The reveal trap
The most dangerous project shape is the long silence followed by the big reveal. Six weeks of assumptions, presented all at once, defended all at once. When it's wrong — and first versions are always partly wrong — the rework costs more than the build.
The thin slice
Instead, I ship the smallest end-to-end path in week two: real data, real users, one workflow. It's ugly and it's honest. Stakeholders react to something real, and every subsequent week thickens a slice that's already proven.
- ✓One workflow, fully working — before the second workflow starts.
- ✓Weekly demos — no status decks, just the product.
- ✓Cut visibly — a parking lot list everyone sees, so scope cuts are decisions, not secrets.
What gets cut (and kept)
Cut admin polish, exotic edge cases, and anything describable as "while we're at it". Keep auth, validation, monitoring, and the rollback path — the unglamorous parts that let you ship weekly without fear.
The compound effect
Teams that ship small learn fast: which features matter, which estimates lie, which users to listen to. After three months they haven't just built software — they've built judgment. That's the asset that outlasts any single release.