The Red-Day Checklist.

🖨 Print / save as PDF

• 1 · Confirm the drop is real. Compare today against the trailing 3- and 7-day windows, not just yesterday. One noisy day on thin spend is not a crash.
• 2 · Check the funnel before the ads. Place a test order. Checkout working? Site speed normal? Payment gateway alive? A funnel problem dressed as an ad problem is the most expensive misdiagnosis.
• 3 · Check tracking. Pixel firing, events landing, no attribution window change, no iOS/platform update mid-week.
• 4 · Check stock and offer. Hero product in stock? Price or shipping changed? A dying offer looks exactly like dying ads.
• 5 · Diagnose the type. Over-crowding (too many ad sets per audience)? Over-scaling (budget pushed too hard)? Faded creative (hero declining for days)? Or a bad prior rebuild?
• 6 · Cut the bleeders — today. Anything clearly below the account's floor gets paused now. No mercy days.
• 7 · Protect the engine. Do NOT touch campaigns that are still at or near their normal ROAS. Never cut the proven engine >30% on a red day.
• 8 · Rebuild forward, not backward. Fresh campaigns beside the wreck — never reopen crashed entities into learning with broken signal.
• 9 · Stack the rebuild with proven material only. Dormant past winners + still-healthy creatives recreated fresh. Zero new concepts on a red day — they appear in zero successful recoveries.
• 10 · Log the crash. Type, cause, fix, result. Every crash that becomes a rule makes the account harder to break.

Where this checklist comes from

This isn't theory — it's the compressed version of a year of real firefights on a live DTC account: 11 crashes, each diagnosed, fixed and documented, with a median recovery of 1–2 days (including a 1.62x rock-bottom day recovered to 4.92x in a single day). The full method, with the numbers behind every rule, is here: How I recover a Meta ROAS crash in 1–2 days →