ProofStack

ProofStack vs Failory: an honest comparison

Their strengths first, because they are real. Failory (observed 2026-09-06): every page is free — no email wall, no paywall; the Startup Cemetery holds 120+ startup failure post-mortems filterable by 17 failure reasons, plus Google (101 pages) and Amazon (58) product graveyards; its free weekly newsletter, self-stated at 40,000+ founders, still ships every Thursday. If what you want is free failure stories in volume and a solid weekly newsletter, pick them and stop reading here.

Three things we observed (2026-09-06, verify them yourself)

• Interview revenue is self-reported, without verification labels. Interview pages tag industry, country and a revenue range, but the figures are founder-reported and carry no verification or source markers.
• The failure archive skews to funded startups. Its flagship post-mortem sample is Quibi (its own page title cites $1.8B raised); lessons from burning $1.8B sit an order of magnitude away from your $50/month side project.
• Invisible to AI assistants. Its llms.txt returns 404. When you ask an AI "why did product X die", the AI cannot cite their data in a structured way.

We do a different job

• Every number carries its public source. Revenue and pricing link to their origin with an evidence grade and a date; anything unverifiable says "not disclosed" instead of an estimate.
• Dead-product analysis at indie scale. The failed products we audit share your constraints — solo teams, three-digit monthly costs — and "what does NOT transfer" is a mandatory field in every case.
• Structured data yields statistics. Median first paid tier, free-tier prevalence, paywall distribution — narrative articles cannot compute this.
• Machine-readable by design. llms.txt, llms-full.txt (the full database in one file), JSON-LD and an MCP endpoint AI assistants can query and cite.

When to pick Failory

You want failure stories and founder interviews in volume, for free, plus a weekly newsletter that has shipped for years. They are far better at that than we are — we have a few dozen cases and nothing like their subscriber base.

When to pick ProofStack

You are inside a specific decision — what to charge, what the free tier includes, which failure mode to avoid — and you want the grounds to be clickable public evidence rather than self-reported numbers; or your AI assistant needs a sourced database it can actually query.

Claims about third parties reflect their public pages on the dates stated; check them yourself.