noique/cross-border-ecommerce-skills/backlink-kol-extractor

Backlink KOL Extractor

Extract KOL/influencer/media prospects from Semrush competitor data using a three-step methodology: domain pattern matching, cross-competitor validation, and social media handle extraction.

Quick Start

Run the extraction script on a directory of Semrush xlsx exports:

python3 scripts/extract_kol.py <input_dir> [--output results.csv] [--min-sources 2] [--min-traffic 500]

Input: Directory with Semrush xlsx files organized by domain subdirectories:

input_dir/
  competitor1.com/
    competitor1.com-backlinks.xlsx
    competitor1.com-backlinks_refdomains.xlsx
    competitor1.com-organic.Competitors-us-*.xlsx
  competitor2.com/
    ...

Output: CSV with columns: domain, category, priority_score, traffic, backlinks, source_count, sources

Workflow

Step 1: Gather Data

Ensure Semrush data is downloaded for 3+ competitors in the target niche. Key files:

• *-backlinks_refdomains.xlsx — Primary source (which domains link to competitor)
• *-backlinks.xlsx — Secondary (specific URLs, can extract social handles)
• *-organic.Competitors-*.xlsx — Tertiary (traffic overlap sites)

Step 2: Run Extraction

python3 scripts/extract_kol.py ~/Downloads/semrush/_project_tactibeaver/

The script scans all xlsx files recursively, categorizes domains into KOL/Review, Media/Press, Forum/Community, Affiliate/Deal, cross-validates by counting how many competitors each domain links to, scores by priority, extracts social media handles from backlink URLs, and outputs sorted CSV.

Step 3: Interpret Results

Priority tiers:

• Score 70+: Core industry sites — prioritize for outreach
• Score 50-70: Important channels — include in outreach list
• Score 30-50: Worth monitoring — add to watchlist

Step 4: Expand via Linktree

For domains with linktree URLs found, visit linktr.ee/{username} to discover full social media presence.

Advanced Usage

Custom thresholds:

python3 scripts/extract_kol.py ./data --min-sources 3 --min-traffic 1000

Methodology Details

See references/methodology.md for category detection keywords, scoring formula, social media URL patterns, and filter rules.