Small bets only work if you stay long enough to find out.
This morning I posted this on X:
"ship lots of small bets and see what sticks" has a catch nobody mentions.
a bet only sticks if you stay long enough to find out. i built a lot of side projects this year and didn't stay with any of them.
fewer bets. longer stays.
This is the longer version: what went wrong with the way I ran small bets, how I now try to tell a slow idea from a dead one, and six rules you can use if you build on the side too.
What small bets promise
Small bets is the indie-hacker idea, popularized by Daniel Vassallo, that instead of putting everything on one big idea you run several small, cheap ones and let the market tell you which to keep. I still like the idea. Cheap experiments are how I got to Steady and ReplyWisely, the two products I built for a problem I had that morning.
The catch is in the last part of the sentence. The market can only tell you which bet to keep if each bet runs long enough to give an answer. Leave before that and you did not place a bet. You took a sample, and a short sample mostly measures how many people happened to see it.
What my year actually looked like
On November 17 I wrote a longer post about it. The core of it:
For the past 12 months, I've been building a lot of side projects, but I didn't stay with any of them. I thought that was because the ideas were bad. Yes, most of them were. But the bigger pattern I'm starting to notice is that I didn't want to get too close to any one idea, product, or startup.
That was uncomfortable to write, and it changes how I read my own postmortems. Some of the kills were right, and fast was the right speed. CampfireX ran eight days, got 200 clicks and zero customers, and pre-pay validation said no clearly. That was an answer.
Other exits were not answers. In the same post I admitted that plenty of ideas "went straight from idea to graveyard" without being built at all. Those never got the chance to say anything. So the problem was not that I placed small bets. It was that I kept leaving the table before the bet paid out either way.
Telling a slow idea from a dead one
This is the question I posted on November 17:
the hard part of a pivot isn't starting over. it's telling a slow idea from a dead one.
the question i keep asking myself: is anyone pulling yet, or am i still the only one pushing?
Pushing is everything you start: posts, launches, DMs, feature releases. Pulling is anything that happens without you starting it. Signs of pull, roughly from weakest to strongest:
- someone comes back a second time without a reminder from you,
- someone asks for something the product does not do yet,
- someone shares it without being asked,
- someone pays.
Steady is my smallest real example. Two people paid for an annual subscription to a fasting app I built in about a week. In the roller coaster post I wrote that this "is not a reason to quit. It is a signal to keep going." Two is small. It is still pull.
A dead idea has had a fair run and still has no pull: you are the only one pushing, and when you stop, nothing happens. A slow idea has some pull, just less than you hoped, or has not had a fair run yet. They feel the same in a bad week. They are not the same decision.
Why fewer bets matters more with a day job
Yesterday I put it this way on X:
with a day job and kids, you don't get to try twenty. what's missing isn't ideas. it's knowing which problem you'd still care about in year two.
I build at 5am and in the margins of weekends. At that pace, every extra bet comes straight out of the time the others needed to give an answer. Someone building full time can spread across ten small bets and still give each one a fair run. Someone building on the side who spreads the same way gives each one almost nothing, then reads the silence as a no.
Six rules for fewer bets and longer stays
- Count your open bets. Write down every product, project and idea you spent time on this month. If you cannot touch each one every week, the list is too long. Keep the one or two you would still think about on a bad week, and park the rest on purpose instead of letting them fade.
- Write the answer you are waiting for, and the date. Not "see how it goes." Something you could check: a number of people who come back, a number of replies from the right reader, a first payment. Then a date to check it.
- Name your pull signal before you start. If you define it after the numbers come in, you will define it to match your mood.
- Kill early only on a clear no. You asked for money and nobody paid. The thing only works with a crowd you do not have. Those are answers. "It feels slow" is not.
- New ideas go on a list, not into a repo. At least until one of the open bets reaches its date. Most ideas on the list will look less shiny in a month, and that is the point of the list.
- On the date, make one of three calls: stay, narrow, or stop. Narrow is the one most of us forget. For niches on X, I wrote the full method as the stick-or-quit protocol on the NicheSlap blog.
None of this makes the first bet a better idea. That part is the filter in the two questions I ask before I pick a problem worth solving. These rules are for after you have picked: they make sure the bet lasts long enough to tell you something.
If you are earlier, or later, than this
If you do not have a problem worth staying with yet, start one step back. I built IkigaiNiche as a free guided way to find a person you would be glad to help and a problem of theirs worth solving.
If you already have one and you post about it on X, the same pattern shows up in your posts. I switched niches on X three times in seven months, and each new niche was already narrow. I wrote about that in niche down on X: narrowing is easy, staying is the hard part.
Related: Why I killed CampfireX after 8 days · Steady: two paid subscriptions · The roller coaster of wanting to quit the day job
Follow the build: @chams_builds on X