sqp-priority-rank - SKILL.md Agent Skill

name: sqp-priority-rank description: Score and rank an Amazon ASIN portfolio by CRO opportunity using SellerApp keyword data, BSR, and reviews. Wraps and extends the existing `cro-priority-tracker` skill. Use when running `sqp-priority-rank` on a portfolio, when deciding which ASIN gets research budget next, or when a client has 20+ ASINs and needs triage. Defaults to the Keplo scoring formula: Data Strength 40 / $ Impact 30 / Effort 15 / Speed 15.

sqp-priority-rank — Portfolio CRO Opportunity Scoring

When a client has 20-100 ASINs, you can't optimize all of them. This skill scores each on opportunity size + data confidence and outputs a ranked work order.

Companion to existing cro-priority-tracker skill. This version pulls SellerApp data automatically; the existing one accepts uploaded SQP/Ads/Business Reports CSVs. Use both together when you have first-party SP-API data + want to enrich with SellerApp keyword/competitor signals.

Invocation

sqp-priority-rank {ASIN1,ASIN2,...}
sqp-priority-rank --portfolio={path-to-asin-list.csv}
sqp-priority-rank {ASIN list} --weights="data:40,impact:30,effort:15,speed:15"   # tweak weights

Output

/tmp/cro-research/portfolio-priority-{date}.md

Tools Used

n8n MCP wrapper.

Step	Tool	Purpose
Per ASIN baseline	`Get_Product_Details` (ratings + price + potential)	BSR, sales estimate, rating count
Keyword landscape	`Keyword_Research_V2_Reverse_ASIN`	Top keyword volume + relative_score
Quality gate	LQI flag (optional)	Section weakness for effort estimate

Phase 1 — Pull (batched)

Get_Product_Details accepts up to 20 ASINs/call — chunk the portfolio. Reverse-ASIN goes one at a time, batch where possible. Token-aware (per cro/sellerapp-api.md).

Phase 2 — Score per ASIN (4 dimensions)

Dimension	Weight	Inputs	Calculation
Data Strength	40%	Review count, sample size in keywords, BSR stability	High when ≥100 reviews + ≥5 high-vol keywords
$ Impact	30%	Sales estimate × delta-CVR potential	Higher BSR rank × higher current CVR gap = larger impact
Effort	15%	LQI section weaknesses, image count, copy quality	Inverse — low effort = high score
Speed	15%	Listing changes only, no MOQ/photoshoot needed	High when AI-generatable (fits Higgsfield)

Total score 0-100. Tag ranking confidence (High / Med / Low).

Phase 3 — Output

# Portfolio Priority Ranking

**Date:** {date} | **ASINs scored:** {N} | **Weights:** {weights}

## Top 10 — Work Order

| Rank | ASIN | Title | BSR | Sales/day | Score | Confidence | Recommended next skill |
|------|------|-------|-----|-----------|-------|------------|------------------------|
| 1 | B0... | ... | #4,123 | ~250 | 87 | High | `asin-deep-research` then `main-image-pipeline` |
| 2 | ... | ... | ... | ... | ... | ... | ... |

## Score Breakdown — Top 5

### 1. {ASIN} — Score 87
- Data Strength: 38/40 — {N reviews, K high-vol keywords}
- $ Impact: 26/30 — {sales estimate, CVR gap}
- Effort: 13/15 — {LQI weaknesses, what to fix first}
- Speed: 10/15 — {needs photography or AI-only}

## Bottom 10 — Defer

{Same table — these are low-priority, may not be worth research budget}

## Combined View (if Sessions/CVR data uploaded)

If client provided SQP / Business Reports, layer in actuals via `cro-priority-tracker` and re-rank.

Reference Files

Existing skill: cro-priority-tracker (companion — handles SP-API/CSV inputs)
~/.claude/knowledge/cro-methodology/decision-framework.md (scoring formula source)
~/.claude/skills/cro/data-analysis.md
Vault: CRO-Knowledge-Base/04-data-analysis/business-reports.md

Quality Bar

Top 10 + Bottom 10 both shown (don't just list winners)
Each top-5 has score breakdown by dimension
Recommended next skill is specific (skill name + ASIN)
Confidence tagged (low confidence = caveat the ranking)

Common Mistakes

❌ Ranking by sales alone — high sales + healthy CVR = no opportunity
❌ Ignoring effort/speed — best-scoring ASIN may need a photoshoot you don't have
❌ Using SellerApp data when client has cleaner SP-API data — use both, layer

Auto-Triggers

New client onboarding (after listing-quality-audit)
Quarterly portfolio review
User asks "which ASIN should I work on next" / "rank my catalog"