Products are built through hundreds of decisions. Here are the ones that mattered most — what I chose, why I chose it, and what I gave up to get there.
14 of 14 decisions
The scoped chat assistant cannot write to the database. It can only emit structured proposal cards (create_story, edit_prd_section, etc.) that the user must explicitly Apply.
Stored complex artifacts as JSON-in-TEXT columns, parsed at read time. Versioning was added only on PRDs, where the stakes are highest.
Fixed the landing copy first, then staged the real privacy work (auth → envelope encryption → optional local-model provider) as separate, deliverable phases.
Delayed real-time features to v2, shipped with 15-second delayed quotes in v1.
Built native iOS first, Android later.
Shipped with 7 core features, parking lot for rest.
Built custom notification system.
Launched completely free, added premium tier after 6 months.
Sunset the Daily Health Tip feature after 3 months' notice.
Launched without A/B testing, added it 3 months post-launch.
Turned down enterprise deals, stayed focused on SMB.
Prioritized building an automated onboarding flow instead of scaling support headcount.
Scoped recurring patterns to the most common use cases first and deferred edge cases.
Redesigned permission flows and visibility logic to reduce ambiguity and edge-case errors.