At last, the final chapter. In the previous Chapter 5, "Test it and ship it," your product went live. Congratulations — but this isn't the end. Launch isn't the goal; it's the real start. This chapter covers Phase 5, growing what you built into something that's "used, earns, and can be sustained." Along the way, let's also lock down the five pitfalls solo developers stumble into, together with their countermeasures.
Don't stop at building. Get it used, and shape it to last
Launch isn't the goal, it's the start
Many indie projects settle for "I got it launched!" and stop there. But a product no one uses is nearly the same as one that doesn't exist. The real learning begins after you launch. Someone uses it, gets stuck, is delighted, drops off — those reactions are the finest teaching material, telling you "what to fix or build next."
So this chapter's motto is "grow it after you ship it." Before flashy feature additions, first get just one person to use it and listen to their voice. From here, the growth loop begins to turn.
✅ Growing means "turning reactions into fuel." Before chasing visitor counts or download numbers, gather even one more piece of what the person who used it felt. Numbers are the result; voices are the cause. Grasp the cause and how to grow it comes into view.
Get "your first 10 users" to use it
Say "marketing" and you tend to imagine "going viral on social media" or "running ads," but the first step in indie development is much humbler and surer: the 10 people within your reach. Delivering directly to them. Rather than being ignored by 10,000, having 10 use it seriously gives you learning that's dozens of times richer for what comes next.
Hand it individually to friends, colleagues, family — people likely to have that problem — with "I made this, try it." Their reactions are the most honest.
Share "I made something like this" in a community or on social media where people with that problem gather. Post it not as an ad but as sharing a solution.
Always ask "where did you get stuck?" Have AI fix the inconvenience you spotted right there, and get them to try it again the next day.
This small round trip of "10 people → fix → have them try again" polishes the product all at once. When the first 10 come back, or start saying "can I tell my friends about this?", that's the signal to move to the next phase.
Marketing — get found via search and AI answers
To spread beyond your reach, you need paths where people "find you from their side." The two leads today: traditional search engines (SEO), and AI answers (AEO = AI search optimization) like ChatGPT and AI Overviews. People no longer just "Google it"; they now "ask AI." A design that gets you found by both is important.
Prepare an explanation page that answers users' problems. The clearer the writing for AI to read, the more likely it's cited in AI answers.
SEO/AEO in the AI search era →Place an llms.txt and you can summarize "what this site is" for AI. Easy to add even for a personal site.
What is llms.txt →For blurbs, social posts, and screenshot captions, have AI draft them to put out volume. People find you through "volume × consistency." Do just the final polish yourself.
💡 Social posts are AI's forte too. Ask AI "give me 5 X-post drafts for this product, aimed at people with the problem, in different tones," and your posting gets off to a fast start. You just pick a draft and refine it. Marketing isn't about talent; it's a contest of mechanisms you can keep going.
Monetization and pricing design
To keep a product going, at some point you need a flow of money. Server fees and AI usage costs recur monthly, so bleeding losses forever breaks your spirit. That said, you don't need to earn big right away. First, learn the "patterns of earning" and start small with charging.
Monetizing with AI has standard patterns. Grasp the big picture, then choose the path that fits you.
How to start an AI side hustle →A realistic path from no capital and no track record. Concrete examples of "start small and stack it up."
Earn from home from zero →For personal products, free to try + paid for part is the standard. Even a few hundred yen a month is big if it confirms "worth paying for."
When stuck on pricing, the starting point is being able to say in one line: "how much is usable for free, and what do people pay for." For example, "free by default, charge past X uses per month," or "only the convenience features are paid." Getting your first paying user is the biggest wall of monetization, and the best proof.
💡 Rather than selling high, first "get 1 yen." One person who pays teaches you the product's value far more clearly than 1,000 free users. You can raise the price later. First, confirm that "there's someone who pays."
Operations and living with AI costs
The flip side of revenue is cost. A product with AI built in especially piles up AI usage fees the more it's used. "The bill shot up before I noticed" is an indie-dev cliché. Regard cost as something to watch from the moment you launch.
Set "notify me past X yen a month" first. A safety device that stops runaway charges early. Essential especially during development.
Use a cheap model for drafts and light processing, a high-performance model only for the finish. Using them by role cuts costs a lot.
Cost optimization guide →You may hit the dev tool's own usage cap. Know the mechanism and workarounds and you won't panic.
When you hit the usage limit →The key is "cutting waste." Are you throwing the same input at AI over and over? Are responses too long? Is there processing you could cache? Consult AI with "suggest ways to lower this code's AI-call costs" and it'll help hunt down where to trim.
5 stumbles of solo dev × AI
Finally, here are the 5 pitfalls where many people trip in the same spot in solo dev × AI, paired with countermeasures. Just avoiding these five greatly raises your completion rate.
Piling on features gets complex, and you run out of steam before launch. The biggest cause of indie-dev failure.
✅ Fix: narrow to one feature and ship an MVP. Ask yourself "is this really needed now?" each time, and cut when unsure.
Hard-coding an API key or password and publishing it as-is. A cause of hijacking and huge bills.
✅ Fix: secrets into environment variables. Before launch, have AI also check "no keys or passwords mixed in."
AI is wrong in plausible ways. Paste without verifying, and you embed code that doesn't work or is dangerous.
✅ Fix: always run the output and check it yourself. Ask "why did you write it this way?" and confirm the reasoning.
Keep developing without checking the bill, and be shocked by an unexpected amount at month's end. A heavy blow for an individual.
✅ Fix: set a cap alert first. Draft with a cheap model, and periodically review wasteful calls.
Leaving it unreleased because "it's not perfect yet." A product that never reaches the world is the same as none.
✅ Fix: launch at 60 and get one person nearby to use it. Approach perfection after launch, watching reactions.
As you'll notice looking at the five, the roots boil down to two: "build small and ship first" and "don't over-trust AI; judge for yourself." Keep these two and you avoid most big failures.
How to keep going — keep spinning small laps
Up to here, we've walked the five phases: decide → prepare → build → ship → grow. The last thing to convey is that this isn't a straight line that ends once, but a loop you spin again and again. The reactions gained in the grow phase connect to the next "decide." Keeping this loop spinning small and fast is the one and only knack to sustaining indie development.
Don't hoard big overhauls. Fix one thing, get it used again. A stack of small improvements is strong.
Note the marketing that worked and the features that landed. They become assets for your next product.
Finish one, move to the next idea. The more you spin, the faster and better you get.
- Launch isn't the goal, it's the start. First get your 10 people nearby to use it, and polish with their voices as fuel.
- Marketing: get found via SEO/AEO (AI search optimization) + llms.txt, and mass-produce posts with AI.
- Monetization: learn the earning patterns and start small with charging. First make "one person who pays."
- AI costs pile up. Live with them via cap alerts, drafting with a cheap model, and cutting waste.
- Avoid the 5 stumbles (over-building / exposing secrets / face-value trust / ignoring cost / not launching) with countermeasures.
With that, all six chapters — you've walked the whole map of indie development. But merely gazing at a map creates nothing. What matters is starting one small lap now. Write down one "thing you want to build" on paper. That is the starting point of your first loop.
🚀 Now it's your turn. You don't need a perfect plan. One small idea and AI as your partner are enough. When in doubt, move your hands — that's the motto that works best in indie development. You can go back to Chapter 1, "Introduction," once more to reconfirm the big picture, or start building your first feature right now. I'm truly rooting for the day your product reaches the world.