Six months of side-hustle app development with AI: 9 ideas, ¥16,000 take-home pay, ¥600,000 loss. I investigated why they did not reach people or get chosen
It has been six months since I started a side hustle making apps with AI alongside my day job.
Here are the results first. I planned 9 apps and managed to release 4. I have just submitted the 5th for review. My take-home pay for those six months was roughly ¥16,000. After the MacBook and monthly AI subscription costs, I made a loss of roughly ¥600,000.
'AI has made making things easy, but reaching people is hard.' I see some version of that on X almost every day. Frankly, I am tired of hearing it. I worked in marketing for 10 years. My point is this: an observation is no use to anyone unless it says why it did not reach anyone.
In this article, I look back over the six months in chronological order, checking my own numbers and public data whenever I hit a wall to ask why. To give you a little of the conclusion upfront, there was only one moment in those six months when I broke through. That one time showed me everything I was missing.
The plan I made in March and the reality by October
Before I began, I wrote a strategy memo. Put that beside where I am now, and this is what it looks like.
| March plan | Results by the end of October | |
|---|---|---|
| Releases | 1 a month, 3-6 by year-end | 4 released |
| Income | Over ¥10,000,000 a year from a side hustle | Take-home pay: roughly ¥16,000 |
| Fixed costs | Under ¥5,000 a month | A loss of roughly ¥600,000 after the MacBook and AI subscription costs |
| Posting | 20 posts a month on X, 1 note post in the month of each release | 6 note posts, March-May. I kept X going until August and reached 862 followers. Even so, I could not grow my readers |
On the other hand, this is what I made with AI and put out into the world over those six months.
| What I made with AI and put out over six months | Number |
|---|---|
| Apps released | 4 (+1 under review). Store listing pages in up to 18 languages |
| App Store screenshots | 523 |
| App Store preview videos | 64 |
| Promotional videos released on YouTube | 38 |
| My own site, strataix.net | 148 pages (5 app pages in up to 18 languages, plus 73 pages of terms and privacy policies) |
| Press releases | 6 (10 languages at first, then 18) |
| PR emails | Roughly 780 |
| note | 6 |
| Automations running behind the scenes | 15 |
| Records of work built up with AI (commits) | 2,323 |
This is the volume I put out alone, alongside my day job, as someone who cannot write code. When it came to making things, AI really did let me do something close to what a one-person studio can do. What I lacked was money, and readers. I will look at where that gap came from, one step at a time.
The PR that went nowhere in Japan, the two weeks when it sold only in Germany, why roughly 780 emails led to 3 pieces of coverage, the real reason it was seen but not picked, and the near-resigned acceptance of being called a 'copycat'. The full account of those six months is published on note in Japanese.
Read the rest on note, in Japanese →I have also published part two.Six months of side-hustle app development with ¥16,000 take-home pay, part two. It was the App Store, not social media, that sold them
The apps mentioned in this article are still available on the App Store.