Starting
Doubt Costs More Than Failure
Ask someone what they regret and you will usually hear about something they did. A job they took, a message they sent, a relationship they stayed in. Ask them about their life as a whole, from a distance, and the answers change shape entirely.
They stop talking about their mistakes and start talking about the things they never tried.
The two kinds of regret
Thomas Gilovich and Victoria Medvec studied this directly in the mid-1990s, and found a clean temporal pattern. Regrets over actions — things done badly — dominate in the short term. They sting immediately and sharply. But over a longer horizon the balance reverses: when people are asked about their biggest regrets across a lifetime, the majority concern inactions. In their American samples, roughly two-thirds of long-term regrets were about things people did not do.
The explanation they proposed is worth more than the statistic. Action regrets tend to close. You did the thing, it went wrong, there were consequences, you dealt with them, and time did what time does. The story has an ending, and you can repair it, apologise for it, or at minimum learn something specific from it.
Inaction regrets never close, because nothing happened. There is no outcome to process, no lesson to extract, and no bottom to the imagined alternative. The version of your life where you tried stays permanently open and permanently flattering, because it never had to survive contact with reality.
A failure is an event. It ends. Doubt is a subscription, and you keep paying it.
What failure actually costs
Failure is loud, which is why we overestimate it. Doubt is silent, which is why we underestimate it, and the accounting is completely lopsided.
A real attempt that goes badly costs you some time, some money, possibly some embarrassment for a fortnight — and it buys you information you cannot obtain any other way. You learn whether you actually enjoy the work, what the real bottleneck is, how long things take, and who might help. That information has a price, and a failed attempt is the cheapest place to buy it.
Doubt costs you the attempt, keeps the information, and charges rent. It occupies the same mental space for years, resurfacing at odd hours, quietly informing you that you were the kind of person who did not try. There is no receipt for it and no point at which it is paid off.
You have not failed. You stopped before you began.
There is a specific state that gets misfiled as failure. You had the idea. You thought about it for months. You researched, planned, refined, and mentally rehearsed. Then you did not do it — and eventually you concluded the whole thing had not worked out.
Nothing was tested. No version of it exists in the world. What actually happened is that the project stayed at the planning stage until it quietly expired, and then you filed the expiry under "I tried."
This matters because the two are treated very differently by your own mind. A tested failure gives you evidence and a next step. An untested plan gives you a permanent maybe, which is far heavier to carry and teaches you nothing.
Do it badly, on purpose
The instruction that breaks the deadlock is not "be braver." It is much narrower: make a real version of it exist, at whatever quality you can manage today.
Publish the rough one. Send the imperfect message. Make the first client call before you have the perfect script. Record the video with bad lighting.
Doing it badly is not a lowering of standards. It is the only way to convert an unanswerable question — would this have worked? — into an answerable one, and answerable questions are where all progress comes from. A bad attempt gives you a real result to improve. A perfect plan gives you nothing but more planning.
The other benefit is that the fear shrinks on contact. Most of what you dread is imagined and, being imagined, has no ceiling. A real outcome, even a poor one, is always smaller than the version you have been rehearsing. Almost everyone who finally does the thing reports the same thing afterwards: it was much less dramatic than expected.
What to do with this
- Run the long-horizon test. Not "what could go wrong," but "in ten years, will I regret trying this more than never knowing?" The honest answer is nearly always the same.
- Shrink it until it is testable this week. Not the business — one conversation. Not the book — one chapter, shown to one person.
- Set an expiry on the planning. Give yourself a date after which the plan gets executed at whatever state it is in. Planning expands to fill whatever time you allow it.
- Write down what a failure would actually cost. In real terms: money, weeks, who would notice. The number is usually much smaller than the feeling, and the feeling shrinks once the number is on paper.
- Count an untested idea as a cost, not a neutral. It is not sitting there safely. It is charging you something every month it stays unexamined.
You have not failed at most of the things you feel you have failed at. In many cases nothing was ever attempted — the attempt was replaced by a long, careful, exhausting period of doubt, which is the more expensive of the two by a wide margin.
Try it badly. At least it is real.