Every project that worked in my career had the same shape. It took me 20 years to see it clearly enough to name it.
I spent a week understanding the platform before writing code. The team had been building for months without asking why.
A multi-year rewrite had failed. I didn't identify the need — but when the greenfield approach collapsed, I mapped the full system and built the automation against constraints. It saved 10,000 man-days per year.
Nobody had asked why infrastructure costs were so high. I found it alone. Mapped the spend. Redesigned the allocation. The answer saved $360,000 per year — recurring. The automation I built became the platform default.
Different industries. Different decades. Different entry points. Same shape: the problem was unclear, misunderstood, or being solved wrong — and the method made it visible.
The shape had five parts. I saw them again and again.
Before touching anything, question the problem statement. Most problems are symptoms of deeper problems. Most teams skip this step entirely — they start building before they know what they're building.
Different people know different pieces of the problem. Nobody has the full map. Scope is where the pieces connect — and where the gaps everyone assumed were filled become visible.
Architecture and process as one system. The monitoring, the rollback plan, the process changes — all part of the deliverable. A good solution in a bad process still fails.
Not training. Working side by side. Soft handoff, not hard cutover. The "why" document, not just the "how" runbook. The person who takes over was part of the solution from the start.
The real test: did it hold after you left? At one company, the automation became the platform default — teams must opt out. At another, people freed up from the automation expanded into entirely new business areas. The best lift isn't just "it still works." It's "it changed what the organization does by default."
Every step that gets skipped has a cost.
Build the wrong thing. A team spent months solving a problem that wasn't the real problem.
Miss critical dependencies. A migration that looked simple involved hundreds of tenants.
Create debt that compounds. A performance fix without monitoring means the next failure is invisible.
Solution dies when you leave. A brilliant redesign only one person understands is a single point of failure.
No lasting impact. Just a line on your CV and a system that decays. The people didn't grow. The next problem is just as hard.
Nobody had asked why monitoring infrastructure cost so much. I found it alone. Mapped the spend across Prometheus. Redesigned the allocation. The savings are recurring — active as long as the platform is live. The automation became the foundation for ScaleOps integration. Both approaches now enabled by default — teams must opt out. 100% uptime during Super Bowl.
The owner knew there were losses. Not how much. Not where. His own employees couldn't tell either. They had tried building tooling themselves — it didn't catch the leak. He contracted me to find it. I scoped the full chain: game machines, ATM integration, cash flow, audit trail. Built observability and audit tooling that made cash visible for the first time. The answer: €20,000 per day, and now they knew exactly where. Built both business and technical processes. Platform runs without me. The observability layer became the default. Losses tracked permanently.
Different entry points. Different problems. Same five steps. Same shape.
I formalized QSAIL after 20 years of noticing this same shape. Across smart TVs, scientific instruments, enterprise SaaS, fintech, automotive, gambling.
The method came from both doing AND learning. 20 years of practice across 10+ industries. MIT Sloan — Management, Innovation, and Technology. Both matter. Practice alone is habit. Education alone is theory. Together they create credibility.
At some point I realized: I'm not following someone else's method. I built my own.
Most teams discover they skipped a step only after the project fails. If you already see it — that's the right time to talk.