The principles and frameworks I use to turn complexity into product decisions - before a single line of a PRD gets written. Click a module to expand.
How I move from chaos to a usable problem statement - before proposing any solution.
Why understanding the workflow matters more than collecting a feature wishlist.
How I choose what to fix first when everything feels broken and urgent.
Why rollout, training, trust, and launch readiness are part of product - not "operations' problem."
How I identify where AI actually improves decisions and communication - not where it's just novel.
How I balance speed, trust, and controls - especially in regulated or high-stakes environments.
Product sense isn't just retrospective - it's how I evaluate a product I don't own. Here's a real, unsolicited teardown of Salesforce Flow Builder, an enterprise workflow-automation tool, using the same lens from Module 02 and 04.
Flow Builder gives admins a visual, no-code way to automate business processes across Salesforce objects - genuinely lowering the barrier for non-engineers to ship real automation, which is the right instinct for an enterprise platform.
Debugging a failed Flow in production is opaque - error messages point at symptoms, not root cause, and there's no built-in "replay with sample data" for admins to safely test edge cases. That is exactly the adoption-killer I flag in Module 04: a system people learn to route around, not trust, the moment it fails silently once.
Ship a "Flow Debugger Replay" mode: capture the exact record state that triggered a failure, let admins re-run the Flow step-by-step against that captured state in a sandboxed view, and surface the failing condition in plain language, not a stack-trace-style error. Success metric: reduction in support-ticket time-to-resolution for Flow failures, not just fewer failures.
Building deep replay tooling competes for the same engineering capacity as new Flow actions/connectors. I'd argue reliability tooling wins first: every enterprise workflow tool eventually loses trust to opacity before it loses trust to missing features - this is the same governance-over-speed tradeoff from Module 06, applied to a product I don't own.
This is an unsolicited, outside-in teardown based on public product experience - not confidential information from any employer.
Beyond the live teardown above, I regularly run myself through classic PM interview prompts as deliberate practice - the same way an engineer drills coding problems. These three are self-practiced exercises, not client deliverables, but they show the same reasoning muscle: structuring ambiguity, picking a lens, and being explicit about what I'm assuming. Full write-ups are downloadable below.
I anchored the analysis in Microsoft's actual mission - empowering people and organizations to achieve more - so the acquisition target had to earn its place by expanding real use cases for existing users, not just be a recognizable name. I scoped the analysis to a few major Microsoft business lines, deliberately left finance and regulatory diligence out of frame, and argued explicitly that the "best" target is not the most famous one - it's the one with the tightest strategic fit.
Download full write-up (.docx) ↓I started by naming Spotify for what it actually is - a two-sided platform of listeners and creators, not just a listening app - then built a funnel-style metrics framework that tracks both sides in parallel instead of defaulting to a listener-only view. The write-up breaks metrics out by funnel stage, starting with acquisition (downloads, CAC, conversion to activation), and checks whether growth on one side is compounding growth on the other.
Download full write-up (.docx) ↓I picked Fitbod deliberately because it's genuinely both for me - it removes workout-planning effort and personalizes programming around goals, equipment, and history, which makes it practical for a schedule that changes daily. But I hold real friction with it too, and the write-up walks through that tension in detail and where I'd take the product from there, rather than only listing praise.
Download full write-up (.docx) ↓