I've founded and closed multiple companies over 14 years. Each time, hard-won lessons just disappeared into nothing. Medicine advanced when doctors started doing systematic autopsies. Aviation safety improved after black boxes became mandatory. But for organizations? 90% of startups fail, and we collect almost no structured data about why. SOIL collects "autopsy" data from founders of dead companies - functional state at peak, financial trajectory, timeline of events, environmental factors, founder context. Framework-agnostic: we gather data first, then analyze through multiple lenses to see which theories actually predict failure. The goal: build enough cases to identify patterns, early warning signals, eventually predictive models. Each company gets a digital memorial (cenotaph) - you can browse existing ones without signing up. Looking for the first 100 founders willing to document their dead companies. Feedback on the approach welcome.