← ClaudeAtlas

kai-retargetlisted

Design retargeting and remarketing campaign architecture across platforms — audience segmentation, creative strategy, frequency caps, and platform-specific setup with ad policy compliance. Use when "retargeting", "remarketing", "retarget", "re-engage visitors", "abandoned cart", "pixel setup", or any request to bring back visitors who didn't convert.
cgallic/kai-cmo-harness · ★ 47 · AI & Automation · score 73
Install: claude install-skill cgallic/kai-cmo-harness
## Objective A retargeting architecture that is ready to build in the ad accounts: audience segments defined by intent and recency, creative matched to each segment's intent level, frequency caps, exclusion rules, sequence timing, budget split, and platform-compliant ad copy. The plan states which pixel events and URL rules define each audience, so someone can implement it without guessing. Retargeting fails in two ways — showing the same ad to everyone regardless of how close they got, and showing it so often that it burns the brand. Segmentation and frequency caps exist to prevent both. ## Done when Work type `paid-ad-campaign` — floor **E5/C4/O4** (`harness/eco-floors.yaml`), `spend_authority: true`. - **E5** — the platform returns ad object ids and a read-back of the live entities confirms targeting, budget, schedule, and creative match the approved bundle field for field. **Hard stop: live-account mutation without recorded human approval is never SHIPPED, whatever the rest of the evidence shows.** - **C4** — every platform's policy reference was loaded before the ad was written and the ad was checked against it; Four U's at **10/16** (ad threshold), `banned_word_check` clean, `mutation_risk_lint` clean; frequency caps set per segment; converters and exclusion lists verified in place. Max 2 auto-retry cycles, each naming the specific failure. - **O4** — CAC, CPL, ROAS, or qualified-lead rate against a threshold declared before launch, read from the ads connector no e