A — Assess the Metric (decide the next move)

Don’t “add anyway.” Decide with a number.

Target vs Actual (example)

Target: ≥60% “useful” in 24h
Actual: 42% (n=19)
Decision: FAIL → tighten Job Story and landing
Next experiment: clarify input constraints; rename primary button

Decision tree

• PASS → add one adjacent feature (history or export).
• FAIL → fix one thing this week: value prop | tighter job | UX path | prompt/model.

Why binary helps

• Prevents “maybe later” creep.
• Makes scope predictable.
• Teaches you where leverage is (copy? prompt? flow?).

One‑pager update checklist

• Keep the Not‑Doing list intact.
• Write what you’ll test next (one sentence).
• Set a new metric or same metric with a clearer path.

Examples + rubric → https://shipwithai.substack.com/p/pasta-al-dente-ship-a-one-feature?utm_source=devto&utm_medium=social&utm_campaign=issue_pasta_al_dente_week