The most important product decision is what you don't build

The central claim is that the most consequential product decision is deciding not to build. Concrete examples are a document hub and a notifications centre: superficially simple features become sprawling platforms once stakeholders ask for tagging, archiving, multi-channel sharing, security rules and compliance handling for legacy PDFs and persistent communications. Those additions not only balloon initial development but create long-term maintenance burdens - platform fragility across OS updates, ongoing authentication and permission work, and continual feature creep that turns a narrow need into a Google Drive or Gmail clone.

The remedy is disciplined subtraction. Make decisions by mapping each communication use case to what actually needs to be said and when, prefer simple or existing mechanisms over bespoke platforms, and show long-term running costs rather than just build costs to shift stakeholder incentives. Cultural friction matters: organizations reward creation, not deletion, yet Gerry McGovern’s Top Tasks and experiments by Klotz et al. show removal often improves outcomes. With cheap execution today, resist additive reflexes; use automation (agents) to prune and maintain, and concentrate engineering talent on a few high-impact products à la Jobs’ 1997 focus. Be choosy about additions and militant about removal so the product truly serves customers rather than the organisation.