Liquor Product Matching Specialist
Budget: $30 – $250 USD
Product Matching Task – Business Rule Summary
We are building a master product catalogue for a liquor retail comparison platform. The goal is to match each master product to the best-matching item from supplier catalogues.
Scope:
• Approx. 1800 master SKUs
• 6 suppliers (each with 1,000–2,000 SKUs)
• One-to-one mapping per supplier
Rules for a Valid Match:
1. Exact Volume Match: e.g., 750ml = 750ml, 1.5L = 1500ml
2. Exact Unit Count Match: e.g., 1 x, 6 x, 12 x
3. Brand Match: First token (e.g., “KWV”, “Alvi’s”) must match else normalised token (ASCI)
4. Varietal Match: e.g., “Chardonnay”, “Cabernet Sauvignon”
5. No Duplicates: A supplier SKU may only be used once
Soft Matching Logic:
• Token overlap is used to determine closeness of match
• Choose the supplier SKU with the most word token overlap after meeting all hard conditions
Keyword Matching Enhancements:
• Normalize special characters (e.g., cuvã©e → cuvée)
• Remove punctuation
• Match additional keywords like:
o Spirits: Peated, Reserve, 10yo, VS, XO
o Wine: Signature, Nicolas, Estate
Required Outputs:
• Final matched file: includes original master + best match + supplier price
• Adding unmatched items to master
• Generic name and Unique ID for mapping line
We are building a master product catalogue for a liquor retail comparison platform. The goal is to match each master product to the best-matching item from supplier catalogues.
Scope:
• Approx. 1800 master SKUs
• 6 suppliers (each with 1,000–2,000 SKUs)
• One-to-one mapping per supplier
Rules for a Valid Match:
1. Exact Volume Match: e.g., 750ml = 750ml, 1.5L = 1500ml
2. Exact Unit Count Match: e.g., 1 x, 6 x, 12 x
3. Brand Match: First token (e.g., “KWV”, “Alvi’s”) must match else normalised token (ASCI)
4. Varietal Match: e.g., “Chardonnay”, “Cabernet Sauvignon”
5. No Duplicates: A supplier SKU may only be used once
Soft Matching Logic:
• Token overlap is used to determine closeness of match
• Choose the supplier SKU with the most word token overlap after meeting all hard conditions
Keyword Matching Enhancements:
• Normalize special characters (e.g., cuvã©e → cuvée)
• Remove punctuation
• Match additional keywords like:
o Spirits: Peated, Reserve, 10yo, VS, XO
o Wine: Signature, Nicolas, Estate
Required Outputs:
• Final matched file: includes original master + best match + supplier price
• Adding unmatched items to master
• Generic name and Unique ID for mapping line