Ship small, then ship again

We keep a private list of the releases we regret, and it has one thing in common from top to bottom: size. Every entry was a big bang — months of work, a branch with its own weather system, a launch day that needed a countdown. Nothing on the list is small. That is not a coincidence, and it took us embarrassingly long to treat it as data.

Big releases hide; small releases tell

A large release is a warehouse of assumptions that all get tested on the same afternoon. When something breaks — and something breaks — the failure is an archaeology project. A small release carries one or two assumptions, so reality grades it overnight and the feedback still has your fingerprints on it. You know exactly what changed, because almost nothing did.

The compounding effect is what surprised us. Shipping weekly does not just find problems earlier; it changes what you build. Work gets designed to be shippable in slices, which forces the honest question every slice must answer: what does the user get this week? A feature that cannot answer that for six consecutive weeks is usually not a feature. It is a hope.

Cutting scope is not cutting corners

The pushback is always the same: users deserve the whole thing. They do — eventually. The distinction we hold is that scope is negotiable and quality is not:

• Cut scope by shipping the import before the export, one integration before nine, the happy path before the power tools.
• Cut corners and you ship the whole surface with thin ice everywhere — no error states, no empty states, a spinner where the care should be.

Scope cuts are visible and honest; a user can see the edge and ask for more. Corner cuts are invisible until someone falls through, at two in the morning, in production.

Ship the staircase one solid step at a time. Nobody has ever fallen through a step that existed.

The calm is the point

The unadvertised benefit of shipping small is emotional. Launch days used to be adrenaline events here — pizza, dashboards, a rollback script open in a second terminal. Now shipping is boring, in the way a good bridge is boring. The nervous energy went somewhere better: into the work itself, a week at a time.

We still do big things. We just no longer do them in one move. The cathedral goes up stone by stone, each stone inspected by daylight — and the scaffolding never becomes the story.