What’s the likelihood of success of any given startup? Let’s say 10%. That’s a 90% failure rate. Why these numbers? I’ve stolen them from here but there’ are a number of references for this “90% of startups fail” number. It seems correct. Startups are hard. I’ll never understand those 19-year-olds that hit a home run on their first at-bat. I was never one of them.
Since February 2019 I’ve launched 8 products. “Launched” and “Product” here are used leniently.
I’ve put most of my time and effort into SugarKubes, and it shows in the traffic and the tiny bit of revenue. Each user has been driven more or less exclusively from Indie Hackers, Hacker News and Medium. All other referrals are just digital exhaust from these 3 main channels. Channels are something I didn’t focus on enough before launching.
According to the assumptions (90% failure rate), likelihood of one of these 9 being successful is roughly 65.1%. That’s a little misleading. It should read on average if I repeated this experiment of launching 9 products a very large number of times, I should have a 65% chance of at least 1 of those companies “working”. Working is a loose term, but let’s just define it as Ramen Profitable.
But that’s not what’s happened so far. I’m convinced that some companies need a shit ton of money to get started, and others just need a shit ton of time.
This leaves me with a small number of options to get that likelihood of success number up. Exactly two, they are:
- launch more experiments
- launch more experiments
The right amount of attempts
So I’m choosing to launch more experiments. How many though, roughly, until my likelihood of success is so high that it overwhelms the difficulty of finding the “right idea”.
A safe number is 35. There are only marginal gains after that. Though it could take more swings than 35, many more depending on how the cards land.
35 experiments is a lot. There’s almost no way over the course of a year to build 35 full products, so that leaves little room other than building essentially a pretty landing page, authentication, and perhaps one main feature. Anything else just won’t work.
So let’s break down the math of building 35 projects in a year. There are 52 weeks in a year so roughly 0.67 experiments per week, or one launch every 10.4 days.
The problem is that it’s not enough to just build it. If we have 10.5 days, we can’t spend 10 of them building it and 0.5 days telling people about it, that’s not going to work. I don’t have a magic number of how many days to spend marketing but it should be more than 1, likely closer to 2. For roundness, let’s say 3 days. That means one launch per week with 3 days of marketing (No more writing code except for, perhaps, bug fixes).
When to quit an idea
Many people much smarter than I have a lot to say on this. Taking the dumbell approach (an inverted normal distribution) they fall into two camps. Never quit, and quit very quickly.
I don’t know what’s correct. There obviously isn’t an answer. I do know that some launches get traction quickly (does it sustain, is it repeatable, etc) and others like SugarKubes that took 5 months before the first sale (does this mean it’s more sustainable because it took longer, slow growth cause anxiety/depression, etc). Either is fine. Just pick one for each experiment. Some products can be combinatorial (each project feeds into another) or exclusive (it’s doesn’t get any benefit from the other projects). I don’t think it matters that much, just decide and move on.
The Edison Approach
Edison liked brute force. Maybe he didn’t like it, but he did it.
“If Edison had a needle to find in a haystack, he would proceed at once with the diligence of the bee to examine straw after straw until he found the object of his search. I was a sorry witness of such doings, knowing that a little theory and calculation would have saved him ninety percent of his labor.”
– Nikola Tesla
For better or for worse, I’m choosing Edison over Tesla. Taking weeks to hone each idea before building and launching has never worked for me. Not that the brute force approach has either, but I just like the math better. I’ve only ever been good at throwing a stupid amount of hours at something with some kind of immediate feedback mechanism to improve each iteration. That’s all I know how to do.
So I have roughly 25 swings left before that math indicates I should hit something “ramen profitable”. It’s not depressing, I actually find it comforting. It’s not 10 failures or mediocre successes, I just haven’t looked through all the hay yet.