Structured when I need it to be, flexible when I don't.

Every idea I work on goes through the same loop, whether it's a tiny fix or a whole new product. The difference is how much work each step gets. This page explains how I decide what to build, what to test, and what to walk away from.

The loop

This is my process for having structure when I need it and flexibility when I don't. I use Openspec to define how I want to tackle big, ambiguous projects as much as I want to tackle a simple bug fix.

• explore — sit with the idea until it makes sense
• propose — write down what I'm building and why
• apply — build it
• archive — retro and documentation

The tactics

Part of discovery is collecting data, and I'm borrowing tactics from all my eras here: business metrics, automated test coverage, low fidelity prototyping, generative research, etc. All of them give me foundations for data-driven decisions about what's a good idea, and what's just the idea I have to try before moving on.

Having ADHD, however, means I reserve the right to just make something because its fun or I’m curious.

I use scoring to accommodate this really wide aperture of joy and curiousity. Each experiment is evaluated across five dimensions that I use to inform what I work on next — not perfection.

The stack

A few of my first experiments have become core elements of the workflow:

• MVDS — An opinionated design system
• Experiment Hub — a scaffold for startup experimentation