What worked cutting scope to ship in a week

Our roadmap had seventeen items labeled "must have for launch." We shipped four. Login, core workflow, export, billing. Everything else waited. The trick was defining launch as "one user can complete the job start to finish"—not "feature parity with competitors." We wrote that sentence on a sticky note and rejected anything that didn't serve it. Shipping ugly-but-complete let us learn from real usage. Two of the cut features were never built because nobody asked. Three came back smarter after we understood the workflow.