Defining Spec Requirements
Teaching the AI exactly what 'optimization' means for your specific niche.
# Advanced Attribute Strategy: How to Define and Extract Niche Product Specifications
Last Updated: January 26, 2026
Level: Advanced
Reading Time: 9 Minutes
In the world of Answer Engine Optimization (AEO), "Generic" is the enemy. An AI agent looking for a "Monitor" will find a billion results. An AI agent looking for a *"27-inch IPS panel with 99% sRGB coverage and USB-C Power Delivery"* will find three.
To win, you must teach Folki exactly what specific data points define your products. This guide covers how to set up Spec Templates, utilize Global Rules, and strategy for Negative Constraints.
1. The Hierarchy of Attributes
Folki thinks about product data in three layers. Understanding this hierarchy is the key to efficient token usage and high-quality data.
Layer 1: Global Attributes (The Basics)
These are attributes that apply to *everything*, whether it's a brick or a drone.
- Examples: `Country of Origin`, `Material`, `Weight`, `Dimensions`, `Warranty Duration`.
- Strategy: Configure these once in Settings > Attribute Rules. Folki will attempt to fill these for every single SKU in your account.
Layer 2: Category Attributes (The Drivers)
This is where AEO is won or lost. These are specific to a product class.
- For "Footwear": `Size System (US/EU)`, `Upper Material`, `Sole Material`, `Arch Support`.
- For "Monitors": `Refresh Rate`, `Panel Type`, `Response Time (GtG)`, `Vesa Mount Standard`.
- Strategy: You must map your store's "Collection" or "Category" to a Folki Spec Template.
Layer 3: Product-Specific Attributes (The Long Tail)
Sometimes, a product has a unique feature that doesn't fit a template (e.g., a specific "AI Noise Cancellation Processor" in one headphone model).
- Strategy: Folki's "Open Discovery" mode will find these automatically, even if you haven't defined them. However, defining them explicitly in a Template ensures they are never missed.
2. Creating High-Performance Spec Templates
Let's walk through creating a "Killer Template" for the Skincare category.
1. Navigate to Settings > Attribute Definitions.
2. Click New Template and name it "Skincare - Serums".
3. Add Target Attributes:
- `Active Ingredients`: (Type: *List of Strings*) - Crucial for "chem-conscious" shoppers.
- `Skin Type`: (Type: *Enum: Oily, Dry, Combo, Sensitive*) - Limits hallucinations to valid options.
- `Volume`: (Type: *Measurement*) - Ensures unit normalization (ml vs fl oz).
- `SPF Rating`: (Type: *Number*)
- `Cruelty Free`: (Type: *Boolean*)
Why Type Definitions Matter:
By telling Folki that `Cruelty Free` is a Boolean (True/False), you prevent the AI from extracting vague marketing text like "We love animals." It forces the agent to look for a definitive certification or statement, returning a clean `true` or `false`.
3. Negative Constraints (What NOT to find)
Sometimes, knowing what to *ignore* is as important as knowing what to find.
The "Compatibility" Problem
If you sell "iPhone Cases", basic AI scrapers often see the word "iPhone 13" on the page and think the product *is* an iPhone 13.
- Solution: In your template, you can set Extraction Hints:
- *"Target is an Accessory. Do not extract specs of the host device."*
This simple prompt instruction prevents the agent from populating your phone case metadata with the screen resolution of the phone itself.
4. Case Study: The "Gaming Laptop" Impact
A client selling refurbished laptops implemented a detailed Spec Template.
Before:
- Attributes: `Processor`, `RAM`, `Storage`.
- *Result:* Invisible to detailed queries.
After (Configured Folki Template):
- Attributes Added: `GPU TGP (Wattage)`, `Screen Brightness (nits)`, `sRGB Coverage %`, `Keyboard Switch Type`, `Thermal Cooling Info`.
The Outcome:
- Queries for "Bright laptop for outdoor use" began showing their product because `400 nits` was explicitly structured.
- Queries for "Full power RTX 3060" matched because `GPU TGP: 130W` was defined.
5. Dynamic Attribute Injection
What if you don't know what you don't know?
Select the "Allow Dynamic Discovery" checkbox in your template settings.
- How it works: If the agent finds a highly prominent technical spec that is *not* in your template (e.g., a new proprietary technology like "Apple Neural Engine"), it will create a new attribute field on the fly.
- Pros: You capture cutting-edge specs without updating templates constantly.
- Cons: can lead to "Schema Drift" (having 50 different names for similar things).
- Recommendation: Use Dynamic Discovery for "Draft" mode, then review the new attributes quarterly to standardize them into your official Templates.
FAQ: Spec Management
Can I copy/paste templates?
Yes. You can "Duplicate" a template. This is useful for hierarchies (e.g., create a base "Apparel" template, then duplicate it to make "Apparel > Outerwear" with extra fields like `Water Resistance`).
What happens if the agent can't find a required spec?
It depends on your strictness settings.
- Strict Mode: The product remains in "Review Required" status.
- Lenient Mode: The product is approved with the field left `null`.
Does this update my existing Shopify Metafields?
Yes. In the Integration Settings, you map each Folki Attribute to a specific Shopify Metafield Namespace/Key. When you click "Sync," we push the data exactly where you want it.
Conclusion
Attribute strategy is the "SEO Keyword Research" of the future. Instead of researching which keywords to stuff in a title, you are researching which facts machines need to make a decision.
Start small. Build one perfect template for your best-selling category today.