objection-video

name	objection-video
description	Run a ProductPinion Pinion Video using the Objection Test template — shoppers view the listing and answer "would you buy this? why/why not?" out loud. Surfaces hidden objections that text-based reviews don't always reveal. Use when running `objection-video {ASIN}`, when CVR is below category average, or as a sub-step of `objection-killer`.

objection-video — Recorded Shopper Objections

Pinion Videos record shoppers viewing your listing and answering questions out loud — including objections they'd never write in a review. The Objection Test template is purpose-built for this.

Prerequisites

• ProductPinion MCP
• Live Amazon listing URL

Invocation

objection-video {ASIN}
objection-video {ASIN} --shoppers=15        # 15-25 default; videos cost more credits
objection-video {ASIN} --questions="..."    # override default questions

Output

/tmp/cro-content/{ASIN}-objection-video-{date}.md — transcripts, themes, action items.

Phase 1 — Configure Pinion Test

Test type: Pinion Videos → Any URL with Objection Test pre-defined template.

Setup:

• URL: live Amazon listing (us / geo)
• Internal name: {ASIN}-objections-{YYYY-MM-DD}
• Questions (default Objection Test template):
  1. "Walk me through this listing — what catches your eye?"
  2. "What concerns or doubts do you have about this product?"
  3. "Would you buy it? Why or why not?"
• Sample: 15-25 shoppers (videos are higher cost than polls)
• Audience: research-brief

Phase 2 — Wait

Pinion Video tests run longer than polls — typically 24-48 hours. Skill exits after submission, returns Pinion test URL, and instructs user to re-invoke when ready.

Phase 3 — Analyze

Pull:

• Recorded videos
• Auto-generated transcripts
• Pinion's auto-summary (if available)

Decompose objections per ~/.claude/skills/cro/review-analysis.md 8-layer framework:

• Aspect (what part of product)
• Sentiment + modifier
• Whether the listing already addresses it (is it visible / hidden?)
• Frequency across the 15-25 shoppers

Phase 4 — Output

# Objection Video — {ASIN}

**Sample:** {N} shoppers | **Date:** {date}

## Top Objections (ranked by frequency)

| Rank | Objection | % of shoppers | Listing currently addresses? | Recommended fix |
|------|-----------|----------------|-------------------------------|------------------|
| 1 | "Not sure if it's safe" | 60% (12/20) | ❌ | Slot 4 trust-signal image |
| 2 | "Looks small" | 45% | Partial — bullet only | Slot 3 scale comparison |
| ... | ... | ... | ... | ... |

## Verbatim Highlights (under 15 words each, in quotes)

> "{quote}" — {shopper-id}, {context}

## Patterns

- **Hidden objections** (not in reviews): {list — these are gold; reviews are post-purchase}
- **Confirmed objections** (also in reviews): {list — validation of mining}

## Recommended Next Skill

`objection-killer {ASIN}` — full pipeline: review-mine + this output → preempt-image generation → A+ FAQ refresh

Reference Files

• ProductPinion docs — Pinion Videos + Objection Test template
• ~/.claude/skills/cro/review-analysis.md
• Vault: CRO-Knowledge-Base/01-research/review-mining.md

Quality Bar

• At least 15 shoppers (statistical floor for video tests)
• Objections ranked by frequency
• "Hidden vs confirmed" split — the hidden ones are the high-leverage finds
• Verbatim quotes under 15 words each

Auto-Triggers

• objection-killer Phase 1
• User says "why aren't they buying" / "find hidden objections"
• After CVR drop investigation