0 → The language app postmortem
Severo, in standby, gets its honest autopsy. Why did it fail? The core wasn't there: they used it, but it never made you want to play, to learn, and the users said the same, thirty people from a post, two the next day, one the day after, then none. That was the signal. It was a red ocean against Duolingo, a giant. And the deepest error, the one Juan reached reading Tynan Sylvester's Designing Games: it was never a game. They tried to sell it as a Duolingo replacement, but it didn't come close, it was a tool, which is exactly what their ex-partner had felt when he said it was basically a chat. Asked what he'd do differently: focus 100% on gamification from the start, start on the web so everyone can use it, and build a real demo. What he wouldn't remove is the experimentation, mobile taught them Flutter and the Play Store, hard-won scaffolding they now build on.
// trace: where this idea came from
- ↳ video diary @ 40:54 (por qué dejamos Severo, el core no estaba, no daban ganas de jugar ni aprender)
- ↳ video diary @ 41:56 (los usuarios lo confirmaron, 30 de un post, 2 al segundo día, 1 al tercero, ninguno al cuarto; esa fue la señal)
- ↳ video diary @ 48:48 (leyendo Designing Games caí en cuenta que no era un juego; lo vendíamos como reemplazo de Duolingo pero era una herramienta)
- ↳ video diary @ 45:39 (qué haría diferente, gamificación 100% desde el principio, empezar en web, hacer un demo real)
- ↳ video diary @ 47:10 (lo que no quitaría es la experimentación; el móvil nos enseñó Flutter y la Play Store, la andamia queda)
- ↳ Entry 272-1: The pipe with no water (Severo es el caso concreto de la tubería sin agua)
- ↳ Entry 269-2: The AI has no soul (la misma app cuyo problema de "sin alma" ya se había diagnosticado)
Severo, the language app they’d poured seven months into, is in standby now, and Juan finally performs the autopsy plainly. Why did they drop it? Because the core wasn’t there ▸ 40:54. They used it, but it didn’t make you want to play, didn’t make you want to learn. And it wasn’t only their own verdict. The users said the same thing, in the cruelest, cleanest form there is: a post would bring in twenty, thirty people, and the next day there were two, the day after one, the fourth day nobody ▸ 41:56. That retention curve, falling to zero in four days, was the signal.
It was also a red ocean. They were swimming straight at Duolingo, a giant that isn’t perfect but is enormous, the kind of thing you’ll use your whole life whether you like it or not. Competing there, from zero, was always going to be a knife fight against a whale.
no era un juego; era una herramienta →
But the deepest error he only saw by reading, which is telling in itself. Working through Tynan Sylvester’s Designing Games, he started catching everything that hadn’t worked, and the core of it was this: Severo was never a game ▸ 48:48. They’d tried to sell it as a Duolingo replacement, and it didn’t come close, it was a tool. Which is exactly what their ex-partner had felt when he said, near the end, that it was basically a chat, what’s the difference between this and a chatbot? He’d been right.
Asked what he’d do differently if he rolled back to January, when the investment came in and they committed, the answer is a three-item spec. Focus 100% on gamification from the start, draw the whole plan around it, because pretty pastel graphics did nothing; there are games with no color at all, pure black, and people play them because there is something there. Start on the web, so everyone can use it in any browser, instead of shipping Android and locking out iOS. And build a real demo early.
What he would not remove is the experimentation. The detour through mobile taught them Flutter, taught them the whole ordeal of the Play Store, testers, approvals, and that scaffolding is real: it’s why rebuilding Zenota and reshaping Severo’s bones into the next thing wasn’t hard. That, he says, is the beautiful part of programming, you never build from zero. Even a dead product leaves standing walls…