Half-Baked Product: an oven-startup parable about shipping features instead of fixes

A new essay on the developer blog weli.dev uses a deliberately absurd business — a Spanish startup that sells industrial ovens — to describe a failure mode that will look familiar to anyone who has worked on a software product. The founder never fixes the thing his product gets wrong. He adds buttons instead.

The founder, the engineer and the MVP

In the story, a founder who knows the kitchen-appliance market raises money on a plan: sell a more efficient oven to Spain’s pizzerias, pastry chefs and bakers, capture 10% of the market and become a billionaire. He hires an engineer from a prestigious school who has spent ten years building ovens and argues about them on Italian forums, offering him 20% of the company and total freedom. Two months later the MVP exists: enter the flour, yeast and water, and the oven decides when to stop. In theory. In practice two thirds of the bakes fail, but the five prototype customers — two bakers, the engineer’s mother and two enthusiasts — get one perfect cake out of three, and the founder raises five million anyway. Nobody asks whether those five customers would buy again.

Every new button costs more than the last

Sales is the next problem. Small bakeries do not switch: for Juan’s Bakery a 15% efficiency gain is optional, while tomorrow’s bread is not, and Manolo across the street is reasoning in exactly the same way. Big customers are different, and one salesperson reaches Pepepizza, whose kitchens are custom-built and whose operations team wants specific dimensions and a rotating base. Engineering estimates five months; the team ships something in three weeks on lost weekends. Then the feature requests arrive: a candle button for birthday cakes, a fireplace connection, a Ramadan mode. Engineering stops trying to build a good oven and starts adding buttons — each one slower than the last, because every new button has to coexist with all the previous ones.

The original defect never gets fixed

Underneath the new buttons sits the same algorithm from day one, and customers keep returning ovens because bread and cakes still burn one time in ten. Support tries to retain a baker whose bread burns by offering the newest feature. Meanwhile the rotating base that Pepepizza actually asked for stays the second-highest priority, and in the essay’s formulation the second-highest priority never gets done. When Pepepizza finally calls, the founder says “almost finished”. The delivered base turns the wrong way — nobody in product ever recorded that it had to rotate clockwise — and the company loses its most important customer. The customer leaves; the compromise stays in the design forever.

The epilogue is the point. Mario, one of the two engineers recruited from the forums, quits because it is the only way he can see to get a vacation; Luigi stays and now maintains the candle button, which they say is his specialty, and reports “no blockers” in standups. The company is still alive, technically, with a pitch deck where the word oven has become “intelligent baking platform” — and a new young engineer signs for 5% and laughs at an old forum warning to support rotating bases from day one.