Match Criteria

This section allows you to customize how the algorithm matches user input against potential search results. It governs how user queries align with search results, optimizing the relevance and precision of those results based on specific algorithms. Adjusting these criteria helps tailor the search experience to user expectations, enhancing engagement and conversion rates.

Text Match

This is the cornerstone of search functionality, ensuring that the user's text input matches content within your product database. It's the most straightforward form of search, looking for direct text correspondences between the query and database entries.

Example: A customer searches for "blue denim jacket." Text Match will scan the product database for these exact words. Products labeled with "blue denim jacket" in the title or description are prioritized in search results.

Phrase Match

Phrase Match takes Text Match a step further by looking for results containing the entire search query in the specified order. It offers more flexibility by tolerating variations in the query's wording or additional content between terms.

Example: For a query like "leather wallet men," Phrase Match can return "genuine leather wallet for men" or "men's bifold leather wallet," ensuring the core search intent is captured despite variations in phrase construction.

Relaxed Match

Relaxed Match introduces a higher level of flexibility in recognizing matches, allowing for discrepancies in spacing, special characters, or specific alphanumeric sequences.

Example: For instance, if a user enters “CE 3,” a relaxed match analyzer might correctly identify a match with the catalog term “CE3,” acknowledging the alphanumeric equivalence and accommodating variations in user input.

Synonym Match

Synonym Match expands the search's scope by including synonyms defined in the Search Console, broadening the range of relevant results based on a comprehensive understanding of language and context.

Example: If a user searches for "sofa," Synonym Match ensures that results also include "couch" or "settee," recognizing these terms as interchangeable in many contexts.

Stemmed Keyword Match

Stemmed Keyword Match addresses the linguistic variations of words by focusing on their root form. This setting accounts for plural forms, tenses, and other derivations of the root word, expanding the search's flexibility.

Example: Searching for "decorate" will also surface results for "decorating," "decorated," "decoration," and "decorative," encompassing a wide array of relevant content.

Intent Match Criteria

This criterion sets a percentage threshold for how much of the user's query must match with a document or listing for it to be considered a relevant result.

Example: For instance, if set at 50%, the search results must match at least half of the user's query terms to appear in the search results. This helps in filtering out less relevant results, ensuring a more focused search outcome.

Searchable Attributes Priority

This setting allows you to prioritize certain product attributes over others, guiding the search engine on which aspects of a product are most important.

Primary Attributes

Identified as the most crucial factors in matching queries with results, often including product name, category, or brand.

Secondary Attributes

Serve as additional factors that can refine search results, like color, size, or material.

Example: In a fashion store, secondary attributes like "Color" or "Size" help narrow down search results after the primary query, such as "dresses," leading to more specific suggestions like "red evening dress."

Fuzzy Search

Implement a search capability that compensates for user input errors, ensuring a robust search experience even with misspellings or typos. The fuzzy logic operates on an edit distance principle, allowing for character modifications, deletions, insertions, or transpositions within a defined threshold to match the intended search term closely. This tolerance allows the search to still return useful suggestions even when the query isn't perfectly accurate.

Edit Distance

This is a measure of how many character changes (additions, deletions, substitutions, or transpositions) are allowed for a query term to match with a potential result in the database.

Example: With an Edit Distance of 2, a search for "camra" could return results for "camera," as the query can be corrected with two changes (substituting "m" with "e" and inserting "e").

Prefix Length

This specifies the minimum length of the query term's beginning that must be correct before the fuzzy search logic is applied.

Example: If the Prefix Length is set to 3, and a user searches for "refrigirator," the fuzzy search will only consider alterations to the query after the first three correct characters ("ref"), ensuring that results are likely to be relevant variations of "refrigerator" and not unrelated terms that might also fit the fuzzy criteria.

Semantic Search

Semantic Search adds a layer of understanding to the search process, interpreting the meaning and context behind queries. It can interpret queries like "affordable smartphones with good cameras," understanding the concepts of affordability and quality.

Vector Field/s

This is the input field which contains the raw data that needs to be vectorized. This will be used to generate vectors.

Vector Index Field

This is the storage location for the vectors once they have been generated. It serves as a reference point for your semantic search algorithms, allowing them to efficiently query and retrieve vector data during autosuggest operations.

Nearest Neighbor(s)

This refers to the number of used in semantic search where the system finds the 'closest' or most similar results to a query in the vector space.

Usage: If a user searches for "affordable laptops for gaming," the search engine uses the nearest neighbor method to find products that closely match this description in terms of price and functionality.

Minimum Score

The minimum score is a threshold setting that determines how closely a search result needs to match the query in terms of semantic similarity. Only results that meet or exceed this score are displayed to the user.

Usage: You might set a minimum score to ensure that only the most relevant results are shown. For example, setting a higher score for technical product searches ensures that only the most accurate matches are displayed.

Tip: Regularly review and adjust these settings based on user feedback and search analytics to continuously improve the search experience.

Warning: Setting thresholds too high might filter out potentially relevant results, while too low may include less relevant ones, impacting the overall user experience.