In January 2024, our product had 847 configurable settings. We had shipped features at a relentless pace for a year and a half — dashboards, automations, integrations, templates — each one justified by a user request or a competitor's launch. But when we examined usage data, a stark picture emerged: 73 percent of those features were touched by fewer than 4 percent of users. The changelog read like a novel nobody finished.
The Unshipping Process
We started by mapping every feature to a simple grid: effort to maintain against user impact. Some of our most-requested features had become the most expensive to support, not because they were complex, but because they existed in tension with newer capabilities we had layered on top. Over three months, we deprecated 41 features, consolidated 12 overlapping workflows into 4, and replaced 200 configuration options with sensible defaults. The result was not a stripped-down product. It was a focused one.
The best product decisions are often invisible — they're the ones where something quietly disappears and nobody notices, because the path forward became clearer.
Our support tickets dropped 34 percent in the quarter that followed. New-user activation rose by 18 points. The product felt lighter, not because we had removed capability, but because we had removed the friction of choosing between things that barely differed.