Between the buzzer and Scopa Fortuna, I started two other projects. They have one thing in common: they do not exist.
The first, on 14 February, was a Telegram bot to automate selling clothes on Vinted. I had a fifty-five-page specification. The second, on 26 February, was an amateur esports tournament platform with persistent rankings, a twenty-page product document, an architecture, a multi-package workspace. One commit, then nothing.
What I believed
That the more complete the scoping, the safer the build. It is a conviction from my old job: you do not start a project without a brief, without requirements, without a risk list. I had found a project template for the tool, very thorough, with fourteen specialised agents and eight phases, from discovery to launch. It turned any idea into a project folder within an hour.
For the Vinted bot, the discovery phase reached one hundred percent on day one. Eighteen user stories, seventy requirements. The planning phase stayed at zero.
What it was
Two causes, different ones.
The Vinted bot died because discovery discovered something: Vinted has no interface for what I wanted to do, and working around that meant playing against the platform. The documentation showed it clearly. That is a good outcome, in a sense; it is just that I did not need fifty-five pages to learn it, and the template had me spend a day structuring a project before asking the only question that mattered.
The esports platform died because it was too big for one person and too vague to start. The product document was beautiful. It described a sports federation for amateur esports. It did not describe what I could put online on Friday. Scopa Fortuna, started on 18 February with the same template, worked because the design document said exactly what to build, in what order, and I could play on the first evening.
What I take from it
The template was not at fault. It does what it does well: produce documentation. The problem is that producing documentation has become so easy that you can write hundreds of pages without having made a single decision. Before, writing fifty-five pages cost enough that you asked whether it was useful. Not anymore.
Since then, I ask one question before any project folder: what can I put in someone’s hands tonight? If the answer is “nothing, we need to scope first”, I scope by hand, on one page. If the answer is “something ugly that works”, I build, and the folder waits.
Documenting a project no longer costs anything. What still costs is deciding.
Get the journal by email
About one entry a week. What I build, what breaks, what I learn. No promotion, one-click unsubscribe.
You can also follow the RSS feed.