THESIS: E-commerce brands lose money on creative at three moments, and almost everyone only guards the last one. Before production — weak concepts get filmed because nobody filtered them on paper, and every one costs a creator fee, an edit, and a test budget to learn what a 10-minute checklist would have told you for free. During the run — fatigue announces itself in the leading indicators days before CPA moves, but most teams watch CPA, which is the coroner's report, not the diagnosis. At replacement time — the winner dies and the next batch is still being filmed or approved, so the team keeps spending on a corpse because no bench exists. This checklist guards all three: a pre-production filter that kills weak concepts at zero cost, the early-warning signals that fire before the money burns, and the variation-and-pipeline discipline that makes fatigue a rotation instead of a crisis.
CONTEXT: The 2026 stakes: creative is now ~56% of incremental conversions (Meta internal study) — a bigger lever than audience or bid — while the fatigue clock collapsed from 9-day to 3-day refresh cycles and top-quartile teams ship 80–120 variants a month. In that market, producing an unfiltered concept is the single most expensive habit in performance creative: the industry's own benchmark says validate 20–30 hook variants cheaply before commissioning production, and magnify only the 3–5 that earn it. Run Section A on every concept before it's filmed; Sections B–D weekly on everything that's live. Verdict rule at the end.
═══
Score each concept 0–2 per check before anything is filmed. A concept under 8/12 doesn't get produced — it gets fixed or killed on paper.
CHECK 1 — The 3-second silent test
CHECK 2 — The reason-to-care is the buyer's, not the brand's
CHECK 3 — A share/stop trigger exists
CHECK 4 — Native shape on paper