Understanding and influencing result ranking
Introduction
When a user performs a search, the order of the results is determined by a sophisticated ranking system. It's not a simple text match, Luigi's Box uses an ensemble of AI models and a multitude of signals to present the most relevant, useful, and profitable products first.
Understanding how this ranking works is key to optimizing your search. This guide breaks down the standard ranking model and explains the various ways you can influence it, from providing richer product data to making manual adjustments in the Luigi's Box App.
What you'll learn
- The core signals that power Luigi's Box ranking.
- How to influence ranking using product data attributes.
- How to manually override ranking for specific products or queries.
Who this guide is for
- Developers and e-commerce managers looking to optimize search performance.
- Teams promoting new, seasonal, or high-margin products.
- Anyone seeking to improve product visibility in Luigi's Box-powered search.
Prerequisites
- Familiarity with how product data is indexed in Luigi's Box.
- Access to your Luigi's Box App for testing manual adjustments.
1. How Luigi's Box ranks results
The default ranking behavior combines multiple signals. These are evaluated and balanced automatically to ensure the best customer experience.
Full-text match quality
Match quality is the foundation of ranking. Rather than assigning a raw "score," Luigi's Box groups products into match-quality tiers:
- Exact matches
- Matches with typo tolerance
- Partial matches (where some words in the query are missing)
Traditional full-text search engines often rely on numeric scoring systems. Luigi’s Box improves on this by using quality tiers instead of scores, ensuring more predictable and meaningful result ordering.
Availability
Products that are in stock are favored. Availability is ranked in tiers, for example:
- "In stock, available now" will rank higher than
- "In stock, available in 48 hours"
Learn more about the availability ranking field.
Analytics feedback loop
Products that users frequently view, add to cart, or purchase are promoted over time. This allows bestsellers and highly engaging products to rise naturally in the rankings.
Learn more about the analytics feedback loop.
Personalization
Search results can be personalized based on a user's previous interactions or profile data. That means the same query may return slightly different product orders depending on the user.
2. Influence ranking with product data attributes
You can guide the ranking engine by enriching your product data with specific attributes.
There's no need to explicitly "enable" these features. Once you include fields like margin, introduced_at, or boost in your product feed, Luigi’s Box automatically begins using them as ranking signals.
Ranking by freshness
New products often lack sales data, making them harder to rank naturally. To solve this, Luigi's Box includes a freshness boost.
How it works:
Add an introduced_at field to your product feed with the product's launch date (YYYY-MM-DD or full ISO timestamp).
Effect:
Newer products receive a time-limited boost, which gradually decays over 60 days. The system uses a log-decay curve, meaning the visibility advantage fades gradually rather than abruptly.
This helps compensate for the lack of data in the analytics feedback loop, giving new products a chance to surface while building their engagement history.
More on the introduced_at field.
Ranking by margin
You can favor products that are more profitable for your business.
How it works:
Include a numeric margin attribute in your product data.
Effect:
Products with higher margins are given a stronger ranking signal. Think of this as casting a “vote” toward profitability. It does not override other factors, relevance still comes first.
Margin does not hard-sort results. Even a high-margin item won't outrank something that is a better match for the query.
Boosting with an attribute
For the strongest product-level control, you can use the boost field.
How it works:
Set the boost value between 0 (no boost) and 3 (maximum boost) in your product feed.
Effect:
Boosted products can outrank nearly all others, except in cases where match quality dominates. Use this to promote:
- Campaign items
- High-priority seasonal products
- Clearance stock or urgent visibility needs
3. Override ranking in the Luigi's Box App
You can make manual, real-time adjustments without editing your product data. This is useful for campaign launches, emergency fixes, or fine-tuning performance.
These manual methods are sometimes referred to as manual interference in Luigi's Box documentation.
Global product boost
Navigate to:
Catalog management > Boosting
There, you can manually apply a boost to any product. This works the same way as the boost attribute in your data, without needing to touch the feed.
Per-query ranking rules
Navigate to:
Search > Search results customizations
Here, you can pin or demote products for specific queries.
For example:
- Pin a specific "running shoe" to appear first for the query
running shoes - Demote outdated or lower-converting products for branded searches
This is especially useful for:
- Campaign timing
- Partner brand priorities
- Query-level troubleshooting
4. Ranking signal hierarchy
Here's a simplified view of how ranking signals are prioritized in the Luigi's Box engine:
- Match Quality
- Boost (manual or data-level)
- Availability
- Analytics Feedback (user behavior) or Margin or Freshness or Personalization
You may see slight reordering depending on how many signals are present and how they interact, but the overall structure remains consistent.
Next steps
Now that you understand how Luigi's Box ranks and how to influence it, consider exploring:
You're now ready to control your ranking system with both precision and insight.