Data cleaning
AuraScore 81/100

Omnichannel Product Catalog Data Scrubbing Protocol

Standardize messy multi-vendor product attributes, sizing, and categories prior to e-commerce and ERP synchronization.

Use this checklist when onboarding new supplier product feeds or preparing retail master data for omnichannel catalog syndication. It ensures consistent attribute normalization, deduplication, and taxonomy compliance.

Template

Role: Senior Retail Master Data Quality Specialist with deep expertise in catalog harmonization and PIM system hygiene.

Context

  • Target Retailer: {{retailer_name}}
  • Upstream Data Feeds: {{catalog_source_systems}}
  • Master Taxonomy Model: {{target_taxonomy_standard}}
  • Mandatory Attribute Set: {{attribute_schema_fields}}
  • Duplicate Match Threshold: {{duplicate_threshold}}
  • Syndication Endpoints: {{primary_sales_channels}}

Task

Produce an actionable data cleaning and validation checklist to systematically audit, cleanse, and normalize multi-source raw SKU records into clean master product data for {{retailer_name}}.

Method

  1. Establish character encoding and whitespace stripping checks across all raw text fields in {{catalog_source_systems}}.
  2. Define regex patterns and lookup tables for standardizing brand names, department names, and category trees against {{target_taxonomy_standard}}.
  3. Verify unit-of-measure conversions across weight, dimension, and volume attributes to enforce uniform metric or imperial definitions.
  4. Design deterministic validation rules for {{attribute_schema_fields}} to identify nulls, placeholder strings (e.g., 'N/A', 'TBD'), and invalid data types.
  5. Establish exact match and fuzzy string distance criteria to flag potential duplicate SKUs using {{duplicate_threshold}}.
  6. Validate parent-child variant relationships (e.g., color, size matrix) to eliminate orphan variants and conflicting parent attributes.
  7. Formulate final quality gate criteria required before releasing cleaned product batches to {{primary_sales_channels}}.

Constraints

  • Check items MUST specify the cleaning action, target field, and pass/fail criterion.
  • MUST NOT permit automated deletion of records flagged as duplicate without an explicit manual audit status.
  • Include explicit rules for handling localized character sets and special punctuation in product titles.
  • Every checklist phase must feature an ownership assignment tag.

Output format

Return a structured Markdown checklist organized into four chronological sections: 1. Ingestion Sanitization, 2. Attribute Standardization, 3. Variant & Duplicate Reconciliation, and 4. Syndication Readiness Gate. Each section must contain 4-6 checked/bracketed items (- [ ]) accompanied by a brief verification note.

Self-review

  • Ensure all variables ({{retailer_name}}, {{catalog_source_systems}}, {{target_taxonomy_standard}}, {{attribute_schema_fields}}, {{duplicate_threshold}}, {{primary_sales_channels}}) are addressed.
  • Confirm every check item provides an unambiguous pass/fail test.
  • Verify the text contains no extraneous commentary outside the requested checklist.
AuraScore breakdown
81/100Provisional
Instruction clarity15/15 · Strong

Explicit role, a named task, and discrete steps the model can follow.

Context architecture12/12 · Strong

Background, inputs and variables the model needs before it starts.

Constraint engineering12/12 · Strong

Hard boundaries — what the model must and must not do.

Output specification6/14 · Thin

A named, field-level shape for the response.

Reasoning structure10/10 · Strong

Ordered work items that force analysis before an answer.

Model compatibility10/10 · Strong

Length and structure that travel across frontier models.

Token efficiency5/10 · Thin

Signal density — instruction weight without padding.

Reusability7/7 · Strong

Documented variables so the scaffold adapts to new inputs.

Robustness3/5 · Adequate

Quality bar, assumptions and behaviour when inputs are thin.

Observed performance1/5 · Thin

How much real usage the template has behind it.

data-analytics
data-cleaning
retail-consumer-goods
retail
product catalog
data cleaning