Hotel Ranking & Sorting Logic

The Nuitee Connect WhiteLabel interface sorts hotel search results using a weighted scoring system designed to balance price competitiveness, quality, location relevance, and popularity. This document explains how hotels are ranked, how individual factors are weighted, and how Best Seller hotels receive special treatment in the ranking process.

Overview

Each hotel returned in a search result is assigned a score based on multiple factors:

  • Price
  • Star rating
  • Guest rating
  • Number of reviews
  • Distance from the search location
  • Supplier or internal ranking signals
  • Best Seller status (if applicable)

Each factor is:

  1. Normalized to a comparable scale
  2. Weighted based on its importance
  3. Combined into a single score

Hotels with higher scores appear higher in the results.


Scoring Model

Base Scoring Formula (Conceptual)

hotel_score =
  (distance_score × distance_weight) +
  (star_score × star_weight) +
  (price_score × price_weight) +
  (guest_rating_score × guest_rating_weight) +
  (ranking_score × ranking_weight)

Weights may adjust slightly depending on data availability (e.g. missing prices or ratings), but the relative importance remains consistent.


Ranking Factors Explained

Distance

  • Measures proximity to the searched location
  • Closer hotels score higher
  • Strongly weighted to favor geographic relevance

Price

  • Lower prices score higher
  • Normalized across the result set
  • Excluded entirely for Best Seller hotels

Star Rating

  • Higher star ratings score higher
  • Used as a proxy for hotel category and amenities

Guest Rating & Reviews

  • Combines average guest rating and review count
  • Review volume influences confidence in the rating
  • Hotels with many reviews receive stronger signals

Ranking

  • Represents supplier or internal ranking signals
  • Helps stabilize ordering when other signals are similar

Weights: Regular Hotels

For non–Best Seller hotels, the typical weighting distribution is:

FactorApproximate Weight
Distance~50%
Star Rating~30%
Price~25%
Guest Rating & Reviewsup to ~30% (adjusted by review volume)
Ranking~15%

Note: Percentages are approximate and may be dynamically adjusted based on available data (e.g. missing prices or ratings).


Best Seller Hotels

Hotels marked as Best Sellers follow a different weighting model to reflect proven popularity and strong conversion performance.

Key Differences

  • Distance is prioritized
  • Price is excluded entirely
  • Guest ratings and review counts are boosted
  • Best Seller badge adds a strong ranking bonus

Weights: Best Seller Hotels

FactorBehavior
Distance~60%
Star Rating~10%
Price0% (ignored)
Guest Rating & ReviewsBoosted (≈1.5× impact)
Best Seller BadgeAdds a significant bonus

This ensures Best Sellers:

  • Remain highly visible
  • Are not penalized by short-term price fluctuations
  • Benefit from demonstrated guest satisfaction and demand

Best Seller Bonus

In addition to adjusted weights, Best Seller hotels receive an explicit scoring bonus when ranking results.

This bonus:

  • Helps break ties between similarly scored hotels
  • Ensures Best Sellers remain competitive even in dense result sets
  • Reflects historical performance and popularity

Normalization & Fairness

All numeric factors (price, distance, ratings) are normalized across the result set to ensure:

  • No single raw value dominates the ranking
  • Fair comparison between hotels
  • Stable ordering across different destinations and price ranges

Replicating the Sorting Logic

Customers who wish to implement similar ranking logic on their side should:

  1. Normalize each factor across the result set

  2. Apply different weight profiles for:

    • Regular hotels
    • Best Seller hotels
  3. Exclude price when ranking Best Sellers

  4. Boost guest rating impact for Best Sellers

  5. Apply an explicit Best Seller bonus

This approach will closely match the WhiteLabel default sorting behavior.


Important Notes

  • Ranking logic may evolve as data quality and signals improve
  • Exact numeric coefficients are not exposed, but relative weighting and behavior are stable
  • Custom sorting strategies can be applied by customers using Nuitee Connect data

Did this page help you?