Matching engine modes determine the fundamental approach SmartSize AI Fit Recommender uses to generate size recommendations, providing control over algorithm behavior from strict exact matching to sophisticated flexible scoring.

Understanding Matching Engine Modes

The Three Matching Modes

SmartSize AI Fit Recommender offers three distinct matching engine modes, each with fundamentally different recommendation approaches:

1. EXACT Mode

Algorithm Behavior: Requires all enabled measurements to fall exactly within their respective size ranges.

How It Works:

  • Binary Matching - Either finds a perfect match or returns no recommendation
  • All-or-Nothing - Customer measurements must be within range for ALL enabled measurements
  • No Flexibility - Ignores strictness settings in favor of exact boundary enforcement
  • Single Result - Returns only the size with exact measurement matches

When to Use:

  • High-precision Products - Items where exact fit is critical
  • Technical Garments - Performance wear where measurements directly affect function
  • Safety Equipment - Items where incorrect sizing could be dangerous
  • Premium Products - High-end items where customers expect perfect fit

2. FLEXIBLE Mode (Default)

Algorithm Behavior: Uses sophisticated scoring algorithms to find the best overall fit, incorporating strictness settings as pre-filters.

How It Works:

  • Strictness Filtering - Applies measurement strictness as initial filters
  • Distance Calculation - Measures how close customer measurements are to size midpoints
  • Likelihood Scoring - Calculates probability-based fit scores for each eligible size
  • Combined Optimization - Balances distance and likelihood for optimal recommendations

When to Use:

  • General Clothing - Most apparel and fashion items
  • Comfort-Focused Items - Products where customer satisfaction is priority
  • Complex Sizing - Items with multiple measurements affecting fit
  • Mass Market Products - Items serving diverse customer preferences

3. CLOSEST_POSSIBLE Mode

Algorithm Behavior: Finds the closest available size even when no exact matches exist, within size table boundaries.

How It Works:

  • Boundary Checking - Ensures customer measurements fall within overall size table bounds
  • Proximity Optimization - Recommends the size with measurements closest to customer’s
  • No Strictness Application - Ignores strictness settings for maximum matching flexibility
  • Best Available Match - Always provides a recommendation if within table bounds

When to Use:

  • Limited Size Ranges - Products with restricted size availability
  • Broad-Fit Items - Products designed to accommodate range of body types
  • Cost-Sensitive Sizing - When inventory constraints limit size options
  • Maximum Coverage - Situations requiring recommendations for most customers

How Matching Modes Affect Recommendations

EXACT Mode Implementation

Strict Boundary Enforcement 

  Customer Measurements: Chest 36.5", Waist 30.5", Hip 38.5"

Size M Ranges:
- Chest: 36.0" - 38.0" ✓ (36.5" is within range)
- Waist: 30.0" - 32.0" ✓ (30.5" is within range)  
- Hip: 38.0" - 40.0" ✓ (38.5" is within range)

Result: Recommends Size M (exact match found)

No Match Scenario 

  Customer Measurements: Chest 35.5", Waist 29.5", Hip 37.5"

Size M Ranges:
- Chest: 36.0" - 38.0" ✗ (35.5" below minimum)
- Waist: 30.0" - 32.0" ✗ (29.5" below minimum)
- Hip: 38.0" - 40.0" ✗ (37.5" below minimum)

Result: No size recommendation (no exact match)

FLEXIBLE Mode Implementation

Two-Stage Process

  1. Strictness Filtering - Eliminate sizes that don’t meet strictness criteria
  2. Algorithm Scoring - Score remaining sizes using distance and likelihood algorithms

Example with Mixed Strictness 

  Quiz Configuration:
- Chest: Must Be Inside Range
- Waist: Flexible  
- Hip: In Range or Bigger

Customer: Chest 35.5", Waist 29.5", Hip 39.5"

Stage 1 (Strictness Filtering):
- Size S: Chest range 34-36" ✓, qualifies for scoring
- Size M: Chest range 36-38" ✗, eliminated  
- Size L: Chest range 38-40" ✗, eliminated

Stage 2 (Algorithm Scoring):
- Only Size S remains after filtering
- Scores based on distance from measurements to size midpoint
- Result: Recommends Size S

CLOSEST_POSSIBLE Mode Implementation

Boundary Validation First 

  Overall Size Table Bounds:
- Chest: 32.0" - 42.0"
- Waist: 26.0" - 36.0"
- Hip: 34.0" - 44.0"

Customer: Chest 35.0", Waist 28.0", Hip 37.0"
Result: Within bounds, proceed with proximity matching

Proximity Optimization 

  Distance Calculations:
- Size S midpoint distance: 2.1"
- Size M midpoint distance: 1.8" ← closest
- Size L midpoint distance: 3.2"

Result: Recommends Size M (closest overall fit)

Practical Applications by Industry

Fashion and Apparel (FLEXIBLE Mode)

Why FLEXIBLE:

  • Customer Variety - Accommodates different fit preferences
  • Fabric Considerations - Allows for stretch, drape, and style factors
  • Return Reduction - Balances accuracy with customer satisfaction
  • Style Adaptation - Handles loose-fit vs fitted styles appropriately

Configuration Strategy:

  Matching Engine: FLEXIBLE
Strictness: Mixed based on garment type
- Casual wear: All measurements Flexible
- Semi-formal: Chest Must Be Inside Range, others Flexible
- Formal wear: All measurements Must Be Inside Range

Technical and Performance Wear (EXACT Mode)

Why EXACT:

  • Functional Requirements - Proper fit affects performance
  • Safety Considerations - Incorrect sizing can compromise safety
  • Performance Optimization - Exact measurements ensure intended function
  • Customer Expectations - Technical customers expect precision

Configuration Strategy:

  Matching Engine: EXACT
Strictness: Not applicable (overridden by EXACT mode)
Focus: Precise size ranges based on performance requirements

Mass Market and Budget Items (CLOSEST_POSSIBLE Mode)

Why CLOSEST_POSSIBLE:

  • Maximum Coverage - Serves broadest customer base
  • Inventory Constraints - Limited size availability
  • Cost Optimization - Reduces need for extensive size ranges
  • Customer Accommodation - Provides options even with imperfect fit

Configuration Strategy:

  Matching Engine: CLOSEST_POSSIBLE
Strictness: Not applicable (ignored in this mode)
Focus: Broad size ranges with strategic overlap

Strategic Implementation

Choosing the Right Mode

Decision Matrix

Product Type Customer Expectations Size Range Recommended Mode
Designer Fashion High precision Complete EXACT
Casual Wear Comfort priority Standard FLEXIBLE
Athletic Wear Performance + comfort Extended FLEXIBLE
Budget Fashion Value + fit Limited CLOSEST_POSSIBLE
Technical Gear Exact function Precise EXACT
Children’s Wear Growth accommodation Broad FLEXIBLE

Mode Selection Criteria

Choose EXACT When:

  • Fit precision is critical for function or safety
  • Customers expect exact sizing
  • Premium pricing justifies perfect fit expectations
  • Product design requires specific measurements

Choose FLEXIBLE When:

  • Customer satisfaction is primary goal
  • Multiple factors affect perceived fit
  • Size ranges have strategic overlap
  • Comfort and style preferences vary

Choose CLOSEST_POSSIBLE When:

  • Maximum customer coverage is priority
  • Size inventory is constrained
  • Broad-fit products accommodate size variation
  • Cost considerations limit size range options

Mode Migration Strategies

From EXACT to FLEXIBLE

When to Consider:

  • High “no size found” rates
  • Customer feedback requests more options
  • Return rates low but recommendation rates low

Implementation:

  1. Add Strictness Settings - Use Custom garment category
  2. Start Conservative - Begin with “Must Be Inside Range” for all measurements
  3. Gradual Relaxation - Progressively relax non-critical measurements
  4. Monitor Impact - Track recommendation rates and customer satisfaction

From FLEXIBLE to EXACT

When to Consider:

  • High return rates due to fit issues
  • Customer feedback indicates sizing inconsistency
  • Premium product positioning requires precision

Implementation:

  1. Validate Size Ranges - Ensure accurate measurement data
  2. Gradual Transition - Test with subset of products first
  3. Customer Communication - Explain sizing precision improvements
  4. Monitor Closely - Watch for decreased recommendation rates

Advanced Configuration

Combining with Strictness Settings

FLEXIBLE Mode + Strictness Optimization

Strategy: Use strictness to pre-filter sizes, then apply flexible scoring

  Example Configuration:
Matching Engine: FLEXIBLE
Chest: Must Be Inside Range (critical for appearance)
Waist: Flexible (comfort accommodation)
Hip: In Range or Bigger (movement freedom)

Result: Precise chest fit with flexible waist and hip matching

Mode-Specific Behaviors

EXACT Mode:

  • Strictness settings ignored
  • All measurements must be within range
  • Binary match/no-match result

FLEXIBLE Mode:

  • Strictness settings act as pre-filters
  • Algorithm scores remaining sizes
  • Optimized recommendation from eligible sizes

CLOSEST_POSSIBLE Mode:

  • Strictness settings ignored
  • Boundary checking only requirement
  • Pure proximity-based matching

Performance Considerations

Recommendation Speed

EXACT Mode: Fastest (simple boundary checking) CLOSEST_POSSIBLE Mode: Fast (distance calculations only) FLEXIBLE Mode: Moderate (strictness filtering + algorithm scoring)

Recommendation Coverage

EXACT Mode: Lowest (strict requirements) FLEXIBLE Mode: Moderate to High (balanced approach) CLOSEST_POSSIBLE Mode: Highest (maximum matching flexibility)

Customer Satisfaction Balance

EXACT Mode: High precision, lower coverage FLEXIBLE Mode: Balanced precision and coverage CLOSEST_POSSIBLE Mode: Maximum coverage, variable precision

Troubleshooting Matching Modes

Common Issues and Solutions

Low Recommendation Rates

Problem: Too many customers receive “no size found” Possible Causes:

  • EXACT mode too restrictive for product type
  • Size ranges don’t cover customer base adequately
  • FLEXIBLE mode strictness settings too restrictive

Solutions:

  • Consider switching from EXACT to FLEXIBLE mode
  • Review and expand size range coverage
  • Relax strictness settings in FLEXIBLE mode
  • Switch to CLOSEST_POSSIBLE for maximum coverage

High Return Rates

Problem: Customers return items due to fit issues Possible Causes:

  • CLOSEST_POSSIBLE mode compromising fit accuracy
  • Size ranges not reflecting actual garment measurements
  • FLEXIBLE mode prioritizing coverage over precision

Solutions:

  • Switch from CLOSEST_POSSIBLE to FLEXIBLE mode
  • Validate size ranges against actual product measurements
  • Tighten strictness settings in FLEXIBLE mode
  • Consider EXACT mode for precision-critical products

Inconsistent Results

Problem: Similar customers receive different recommendations Possible Causes:

  • FLEXIBLE mode algorithm responding to small measurement differences
  • Size ranges with inappropriate overlap
  • Strictness settings creating unexpected filtering

Solutions:

  • Review size range overlap and progression
  • Adjust strictness settings for more consistent filtering
  • Consider EXACT mode for more predictable results
  • Validate customer measurement input accuracy

Matching engine modes provide fundamental control over how SmartSize AI Fit Recommender generates recommendations, enabling you to align algorithm behavior with your product characteristics, customer expectations, and business objectives.