A product graveyard looks negative only if you think a studio is supposed to perform certainty.

I do not. A studio is a place where ideas are tested against reality. Some survive. Some become useful for a while and then stop earning their keep. Some fail because the timing was wrong, the promise was too soft, the audience was too vague, or the maintenance cost quietly beat the upside.

Hiding those products makes the public story cleaner, but it makes the pattern harder to trust.

Why dead products matter

Dead products are evidence. They show where judgment was wrong, where distribution did not exist, where a feature was interesting but not urgent, and where the builder learned something that changed the next attempt.

That evidence is hard to get from a portfolio. A portfolio says, “Look at the good parts.” A graveyard says, “Look at the pattern.”

The pattern matters more than any single launch. This studio has two tombstones a decade apart, and they rhyme. MeAri (2015) was a finished anonymous app that shipped, found about 2,000 people, and died of its own mechanic: a vote-driven clock that needed a crowd to keep posts alive, in a room that never filled. Needles.AI (2024) generated genuinely good tattoos for free on servers I built by hand, and lost anyway, because the model layer had turned into a capital game I could not outspend. Different decades, different products, same shape — each one staked on the single fight where effort does not beat the missing thing. That is the pattern. No portfolio would have shown it.

What a good tombstone should say

A useful postmortem is not a dramatic confession. It is a clear record. It should answer a few questions without making the reader decode the story:

  • What shipped?
  • Who was it for?
  • What did the product promise?
  • What actually happened after launch?
  • Why did it stop?
  • What changed afterward?

The tone should be direct. No fake humility, no theatrical regret, no victory lap disguised as a failure story. The best postmortem sounds like operating data.

What stays out

The graveyard is not a dumping ground for every abandoned idea. Ideas that never shipped do not need a tombstone. Private sketches do not need public ceremony. The graveyard is for products that crossed the line into the world and created some kind of evidence.

That evidence can be tiny. A failed Product Hunt launch counts. A tool that got real users and then became too expensive to maintain counts. A product that revealed a better product counts.

The rule is simple: if the product taught something that should affect future decisions, it belongs in the archive.

Why keep it public

Public failure creates pressure to explain clearly. It also gives visitors a better read on the builder. Anyone can claim taste after a win. The better signal is whether the builder can look at a weak launch without turning it into either shame or mythology.

Madman Studio keeps the graveyard visible because the graveyard is part of the studio. The dead products are not a separate history. They are how the live products got sharper: Needles.AI’s last lesson — build on top of the frontier, not against it — is the operating rule behind everything live now. The working products sit on top of the model layer instead of fighting it, because a buried one already paid to learn that the other way loses.