Questions that come up
How long should I test something before deciding?
Long enough for the novelty to wear off, which for a place is usually about a month and for work is usually one complete cycle of the thing you'd actually be doing. Two weeks tells you how it feels on holiday. Four tells you how it feels on a normal Wednesday.
What if I can't afford to run an experiment?
Then the experiment is the wrong size. Almost every one of these has a cheaper version: a conversation instead of a trip, one client instead of a business, a week of leave instead of a sabbatical. If you genuinely cannot afford the smallest version, the real constraint is runway and that's the thing to work on first.
Isn't running experiments just another way of putting off the decision?
It becomes that the moment there's no date and no defined result. An experiment with an end date and a stated criterion moves you. An open-ended trial is procrastination with better branding, and the difference is entirely in whether you wrote down what would change your mind.
What if the experiment goes well but I'm still scared?
Then you've converted a decision problem into a courage problem, which is genuine progress even though it doesn't feel like it. Fear that survives good evidence isn't telling you the plan is wrong. It usually needs a date rather than more analysis.
The frameworks under this
Last updated 5 September 2026