Glood’s Frequently Bought Together: Smart Bundle Builder

Glood’s Frequently Bought Together (FBT) feature is a revenue-driving powerhouse that identifies products customers love to purchase together. Available in your Glood Dashboard, FBT combines AI-discovered patterns with manual bundle creation tools, giving you complete control over your upselling strategy.

Algorithm Overview

Core AI Components

1. Association Rule Mining

Our proprietary model analyzes millions of transactions to identify:
  • Support: How frequently items appear together
  • Confidence: Probability of buying B when A is purchased
  • Lift: Strength of the association beyond random chance
# Simplified algorithm representation
if confidence(A�B) > threshold and lift(A,B) > 1.5:
    recommend(B) when viewing(A)

2. OpenAI GPT-4 Integration

GPT-4 enhances FBT recommendations by:
  • Semantic Understanding: Recognizes that “phone case” naturally pairs with “screen protector”
  • Context Awareness: Understands seasonal combinations (BBQ grill + charcoal in summer)
  • Category Bridging: Identifies cross-category opportunities (camera + travel backpack)
  • Description Analysis: Reads product descriptions to find compatibility mentions

3. Google Gemini Visual Analysis

For visually-driven products, Gemini provides:
  • Style Matching: Ensures aesthetic compatibility in fashion bundles
  • Color Coordination: Suggests items that visually complement each other
  • Size Compatibility: Verifies physical compatibility for accessories
  • Brand Consistency: Maintains brand coherence in luxury segments

4. Glood Fine-tuned Models

Our specialized e-commerce models handle:
  • Price Optimization: Balances bundle value with customer price sensitivity
  • Inventory Management: Prioritizes in-stock combinations
  • Margin Optimization: Considers profitability in bundle composition
  • Conversion Prediction: Estimates likelihood of bundle purchase

Data Processing Pipeline

Real-time Signals

  1. Current Session Data
    • Products viewed in current session
    • Time spent on each product
    • Cart additions and removals
    • Search queries used
  2. Historical Patterns
    • Previous purchases by this customer
    • Segment-level purchase patterns
    • Seasonal buying trends
    • Category preferences
  3. Product Relationships
    • Manufacturer recommendations
    • Product compatibility matrices
    • Collection groupings
    • Substitute vs complement classification

Advanced Techniques

Temporal Pattern Recognition

Our AI identifies time-based patterns:
  • Morning routine products (coffee + filters)
  • Seasonal combinations (sunscreen + swimsuit)
  • Holiday gift sets
  • Back-to-school bundles

Multi-objective Optimization

The algorithm balances multiple goals:
  1. Relevance: How well items complement each other
  2. Value: Total bundle price optimization
  3. Availability: Stock levels and fulfillment capability
  4. Profitability: Margin considerations
  5. Customer Satisfaction: Historical review scores

Contextual Adaptation

Recommendations adjust based on:
  • Device Type: Mobile vs desktop bundle sizes
  • Customer Segment: Budget vs premium shoppers
  • Geographic Location: Regional preferences
  • Time of Day: Purchase urgency indicators

Performance Metrics

Success Indicators

  • Bundle Attach Rate: 35% of customers add recommended items
  • AOV Increase: Average 42% order value boost
  • Conversion Lift: 28% higher conversion with bundles
  • Customer Satisfaction: 4.7/5 relevance rating

Quality Assurance

  • Diversity Score: Ensures variety in recommendations
  • Coherence Check: Validates logical product combinations
  • Price Balance: Maintains reasonable bundle totals
  • Inventory Health: Avoids out-of-stock frustration

Implementation Examples

Fashion & Apparel

Viewing: Designer Dress
Recommends:
- Matching Clutch (visual similarity via Gemini)
- Coordinating Shoes (style matching)
- Jewelry Set (price-point alignment)

Electronics

Viewing: DSLR Camera
Recommends:
- Memory Card (necessity detection via GPT-4)
- Camera Bag (protection need)
- Extra Battery (usage pattern)

Home & Garden

Viewing: Outdoor Grill
Recommends:
- Grill Cover (maintenance item)
- BBQ Tool Set (complementary tools)
- Propane Tank (fuel requirement)

Continuous Learning

Our FBT algorithm improves through:
  1. A/B Testing: Continuous experimentation with bundle compositions
  2. Feedback Loops: Learning from actual purchase decisions
  3. Seasonal Adjustments: Adapting to changing seasonal patterns
  4. Category Evolution: Discovering new product relationships

Glood Dashboard Features

Manual Bundle Builder

Create custom bundles directly in Glood:
  1. Visual Bundle Creator
    • Drag and drop products to create bundles
    • Preview how bundles appear on your store
    • Set bundle-specific discounts
    • Schedule seasonal bundle campaigns
  2. CSV Bundle Upload
    primary_product_id,bundle_product_1,bundle_product_2,discount
    PROD123,PROD456,PROD789,10%
    PROD234,PROD567,PROD890,15%
    
  3. Smart Bundle Templates
    • Accessory bundles (phone + case + charger)
    • Complete the look (outfit combinations)
    • Starter kits (everything you need)
    • Gift sets (curated collections)

FBT Configuration in Glood

Section Settings:
  Algorithm: Frequently Bought Together
  Display Mode: Bundle with Discount
  Number of Products: 3
  Bundle Discount: 10%
  Show Individual Prices: Yes
  One-Click Add All: Enabled
  
Advanced Settings:
  Minimum Confidence: 0.3
  Inventory Check: Real-time
  Price Range: ±30% of main product
  Category Mixing: Allow cross-category
  Brand Preference: Mix brands

Bundle Performance Analytics

Track bundle success in Glood Analytics:
  • Bundle Take Rate: % of customers who buy the full bundle
  • Partial Bundle Rate: % who buy some recommended items
  • Revenue Lift: Additional revenue from bundles
  • Popular Combinations: Top-performing bundles
  • Inventory Impact: Stock movement from bundles

Glood’s Unique FBT Features

Dynamic Bundle Pricing

// Glood automatically calculates optimal bundle discounts
{
  "bundle_price": {
    "strategy": "percentage_off",
    "value": 10,
    "min_items": 2,
    "max_discount": 50,
    "show_savings": true
  }
}

Inventory-Aware Bundling

  • Auto-hide out-of-stock bundles
  • Suggest alternatives when items unavailable
  • Priority boosting for high-stock items
  • Pre-order bundle support

A/B Testing Framework

Test different bundle strategies:
  • AI vs Manual bundles
  • Different discount levels
  • Bundle size variations
  • Display formats

Glood FBT Success Stories

Fashion Retailer

  • Setup: Complete outfit bundles
  • Result: 43% increase in AOV
  • Key: Mixed AI + curated bundles

Electronics Store

  • Setup: Accessory bundles for main products
  • Result: 67% bundle attachment rate
  • Key: Smart pricing with 15% bundle discount

Beauty Brand

  • Setup: Skincare routine bundles
  • Result: 3.5x customer lifetime value
  • Key: Educational bundle descriptions

Implementation Best Practices

Quick Start with Glood

  1. Install Glood app from Shopify App Store
  2. Navigate to AI Recommendations → Create Section
  3. Select “Frequently Bought Together”
  4. Use auto-discover or create manual bundles
  5. Customize design to match your theme
  6. Deploy to product pages

Optimization Tips

  • Start with AI-discovered bundles
  • Add manual bundles for new products
  • Test different discount levels (5-20%)
  • Use Glood’s one-click “Add All” button
  • Monitor analytics weekly

Common Configurations

  • High-value products: 5-10% discount, 2-item bundles
  • Accessories: 15% discount, 3-4 item bundles
  • Consumables: 20% discount, bulk bundles
  • Gift sets: Fixed price, curated selection
For step-by-step setup instructions, see our Glood FBT Implementation Guide.