You picked strangler fig. You hired great architects. You have executive support. And 18 months later, you're drowning - releases slower than before modernization started, teams working 60-hour weeks, senior people quitting.

This keeps happening because we've been borrowing from the wrong disciplines. Civil engineering. Manufacturing. Disciplines where steel properties and gravity are predictable. But software doesn't live in that world.

Software lives in systems that are complicated, with rigid code executing inside unpredictable, constantly-shifting business contexts. When you can't predict the future, requirements-driven design fails. Pattern-first thinking fails. And dual-mode operation becomes an exponential death spiral.

This talk shows you why modernization efforts die (hint: it's not bad patterns - it's bad assumptions), and what actually works:

• **Stressor analysis over requirements** - random simulation reveals what survives when business contexts shift
• **Residues over patterns** - designing for attractor movement, not today's abstractions
• **Team boundaries that follow architectural stress** - integrating Residuality Theory with Team Topologies
• **Cognitive load as architectural constraint** - dual-mode isn't "temporary inconvenience," it's where teams drown

You'll leave with red flags that predict failure, a framework that integrates architecture and organization, and a checklist for designing systems that survive when the business context inevitably moves.

Based on real enterprise engagements across European banking, retail, and logistics. No hand-waving. No "it depends." Just what works and what doesn't.