Our moat is showing our work.

Every autopsy draws on a fixed set of primary sources:

• Wayback Machine snapshots of the company's landing page across its life — we look at what the pitch was before revisionism set in.
• Hacker News threads at launch, funding, and shutdown — comment volume and sentiment give us founder-community signal.
• TechCrunch / public press for funding timelines and on-record statements.
• Y Combinator directory for batch, status, and founding team for the YC-indexed 1,000+ entries.
• SEC + state filings for dissolution, last-activity dates, and litigation.

We do not use Crunchbase Pro, LinkedIn Phantombuster, or any paid database in the free and Pro tiers — those cost is passed on only at the Founder/Scout levels. The baseline product is buildable on public data.

The score is the weighted sum of five factors, each rated 0-10:

• Market TAM Trajectory (weight 0.25) — is the underlying demand still growing, flat, or dead?
• Tech Gap Now vs Then (weight 0.25) — what existed in 2026 that didn't exist when the company died?
• Capital Efficiency (weight 0.20) — could a solo or small team ship a working version at 1-2 orders of magnitude less capital than the original company?
• New Distribution Channels (weight 0.15) — are there channels (TikTok, Substack, MCP, app stores) that didn't exist before?
• Solo Founder Fit (weight 0.15) — can one engineer realistically build this, or is domain expertise a moat only incumbents can cross?

Weighted sum, rounded to one decimal. Interpretation bands:

• 0-4: Structurally bad. Negative signal — don't rebuild.
• 4-6: Partial signal. One narrow angle may work.
• 6-8: Worth rebuilding with a specific twist. See the Angle Still Open section.
• 8-10: Red flag. Someone is probably building this right now.

Every entry carries one of three badges:

• Deep — multiple primary sources (Wayback + HN + press) and confidence ≥ 0.75. This is the flagship quality.
• Standard — one or two primary sources, confidence 0.55-0.75. Readable and useful, but flagged as partial.
• Seed — no primary sources beyond the company name and public summary. Confidence < 0.55. These entries carry a "noindex" tag to avoid thin-content SEO penalty, and they exist to be upgraded on demand.

If the autopsy you care about is marked Seed, you can order a $99 Deep Dive and we'll upgrade it inside 48 hours. That constraint is part of the product, not a shame.

Graveyard is not a predictive model. The Revival Score does not tell you whether a specific company you build will succeed. It tells you whether the category has been re-opened by outside forces (new tech, new distribution, new regulation) since the original company died.

We do not scrape founders' personal data beyond what is already public, and we do not publish material that could ground a defamation claim. If you are a founder who has appeared in Graveyard and want a page corrected or removed, email us and we'll respond within 24 hours.