Qianchuan Viral Creative Teardown

Problem to solve

Short-video creative diagnostics often stop at vague opinions: a low CTR may mean a weak first 3 seconds, low completion may mean a dragging middle, and low CVR may mean the selling point is not structured clearly. This skill turns a video asset into a comparable analysis object by first confirming the goal—livestream traffic or product-linked conversion—then combining creative content with delivery metrics.

How it works

• Inputs: video file, optional delivery data such as spend, impressions, CTR, CVR, ROI, and completion rate, plus the target objective.
• Extraction: captures verbatim narration, shot descriptions, on-screen text, and audio cues; when transcription is needed, it can use ffmpeg to extract the audio track.
• Analysis: follows a seven-part framework and applies different standards for traffic-driving creatives versus conversion creatives, such as suspense for livestream entry versus result benefits for purchase intent.
• Output: produces a fixed Markdown report with a shot table and actionable conclusions on strengths, traffic drivers, and optimization.

Boundaries

Without an explicit objective, differentiated judgment is limited. Without delivery data, the report can analyze content only and cannot fully infer traffic causes. Accuracy depends on readable video content or usable transcription, and very short clips, dialects, or noisy audio may reduce extraction quality.