Closed loop
Creative Learnings
Short, business-scoped lessons from past ad performance. Execution learnings steer the matching variant category on the next generate. Structural learnings are for mix review only — they are never forced into a mismatched slot prompt.
How this works
- Scoped to this business only.
- Written only when data clears a strict promotion bar (not day-1 noise).
- Generation receives 0–2 advisory bullets per category slot.
- Campaign brief always beats learnings if they conflict.
Counts
- Execution (gen-safe)
- 0
- Structural (mix only)
- 0
- Active total
- 0
Execution learnings
Injected only into the matching creative category on generate.
No execution learnings yet. They appear after evaluate finds a clear same-category winner with enough volume.
Structural learnings
Cross-category results (e.g. lifestyle slot beat pain slot). Shown here for review — not injected into a pain-slot generation prompt.
No structural learnings yet.