Two options. You have compared them a hundred times. Stay or go. Safe or scary. The sensible promotion or the thing you think about in the shower. You made the pro and con list, probably several, months apart, with the same items in different handwriting.
The list never decides. It cannot. Here is why, and what produces an answer instead.
Why pro and con lists fail here
Those lists work for bounded choices with facts in them. Which apartment. Which laptop. They fall apart on career decisions for three reasons.
- You are comparing two works of fiction. Neither future exists. You are not weighing consultant you against founder you. You are weighing your imagination of each, and imagination is wrong about what days feel like.
- The weights are made up. Is more money worth more than creative freedom? You do not know. That is the thing you are trying to decide. A list without honest weights is anxiety with formatting.
- Identity decisions are not information problems. You do not lack facts. You lack evidence about yourself, meaning how you respond when the new path is real instead of imagined. No amount of thinking produces that. Only contact with reality does.
The loop never ends because you are asking analysis to do a job only experience can do.
First, check it is really two options
Before anything else. Binaries are suspicious. "Stay in my job or start a bakery" often hides a third option that never reached the list because you rejected it in advance. That is what question five is for. Sometimes the real answer is a version nobody framed. Stay, negotiate four days, test the bakery on the fifth. Widen the frame once, on purpose, before you narrow it.
The 90 day test
This is the method that replaces the list. It costs you one season and no resignation letter.
1. Pick the option that pulls, not the one with the better case
Test the path that keeps coming back. The one you reread articles about. The one that made you open this page. The safe option usually needs no test, because you are living it and you already have its data.
2. Design the smallest real version
Not "quit and find out". The smallest test that involves real stakes and real strangers.
- One paying client, found and served, on evenings and weekends
- One project shipped publicly under your own name
- Two honest hours with someone five years into that path. Ask what a Tuesday looks like, what it pays, and what it took from them
- One course with a deliverable at the end, not just attendance
The test has to touch the market. Reading about the path is more thinking. A stranger paying you, rejecting you, or hiring you is data.
3. Decide in advance what counts
Write it down before you start. "If I find one paying client in 90 days and the work gives me energy, that is a yes. If I dread the work by week six, that is a no." Set the threshold in advance. Otherwise you will move the goalposts to match whatever you were always going to do, and you are back on the loop by winter.
4. Review honestly at day 90
Compare against the threshold you wrote, not against how you feel that particular Tuesday. Bring in one honest outsider to read the evidence with you, because you will grade generously in whichever direction scares you.
What if the test fails
Then you won. A failed 90 day test buys you something five more years of wondering never will, which is a closed question. You stop paying the daily tax of what if. And the current path feels different once you have chosen it instead of defaulted into it. The only bad outcome is the one you are in now, which is not deciding, indefinitely.
If it is genuinely 50/50
Three tiebreakers, in order.
- Pick the more reversible option. You can usually return to employment from an experiment. You can rarely run the experiment from deeper inside the safe path.
- Ask the 80 year old. Not which is smarter. Which would you regret never trying? Regret cuts through noise that analysis cannot.
- Follow your attention. Whichever path you read about when nobody is watching is not a coincidence. Attention is the signal that does not lie.
What to take from this
You are not indecisive. You are asking a list to produce evidence it does not contain. Widen the frame once, pick the option that pulls, design the smallest real test, write the threshold down first, and let 90 days of contact with reality make the call.