The Three Data Layers
AI product recommendations rely on three layers of data, each serving a different purpose:
Identity Data
What is this product? Names, codes, categories.
Essential for all products
Specification Data
What are its characteristics? Dimensions, materials, performance.
Required for filtering
Context Data
Where and how is it used? Applications, environments, use cases.
Enables smart matching
Universal Required Fields
Every product, regardless of category, needs these fundamental data points:
| Field | Description | Example |
|---|---|---|
| product_name | Clear, descriptive name | "Milano Lounge Chair" |
| sku | Unique product identifier | "MIL-LC-001" |
| brand | Manufacturer/brand name | "DesignCo" |
| category | Primary product category | "Seating > Lounge" |
| description | Short product description | "Contemporary lounge chair..." |
| primary_image | Main product image URL | "https://..." |
| price | Price or price range | "1500" or "1200-1800" |
| currency | Price currency | "EUR" |
Category-Specific Fields
Different product categories require additional specific attributes. Here are the key fields by category:
Seating (Chairs, Sofas, Benches)
Dimensions
- • width_cm, depth_cm, height_cm
- • seat_height_cm
- • seat_depth_cm
- • armrest_height_cm (if applicable)
Specifications
- • frame_material
- • upholstery_material
- • weight_capacity_kg
- • stackable (boolean)
- • with_armrests (boolean)
Tables (Dining, Conference, Coffee)
Dimensions
- • length_cm, width_cm, height_cm
- • diameter_cm (for round tables)
- • seating_capacity
- • extended_length_cm (if extendable)
Specifications
- • top_material
- • base_material
- • shape (rectangular, round, oval)
- • extendable (boolean)
- • cable_management (boolean)
Lighting
Technical
- • lumens
- • color_temperature_k
- • cri_rating
- • wattage
- • dimmable (boolean)
- • light_source (LED, halogen, etc.)
Physical
- • fixture_type (pendant, sconce, etc.)
- • drop_length_cm (for pendants)
- • ip_rating
- • mounting_type
- • shade_material
Textiles & Fabrics
Performance
- • martindale_cycles
- • pilling_rating
- • lightfastness
- • fire_rating
- • cleanability_code
Composition
- • fiber_content (with percentages)
- • weave_type
- • weight_gsm
- • width_cm
- • pattern_repeat_cm
Context & Application Tags
Beyond specifications, AI needs to understand where products are used. Include these context fields:
Environment Tags
Style Tags
Certifications & Compliance
Professional buyers filter by certifications. Structure these as searchable fields:
Certification Fields
- • fire_rating
- • greenguard_certified
- • fsc_certified
- • leed_points
- • cradle_to_cradle
- • bifma_certified
- • ada_compliant
- • ce_marked
- • recycled_content_pct
Data Quality Checklist
Before submitting data, verify:
- All required fields are populated (no blanks)
- Dimensions use consistent units (cm, not mixed)
- Image URLs are valid and images are high-quality
- Categories follow standard taxonomy
- Prices are current and in correct currency
- No discontinued products included