When keywords are entered in search, the following steps are performed:

1. Retrieve User Profile

To utilize features such as Visitor Targets, a User Profile is retrieved. Geolocation data based on visitor’s IP and other custom profile information are saved.

2. Load and Cache Global and Conditional Boost Rules

All active Workbench rules related to adjusting relevance ranking are retrieved.

3. Load and Cache Personalization Insights

Insights from the Personalization Module (optional feature) are retrieved to adjust relevance ranking. These insights enable 1-to-1 Personalization in search results. In other words, based on a user's activity, specific results can be automatically boosted for that user. Another user that follows the same actions may not see the same results returned to the first user.

4. Load and Cache Visibility Rules

Visibility Rules in the Workbench filter which records would be excluded from the results.

5. Perform Search

a. Load and Cache Stop Words

Words that should be ignored are retrieved.

b. Load and Cache Synonyms

The Synonym list from Workbench is loaded.

c. Load and Cache Spelling Override entries

The Spelling Override list from Workbench is loaded.

d. Load and Cache Learning Intelligence

Intelligence from the Learning Search feature is loaded.

e. Execute Search

i. Tokenization

The Engine analyzes the query string and returns results that match the term(s).

ii. Apply Scoring

A score is assigned to each result, based on the searchable fields (i.e. marked as queried within the dashboard):

Based on the relevancy within these searchable fields, Hawk assigns a Score value to each result:

These Score values are based on the following:

Term Frequency – The higher number of times a term appears within a product’s searchable fields, the higher the score is
Rare Appearance Rate – Rarer terms that match within a product’s searchable fields give a higher influence to the score
Number of Query Terms – The higher number of query terms that matched within a product’s searchable fields will provide a higher score than other products that matched less query terms

(For detailed technical documentation regarding scoring, please see: https://lucene.apache.org/core/3_5_0/scoring.html)

iii. Apply Boost Rules, Personalization Insights, and Visibility Rules

Relevant rules and insights are applied.

Search Field Boost - Each search field can be applied a boost as well. The best practice would be to provide boost to fields that have stronger content such as a product name. This depends on the quality of your data.
Intelligent Search – Popularity can also be applied to influence relevance ranking. Popularity is determined by search behavior and how past customers are interacting with your products.
Boost and Bury – Rules can also be applied for specific triggers. Rules are flexible and can include different criteria to boost or bury. For example, if a customer searches for a specific term, a merchandiser can boost a specific brand of products to the top.
Pinning – Within the preview window, merchandisers can specifically choose which products should be boosted to the top by applying a pin. This can help when you want to merchandise specific products.

iv. Facet Filtering

Selected facets are applied to the results. In the example below, we've selected the Camp & Hike facet under Department, so the results included in this facet are returned: