TABLE OF CONTENTS
- Create a Search Profile
- Searchable Fields
- Relevance Based Ranking
- Query Properties
- Test Your Search
- Default Search Profile (Database)
- Managing/Editing Search Profile
- Changes on the Admin screen (from version 9.7.1)
Navigate to: Marketing > Site Search.
Searches of the web store can be managed from this section. Search Profiles are specific to catalogs. They indicate the rules based on which the search feature would find out the results on the web store associated with the respective catalog.
Note - When a new Search profile is created and no catalog is selected then all the default settings are applied.
Create a Search Profile
Navigate to Marketing > Site Search
Search Settings page opens.
Click Search Profiles.
Click on Add New
Add Search Profile page opens as shown below
Select the Catalog for which the search profile needs to be created.
Administrators need to Provide a unique Search Profile Name.
Search Profile Default Setting
When an administrator creates a new Search Profile selecting only Catalog and creating Search Profile Name then the default settings are applied to the Search Profile.
The default setting indicates the predefined selections that are applicable when a new search profile is created. Following are the default settings
Catalog Name
Search Profile Name
Searchable Fields:
SKU (no boosting value) - selected by default
Product Name (no boosting value) - selected by default
Relevance Based Ranking - nothing selected
Query Properties
Query Type
Multi Match - Cross (to be selected by default)
Operator = AND (to be selected by default)
Autocorrect = No (to be selected by default)
Test Your Search Here
Your search results will appear here
Searchable Fields
The fields added here are used for searching the entered text on the webstore. By default, Product name and SKU are available on creating a new search profile.
Adding multiple Searchable fields
The administrator can add multiple searchable fields under this section. Only the numeric values can be added to boost the relevance of the search. The numeric values decide from which searchable field the search results would be fetched first.
Note - A Boost Score (a numeric value between 1-999) can be added as an input to the searchable field to promote the position of the products in search results where the searched keyword is found in the field whose boosting score is the greater.
Removing the Searchable field
An Administrator can remove the added searchable fields (if required) by clicking on the cross.
Relevance Based Ranking
An administrator can select the sequence for displaying the products on the webstore. An administrator can select the following options from the dropdown to select the sequence of the search results.
Price - This can be used when the Administrator wants to show the results on the basis of Prices associated with the products.
Rating - This can be used when the Administrator wants to show the results on the basis of Ratings shared for the products.
Total Reviews - This can be used when the Administrator wants to show the results on the basis of reviews shared for the products.
Note - The order of the sequence can be decided by selecting ASC/DESC.
Query Properties
Query Property defines how the searchable fields can be searched and shown on the webstore.
Multi-Match: When the Multi-Match query type is configured, the application searches keywords in more than one searchable field to display results on the Web Store.
It has two types:
Best: When Multi Match - Best is selected, all the searched text/keywords must be available combined in any one of the searchable fields.
Example:Searchable fields = Product Name, SKU and Search Keyword entered on Web Store = Dewalt Kit
The behavior of Multi-Match - Best = Search results will validate if “Dewalt Kit” both keywords are available in a single searchable field i.e (SKU/Product) if found together, will include all the associated products in the search results, depending on the operator used.
Cross: When Multi Match - Cross is selected text/keywords can be available across multiple searchable fields to bring the search results.
Example:Searchable fields = Product Name, SKU and Search Keyword entered on Web Store = Dewalt Kit
The behavior of Multi-Match - Cross = Search results will validate if “Dewalt Kit”/ “Dewalt” / “Kit” either/all of the keywords are available in SKU/Product Name and if found, will include all the associated products in the search results.
Operators
In case of two or more keywords, the Operator helps to decide whether the searchable field should have all the entered keywords or can have either of the keywords while displaying the search results on the webstore.
There are two types of operator
The operator OR: Give results where either of the words from the searched keywords exists in the searchable field.
Operator AND: Give results where all the words in the searched keywords exist in the searchable field.
AutoCorrect
When an entered keyword has a spelling mistake, Autocorrect helps in identifying the correct spelling of the keyword and displays the associated product
The maximum number of edits allowed is 2; by default, the value is AUTO:3,6.
The administrator can set the maximum number of edits in the entered keyword.
If keywords length is 0 to 2: no edits are allowed and the word must match exactly
If keywords length is 3 to 5: One edit is allowed
If keywords length is > 5: Two edits are allowed
Note -
If the administrator selects Autocorrect as Yes all the above rules are applicable.
Autocorrect doesn’t work in combination with Query Type => Multi Match - Cross
Test Your Search
The administrator can try the search results under the “Try it yourself” section by specifying a search keyword and clicking on the search icon.
The search results can be validated under this section before applying the changes to the webstore.
Clicking on save, saves all the rules/validations added by an administrator and the changes are visible on the webstore.
System Defined Search Profile (Database)
A Search Profile is available in the Database which is termed as Default Search Profile. A default search profile is used in the following two cases.
If a created search profile is not associated with the store base catalog but is associated with the user/account catalog then default search profile (database) is used.
When no search profiles are created for a specific store-based Catalog
The system-defined setting indicates the configurations that are applicable when no search profile is manually created from the admin application for Stores, Catalogs or when profile and account-based catalogs are used on the webstore: Following are the default
Searchable Fields:
SKU (no boosting value)
Product Name (no boosting value)
Query Properties
Query Type
Multi-Match - Cross (to be selected by default)
Operator = AND (to be selected by default)
Autocorrect = No (to be selected by default)
Managing/Editing Search Profile
The administrator can edit the search profile from this section. Following are the additional details that are available on the admin site.
Create Facets
Under the Edit Search Profile, click the Facets tab.
Click Associate Facets.
Select the Facets.
Click on the required number of attribute names to associate facets.
Once associated, all the associated facets are visible as shown below.
If the administrator decides to delete a particular associate facet then by clicking on the delete button provided the same facet would be deleted.
Associate Stores
Under the Edit Search Profile, click the Stores tab.
Click Associate Stores.
Select the Store(s) on which the search profile should be applicable. Stores having the selected Catalog are displayed.
Click Associate Stores.
Note: Only after the Store association is made can a Search Profile be set as Default.
Changes on the Admin screen (from version 9.7.1)
On the admin screen following tabs will not be available in Search Profile
Marketing > Site Search > Search Profile > Query Properties
Query Type
Match Phrase Prefix
Match
Match Phrase
Minimum Should Match
Enable Accurate Scoring
Elasticsearch Basics
Elasticsearch uses various technical terms/queries for processing the data/keywords entered in order to make productssearchable and share relevant results. A few of the technical terms are used in Znode which are explained below:
Analyzer: Use to share instructions that are given to Elasticsearch on the nitty-gritty of how the data should be indexed/stored.
Tokenizer: Use to decide how Elasticsearch will take a set of words and divide it into separated terms called “tokens”. Token generation depends on the type of token filter used
Token Filter: A token filter is an operation done on tokens that modifies them in some way or another
Following token filters would be used in Elasticsearch for Znode (base)
Lowercase: Use to changes token text to lowercase
Ex: If a text entered = “The Quick Brown Fox” it will be converted to “the quick brown fox”
Synonyms: Use to create a list of words that can be used as an alternative and is saved in the synonym list, when a user enters a search keyword it is then compared with the synonym list to find the match before displaying the search results
Ex: If a text entered = “Brown Fox” and synonym list has Fox= “coyote, dingo” then the search results would display products that have fox, coyote, dingo
Stopwords: Use to ignore all stop words if found in the search keyword. Stop words are usually words like “to”, “I”, “has”, “the”, “be”, “or”, etc. They are filler words that help sentences flow better, but provide a very little context on their own.
Ex: If a text entered = “The Quick Brown Fox” it will be considered as “quick brown fox”
Shingle: Use when tokens are to be generated using the concatenation of the adjacent tokens. Shingles are generally used to help speed up phrase queries
Ex: If a text entered = “The Quick Brown Fox” then tokens generated would be “the,” “the quick”, “quick” “quick brown“, “brown fox” and “fox”
Ngram: Used to break a text into words, this is helpful to create tokens(set of words) that can be used to search the desired output in search results
Ex: If a text entered =“Quick fox” then tokens generated would be [ Q, Qu, u, ui, i, ic, c, ck, k, f, fo, o, ox, x ]
Character Filters: Use to preprocess the stream of characters before it is passed to the tokenizer. A character filter receives the original text as a stream of characters and can transform the stream by adding, removing, or changing characters.
Ex: If the Character mapping is done as "- => &" and a user enters a text 3-3 it will be converted to 3&3 while displaying the results
Stemmer: Use to reduce a word to its root form to ensure variants of a word match during a search.
Ex: walking and walked can be stemmed to the same root word: walk
Fuzziness: Use to identify two elements of text, strings, or entries that are approximately similar but are not exactly the same
Ex: If a text entered = Cat result will also search for Mat, Bat, Rat, Sat etc
Ex: If a text entered = Black result will also search for Lack, Slack etc