Feature Graveyard

Things we built that nobody used. Register one and you get a plot, a certificate and an hour count — so killing it stops being a thing you have to explain.

Register a death

323 full-time weeks buried across 7 of 23 features weighed.

The heaviest tolls

See all
  1. 14 years on a startup with no buyersno demand · chrisaycock208 weeks≈ $1,000,000
  2. 218 months building for users they never metno traction · onagengast59 weeks≈ $350,000
  3. 3A cap table calculator, 1 year of eveningsno traction · calcsam780 hours≈ $100,000
  4. 42 years on a SaaS with zero salesno demand · jcadam780 hours≈ $100,000
  5. 5A hosted SaaS built on the strength of 1 pollno demand · quaffapint390 hours≈ $47,000

≈ is our estimate, not a figure anyone gave — the hours priced at contract rates for the year. How that is worked out

Everyone buried here

What we have learned

Not a ranking — just the pile, so it is a bit less lonely.

23 entries · showing 5

A personalized dashboard

The customer took another client.

Take from it Create an MVP and ship it, don't go too far with a client you don't have yet.

replacedburied 2 Sept 20266 weeksantionmil

0 stones

PyFlow, a Rust package manager for Python

Wrote a Python dependency and version manager in Rust, modelled on Cargo — one tool that handled dependencies, scripts and Python versions together. The problems it solved were real. Almost everyone who saw it came back with the same question: why would I use this when we have pip, pipenv and poetry? Abandoned, because nobody used it.

Take from it A better tool still has to beat the inertia of the ones people already have. The first question back was why anyone would switch from pip.

abandonedtold 12 Oct 2025the__alchemist on Hacker News ↗

0 stones

A company API rebuilt around HATEOAS

Implemented HATEOAS properly across a company's API, the way the papers describe it. Every request then spent processing time assembling the URLs and the list of actions a client could take next. The only things that ever read any of it were the company's own client libraries — written by the same team, entirely under their control — and carrying it made those libraries more complex rather than less. An interesting idea in theory. A nightmare to run.

Take from it The only consumers were the client libraries the same team controlled — so the generality bought nothing, and cost processing time on every request.

shipped, unusedtold 9 Jul 2025gsibble on Hacker News ↗

0 stones

4 years on a startup with no buyers

4 years trying to get a startup to revenue, building for a niche audience that turned out not to want it. Burned out badly enough to stop for the sake of their own mental health, and took a normal job 3 years ago. Says life got dramatically better afterwards: new sports, a house, and no constant sense of having to work more and spend less. The technology was impressive enough that the job search took a couple of months and paid better than anything before it.

no demandtold 25 Feb 20244 yearschrisaycock on Hacker News ↗

0 stones

A system that blocks spam at the source

Their argument was that working out where spam came from does nothing to stop it arriving, so they built a system that prevents it at the source instead. Made it available. Nobody used it and nobody cared. It still works well for them personally.

shipped, unusedtold 8 Jul 2023Lk7Of3vfJS2n on Hacker News ↗

0 stones
18 still buried below