Console user manual

Start using LupaSearch console, analyze how your store visitors interact with the e-shop search, and deliver a convertible user experience.

This user manual provides instructions on how to use a LupaSearch Dashboard.

Follow the information on how it works and use expert insights to improve the search experience of your e-commerce store.

Key definitions

LupaSearch - a highly configurable Software as a Service (SaaS) search engine that delivers real-time search results based on users’ intentions and actions in web applications, including e-commerce stores.

Index - data that search engine refers to when searching for the most relevant results to match the specific query. You can also think of it as a location where all your Documents are stored that serves as a base for the search.

Mapping - a schema that defines the attributes of each field in the Index. That is, how the Documents (and the fields they contain) are stored and indexed. These attributes can be used to configure Search Queries.

Document - any searchable unit that can be queried. Most often, Documents are e-commerce products. You can also use LupaSearch to query other types of units depending on your business needs: product categories, blog posts, tags, documentation, etc.

Search Query - a configuration that describes how data should be queried from your index. For example, you can set query fields, aggregations (facets), initial filtering, sorting rules and many other settings.

Public Query - an execution of your Search Query. Public query expands search query configuration with data that is required to perform a single query, like search text, page number, additional filters and sorting rules.

Synonym - additional search phrases that e-shop administrators can assign to the Documents to improve the search relevance. The terms may contain professional, formal, and culture-specific jargon.

Test your search

Use this section to test your Search and see how it works in real-time before you launch it onto your e-shop.

Pick a Document Query Key and Suggestion Query Key to try out a demo version of your website search. When the product feed loads, start typing in a Search Query into a search field. This section provides the same experience as your website visitor would have.

At the top right, press Configure to set up settings for a search test. Save any applied changes and try out the search.

Indices

Search indices are a frequently updated database that lists all your products and their details in a streamlined format. It allows users to quickly and accurately search your site without the need for the search engine to scan numerous pages each time. Instead, it searches through the index, similar to how a book's index helps you find key terms and relevant pages quickly.

A search index is like a high-speed directory, instantly connecting customers with products. It's built from your entire inventory and constantly updated, ensuring lightning-fast search results every time.

Analytics

LupaSearch includes an Analytics dashboard, which allows you to view search analytics to better understand how your users are using the search.

1. Locate and click on the "Analytics" section in the dashboard.
2. Select “Search index” at the top right of the page
3. Filter analytics results by selecting one relevant key (for instance, Main product search query or Suggestions search query) or all of them in the Search Query dropdown menu.
4. Click on the Date Range and pick the days. You will see the Analytics results within the selected time.

Analytics features include:

• Top 100 most popular queries - gain insights into your audience’s needs by analyzing the top 100 popular queries. Spot trends and adjust your products and strategies to meet customer demand. An extended list of the most popular queries is provided in the monthly reports, which are available in CSV format.
• Top 100 most popular suggestions (if LupaSearch events API is implemented) - analyze the top 100 search suggestions to see how users are using the search. Use these insights to refine your inventory and better align with market demands. An extended list of the most popular suggestions is provided in the monthly reports, which are available in CSV format.
• Top 100 most frequent queries that return no results - configure the search engine to display the top 100 most common queries with no results. This helps you identify unmet customer needs and use those insights to align your offerings with customer demand better, enhancing the shopping experience. An extended list of the most frequent queries with no results is provided in the monthly reports, in CSV format.
• Top item clicks - LupaSearch offers detailed insights into the top item clicks within your e-commerce platform. This feature helps you understand which products are attracting the most attention from your customers, allowing you to optimize your inventory, marketing strategies, and user experience. With real-time data on the most clicked items, you can make informed decisions to drive sales and enhance customer satisfaction.
• Top add to cart - this feature provides valuable insights into which products are most frequently added to carts on your e-commerce platform. This data helps you identify popular items, optimize inventory management, and tailor marketing efforts to boost conversions. By understanding customer preferences and behaviors, you can enhance the shopping experience and drive sales growth.
• Filtering by date - leverage our date-filtering analytics to deliver current, relevant results and make it easier for customers to find what they want.
• Filtering by search index or specific search query - use this feature to gain insights into your e-commerce search performance and how visitors engage with your site.
• Number of searches of a particular keyword - examine detailed stats for each keyword, including total searches, no-results counts, click events, add-to-cart actions, and click-through rates. Use this data to refine your product offerings and improve the shopping experience.
• A/B search campaign testing - move beyond guesswork with AI-driven A/B search campaign testing. Run different configurations at the same time, analyze performance through detailed graphs and stats, and adjust based on the results.
• AI-generated executive summary and monthly data reports - create AI-powered executive summaries from your monthly search performance data, graphs, and trends. Improve data-driven decision-making and optimize e-commerce search results with our detailed monthly reports.
• User journey insights - uncover the full user journey and gain insights beyond traditional search data. Identify confusing terms and frustrating navigation patterns early to prevent sales impact.
• Conversion rate - gain deeper insights into user behavior and see how intelligent search technology boosts your bottom line. Deliver highly relevant search results to convert clicks into sales.
• Indexing analytics - monitor the total processed products through daily imports, updates, and deletions. Use this data to quickly spot issues or unexpected changes in indexing volume.
• Click events and add to cart events - event reporting tracks user interactions on the search results page that aren't automatically registered by the LupaSearch server. Monitoring click and add-to-cart events helps assess search result relevance and understand user behavior.
• Estimated add-to-cart value - understand the approximate revenue generated from products added to carts after users click on search results. This metric reveals the revenue contribution of your search results page to your online store.
• Comparison to previous periods/trends - develop new strategies by comparing current search data with past trends to make informed decisions for your e-commerce business.

Users

Permission controls
Improve security with permission controls, ideal for medium and large businesses. This feature lets administrators grant specific permissions, such as Read Only or Modify, ensuring users access only what they need. It applies to critical areas like Analytics, API keys, and more, minimizing unauthorized access and reducing human error.

For example, a project manager can limit an analytic specialist’s access to only the Analytics section, protecting essential components from accidental deletion.

API security
LupaSearch prioritizes robust user management within its API security framework. Below you can find a breakdown of key functionalities:

Adding new users

1. Authentication: You'll need to be authenticated with valid API credentials (JWT or API key) before performing any user management actions.
2. User Creation Options: LupaSearch offers two methods for adding users:

• Invite User: This sends an invitation email to a designated email address. The recipient must accept the invitation and set their password to become a registered user. This method is ideal for granting access to specific individuals and maintaining control over user creation.
• Create User Directly: This allows you to create a new user account programmatically through the API. You'll define their email address, password (hashed for security), and role within your organization. This method is useful for automation and bulk user creation.

Deleting users

1. Authentication: similar to adding users, you'll need valid API credentials for user deletion.
2. User Identification: you'll specify the user you wish to delete by providing their unique user ID obtained through previous interactions with the API.
3. Permission Check: you can only delete users within your own organization. LupaSearch ensures unauthorized access and prevents accidental deletion of users belonging to other organizations.

Role-based access control

LupaSearch utilizes role-based access control to define user permissions within your organization. This means different user roles have varying access levels to API functionalities. Here's an example:

• Admin: has full access to all user management features, including adding, editing, and deleting users.
• Editor: may have permission to add and edit users, but not delete them.
• Viewer: may have limited access to user information for informational purposes but cannot perform any management actions.

By leveraging invitation workflows, user creation options, and role-based access control, LupaSearch empowers you to manage user access securely and efficiently.

API keys

API Keys are unique identifiers that allow secure access to our platform's services. They enable clients to programmatically integrate and interact with our search capabilities, ensuring only authorized applications can use our API. Using API keys, clients can customize and manage their search solutions while maintaining control over their data and access permissions.

API Keys provide secure, controlled access to our search services, enabling seamless integration and customization. They ensure that only authorized users can interact with our platform, enhancing security and efficiency.

Import templates

Import templates make it easy to import user data feeds. This process speeds up data extraction, allowing LupaSearch to access data from APIs, XMLs, and other public sources efficiently.

Permission controls can also be applied to import templates for increased security.

Documents

A document is any searchable item in your index. Most commonly, documents represent products - but you can also index:

• Categories
• Tags
• Blog posts
• Documentation
• Any other website content

Requirements

Before importing documents, ensure:

• Your index has a defined mapping - a schema that configures document fields and their data types
• Every document includes an id field

Document import

The Document Import feature simplifies product and content management by allowing you to upload a JSON file directly from the dashboard to:

• Create new documents
• Append additional documents
• Replace existing documents

This enables immediate updates without waiting for PIM synchronization and can act as a backup mechanism for quickly restoring or correcting product data.

Document structure

Each document represents a searchable unit in LupaSearch. While products are the most common use case, any structured content can be indexed and queried.

To format documents correctly, follow the guidelines in Documents

Mapping

Mapping defines how your products (or other relevant searchable objects) are stored and indexed. You do not need to describe all attributes of your dataset. Make sure to define attributes you want to use later in the search configuration. You can add various attributes like text, numeric, and other data types.

Consult with LupaSearch product consultants if you have any questions or concerns regarding effective mapping. Of course, you can always set it up on your own. Find out instructions in the Mapping.

Synonyms

Synonyms are the feature that enables the search engine to create equivalences among different search queries. Even if your product feed does not include all possible search query variations, you can link groups of words to each other as synonyms.

When a user searches for a product in his own words, the Synonym feature will return the result of an equivalent product. (For instance: “trousers” = “leggings”; “shoes” = “boots”, “cake” = “pie”, etc.)

LupaSearch offers three distinct types of synonyms to enhance customers’ search experience:

1. Manually Added or Uploaded Synonym.
   Users can manually add synonyms or upload predefined lists through the dashboard. This allows for tailored control over specific terms and their associated synonyms.

2. Similar Phrase Synonym Suggestions.
   The system leverages user behavior and search trends from a specified timeframe to automatically suggest synonym phrases. These suggestions are intelligently generated, and users can easily approve them in the LupaSearch console.

3. AI-Generated Synonyms with OpenAI.
   LupaSearch has integrated the powerful OpenAI algorithm for advanced synonym generation. OpenAI automatically provides synonyms for user queries, understanding word distances and presenting more accurate search results. The AI-generated synonyms feature brings new opportunities for an even more precise search experience:

   1. Suggest synonyms as Similar Queries:
      In case of zero results for a user's query, the AI-generated synonyms suggest alternative queries more likely to yield results. For example, if a user searches for "cupcakes," suggestions like "cakes," "muffins," or "desserts" may be provided.
   2. Suggest synonyms as "Did You Mean" Suggestions:
      In case of zero results for a user's query, the AI-generated synonyms will be suggested as a “Did you mean” option.

   This feature is available in all languages supported by LupaSearch, ensuring a comprehensive and intelligent synonym management system for your site search solution.

Adding synonyms

1. Locate and click on the "Synonyms" section in the dashboard. Check if the correct “Index” is selected in the top-left corner.

2. Use the "Add" button to include specific terms relevant to your search queries.

3. Enter both the primary term and its synonym(s) in the provided fields. LupaSearch allows adding two types of synonyms: equivalent and directional.

   1. Equivalent synonyms - add similar words or phrases with similar meanings according to the context of the products in your e-shop. Consider these synonyms as interchangeable words that do not change the meaning radically.
      For instance, if you add the words “small”, “little”, “miniature”, “narrow”, and “short” as synonyms, the user query "small fridge" would match products with names like: “small fridge”; “fridge - short”; narrow”, “compact fridge”.
      Similar results would still be present for similar user queries, like a “little fridge” or “miniature fridge”. However, exact matches (like “a small fridge” in the example above) will always be displayed at the top of the search results, unless these rules are overridden with your custom query boosting.
   2. Directional synonyms - add jargon and slang words or phrases that your potential buyers would use to call the products (but you would not add them to the product feed). These synonyms are applied one way only and override the jargon words.
      For instance, if you add “kicks” as a directional synonym for “shoes”, it would match the user entries like: “shoes”; “sneakers”.
      However, the search query "sneakers" would never return results with slang or jargon words (that are not formally used to call the product). Therefore, consider directional synonyms as a one-way synonym replacement.

4. Click “✓” to confirm. A change is applied to the live production environment (usually instantly, however sometimes it might take up to 30 seconds for a change to be visible).

Alternatively, you can add synonyms using the API. Follow the instructions in the Synonyms.

Bulk adding synonyms

1. Locate and click on the "Synonyms" section in the dashboard. Check if the correct “Index” is selected on the top-left corner.

2. Click “Append New Synonyms” button. A dialog will appear allowing you to select and upload bulk synonym entries from a text file directly through the dashboard. Input a list of term-synonym entries (each synonym entry pair should appear on a new line), separating them by

   1. “,” for Equivalent synonyms - interchangeable words that do not change the meaning radically and are often present in your product feed;
   2. “=>” for Directional synonyms - jargon and slang words or phrases, override the jargon words. Phrases on the left side of the “=>” usually should not appear in your product feed.

3. Validate and manage the uploaded synonyms list. The dashboard allows listing, filtering, deleting, and adding new synonym entries.

Replacing synonyms

1. Locate and click on the "Synonyms" section in the dashboard. Check if the correct “Index” is selected in the top-left corner.

2. Click “Replace Existing Synonyms” button. A dialog will appear allowing you to select and replace existing synonym entries by uploading a text file using the same format as with “Bulk Adding Synonyms”.

3. Input a list of term-synonym pairs, separating them by

   1. “,” for Equivalent synonyms - interchangeable words that do not change the meaning radically and are often present in your product feed;
   2. “=>” for Directional synonyms - jargon and slang words or phrases, override the jargon words. Phrases on the left side of the “=>” usually should not appear in your product feed.

4. Validate the uploaded synonyms list. The dashboard allows listing, filtering, deleting, and adding new synonym entries.

Exporting synonyms

1. Locate and click on the "Synonyms" section in the dashboard. Check if the correct “Index” is selected on the top-left corner.

2. Click the "Export Synonyms To File" button.

3. The system will generate a file containing the full list of created synonyms.

4. Save the file to your local device or download it as prompted.

5. You can use this file to import synonyms to other indices or projects in LupaSearch within your organization.

Approving suggested synonym

1. Locate and click on the "Synonyms" section in the dashboard. Check if the correct “Index” is selected in the top-left corner.

2. Navigate to the “Show similar phrase synonym suggestions" section.

3. The system actively detects and suggests possible synonym phrases by analyzing user behavior and search trends. A new list of synonym suggestions is automatically generated weekly.

4. Review the AI-generated synonym suggestions within the designated dashboard section.

5. Click “⊕” in the suggested similar phase line and approve the suggested synonym phrases based on their relevance to your search context. Approved synonyms are automatically added to the bulk list of synonyms and applied to the live production environment (changes might take up to 30 seconds to appear).

Enable AI-generated synonyms with OpenAI

1. Locate and click on the "Search Queries" section in the dashboard.

2. Select any search query that is used in your search and select “Edit search Query”✐ action

3. Find “AI Synonyms” section in the configuration panel and

   1. Enable AI Synonyms;
   2. Enter a numerical value for the threshold. This value signifies the maximum number of results allowed for the original user query before AI Synonyms come into play.
      For instance, adding the number “5” as a Threshold extends search results with AI synonyms only if the original user query returns less than 5 product results.
      Adjust the threshold based on your preference for balancing precision and recall in search results. A lower threshold may trigger AI Synonyms more frequently, while a higher threshold may limit their application to specific cases.
   3. Select behavior: “Similar queries” - AI synonyms will be returned as additional similar queries in the response; “Did you mean” - AI synonyms will be added into the main results with the same format and logic as “Did you mean” spellcheck results.

4. Click the “Test” button and safely test how the new configuration works directly from the search test sidebar on the right, without affecting any queries that are running in production.

5. Click “Publish” and AI Synonyms will be instantly applied to the live production environment.

Search queries

Search Query is a configuration that describes how data should be queried from your index. For example, you can set query fields, aggregations (facets), initial filtering, sorting rules and many other settings.

Flexible search query configuration

You can have an unlimited amount of search query configurations - for each of your data sources. You can search your data in a variety of ways with our flexible query configurations - and even edit while it’s in production - with zero downtime or redeployment.

Each Query can be configured to search your data in several ways. Examples of use cases include, but are not limited to:

• Searching for documents by only title or name;
• Searching by unique identifier;
• Searching by both name and description;
• Configuring whether all user query words should match the document;
• Adding predefined facets, filters, sorting, and pagination settings.

For more information, click here: Query Configuration

To access rule configuration, click the following button in the Query List table:

Then, go to the Select fields - which control which fields of your document are returned together with the search results.

Next, go to the Query fields, which control which fields of your document are used to perform the search, along with the provided weight - field 'importance' value. Add the relevant field and the boost factor as seen in the screenshot:

Then, you will find several options for boosting your search results: numeric boost, value boost, and statistical boost, as well as exact phrase boost.

You will also find other configuration options in the screenshot below:

Then, press Publish to save your configuration or Test to see how it works before saving the final configurations.

Query rules

To create rules, go to the Manage Rules section and click on it as in the image below:

Then, you can create new rules by clicking on the “Create new” button”:

You will see a Rule section, where you can find several options for setting up rules.

When you click on “Add” in the “Rule condition” category, you will find the rule search text condition:

Rule conditions are used to describe the conditions when rule action should be applied. Each rule can have multiple conditions, which can be matched in two different ways (see Rule options section).

The next step is to click on the “Publish” or “Test” button, depending on what you want to do.

Custom ranking

Custom ranking functionality helps reorder products for each search query.

1. Click “Search queries” in the console menu.

2. Find your search query in the table and click Custom Ranking.
   Different search queries can be for different page functionality, like different languages - if you are not sure which to choose, contact us at [email protected]

3. Create a new ranking rule:

4. Just type in any search phrase (it is possible to add multiple) you want to change ranking for, and click Enter:

You will see the current ranking for selected search query!

5. Changing the order

You can use the mouse to drag and drop the current search ranking.
You can use double arrows to move chosen items to the very top or very bottom of the list

6. (Optional) Adding more items

If you do not see the required item in the top 30, you can use “Add Items To Search Results” to include additional items to phrase ranking:

You can use the search functionality to find the required item.

Click on items you want to include and add them to the results.

Set the affected item limit

By default, full ranking of 30 items will be fixed in place and will always return these same results for the user. If you want to only affect the top 5 or top 10 results, you can use this setting:

And type in the number of items you want to “fix in place”:

7. Save the changes

Finally, to apply the rule changes, you just need to click the “Create” button:
You will notice that the term you wanted to rank higher, will appear in the first result(s) in the search results as in the above screenshot.

Marking items as boosted (sponsored)

In response to EU regulations and client preferences, there is an option to manually label sponsored products on the search results page.

LupaSearch can mark products as “sponsored/boosted” when a custom ranking or boosting rule is used. This will help your e-shops more easily align with legal requirements when sponsored items have to be marked in the search results. Based on your business preferences and needs, LupaSearch or your IT team can also create and add a custom label design.

Marking boosted items in custom ranking

1. Find your main search query to apply custom ranking to:

2. Create a new custom ranking rule by choosing a search keyword(s)

3. Reorder items manually, or add new ones

4. (Optionally) Limit number of items affected by custom ranking

5. Turn on the “Mark matching items as boosted” setting and save settings

6. Save the custom ranking rule

7. Observe the results in the demo environment:

Marking boosted items in custom rules

1. Find your main search query to apply a custom rule to:

2. Create a new rule (or choose old one if desired rule is already created)

3. Fill in the rule condition (choose when the rule is applied)

Example: only apply the rule when the user is searching for the exact keyword “consoles”

4. Choose rule action (Numeric Boost, Value boost or Must Include).

Example: boost items that have brand “Microsoft”

5. Turn on “Mark matching items as boosted” setting:

6. Click “Test” button to observe the results

Notes for developers

Each boosted product in its public query response data will have a new special field _boosted: true. You can use this field to apply any design changes to the relevant product card.

Plugin configurator

Quickly integrate powerful search functionality into your e-commerce store with our easy-to-use client plugin. Enjoy a clean, default interface and essential LupaSearch features without the complexities of custom integrations. Simply add one script and you're ready to enhance your customer search experience across desktop and mobile devices.

GenAI

Offer conversation-like search experience to deliver hyper-relevant search results.

Our GenerativeAI Search delivers precision unmatched by traditional search. By understanding the context of your query, we provide highly relevant results, filtering out irrelevant information. Whether you're searching for travel destinations, products, or services, our AI-powered technology ensures you find exactly what you need.

Cutting-edge technology is always evolving. Gather feedback and iterate fast

This cutting-edge technology is constantly evolving, and we're excited to offer you early access to its potential through Lupa Gen-AI, which is currently a Minimum Viable Product (MVP). This initial version focuses on delivering key features to enhance on-site search, providing personalized and efficient results. As an MVP, it allows us to gather valuable feedback and iterate quickly to meet evolving customer needs.
Front-end solution from the customer for enhanced on-site search experience

Integrating Gen-AI for on-site search involves a seamless connection with the customer's front-end solution. This collaboration enhances the search experience by delivering personalized and dynamic results directly to users.

By bridging Gen-AI with front-end solutions, e-commerce businesses can create an intuitive and efficient search experience that meets customer expectations and drives sales growth.

Enjoy conversation-like search that converts

Experience the future of search with LupaSearch's GenAI. Chat like you would a friend, get product results instantly. Our AI understands your needs and delivers tailored recommendations. For example, ask: “I just devoured 'Dune.' It's incredible! Any recommendations for similar reads?", “What ingredients do I need to make a creamy champignon soup?”

Lupa GenAI simplifies complex search queries

Lupa GenAI simplifies complex search queries by handling multiple criteria like price, brand, color, size, rating, and availability. For example, ask: “Show me red dresses under $50 with at least 4 stars in stock,” and the AI will find matching products.

It learns from user feedback to offer a personalized approach

LupaSearch GenAI learns from your choices, offering tailored suggestions based on your preferences.

Recommendations

Recommendation module can be used for:

• Similar/Related Products/You May Also Like
  • Based on product similarities (e.g., attributes like color, size, or brand)
  • Displayed on product page
• Alternative Products
  • Based on product category and custom rules (e.g., inventory levels or higher-margin alternatives)
  • Displayed on product page
• Trending Products / Popular Products /Best-Selling Products
  • Based on search clicks, add to cart, or/and orders information.
  • Displayed on homepage, category page (in the same category)
• Customers Also Bought / Frequently Bought Together
  • Based on cart and order item combinations (e.g., complementary or bundled products)
  • Displayed on product page, shopping cart and pop-up
• Customers Also Viewed
  • Based on visitor behavior data like item clicks.
  • Displayed on product page
• New Products
  • Based on products tagged as "new" in the product feed or catalog updates.
  • Display on homepage, category page

Recommendation queries can be configured in LupaSearch Console. To access configuration, select the desired index and choose "Recommendation Queries" in the main menu: Recommendation Queries.

When creating/editing a query you will have the ability to test your recommendation configuration without publishing it - in the same way as with regular search queries.

Recommendation queries overview

Recommendation queries allow for flexible product suggestions. Here are some examples:

• Suggest similar products based on attributes.
• Recommend products based on user actions during an on-site search session.
• Suggest products frequently bought together.

Sources for Recommendations

A recommendation query can utilize one or multiple sources. LupaSearch supports the following sources:

1. Similar attributes. This source recommends products based on their static attribute similarity. It uses configured query fields and facets derived from a related documentQueryKey. It does not have additional options.
2. Same session clicks. This source recommends products based on user behavior within a single search session. If a pair of products is frequently clicked by multiple users within the same session, these products will be considered related. This source requires the implementation of the Events API. Available options include:

• Weight: the weighted importance of this source for recommendations.
• Max age days: the maximum session occurrence age in days to be included in recommendations.
• Min. occurrences: the minimum number of co-occurrences of two item clicks in the same search session for the items to be related.
• Top session count: the number of top sessions with the given product ID to include in calculations (larger values might affect performance).

3. Clicks by the same user This source recommends products based on user behavior across multiple search sessions. If the same pair of products is frequently clicked by multiple users, they will be considered related. This also requires the Events API implementation. Available options include:

• Weight: The weighted importance of this source for recommendations.
• Max age days: The maximum item click age in days to be included in recommendations.
• Min. occurrences: The minimum number of co-occurrences of two item clicks by different users for the items to be related.
• Top session count: The number of top sessions with the given product ID to include in calculations (larger values might affect performance).

4. Bought together recommendations This source recommends products based on occurrences of product pairs in a single order (user cart). This requires the Events API implementation with conversion events. Available options include:

• Weight: The weighted importance of this source for recommendations.
• Max Age Days: The maximum conversion age in days to be included in recommendations.
• Min. Occurrences: The minimum number of co-occurrences of two items in a cart for the items to be related.

Different sources can be combined in a single query by assigning each source a desired weight.

Manually editable Recommendation Queries

You can manually create item rules based on specific business-relevant criteria such as brand, category, color, and more. Additionally, you can designate particular products to be recommended alongside others.

This feature allows for highly targeted recommendations. For instance, you can set a specific query for “iPhone 14 Pro cases” and configure the e-commerce search to suggest your three best-selling phone cases on all iPhone 14 product pages.

Redirections

With the help of LupaSearch, shoppers can find relevant product results using site search. However, there are times when you might want a specific search query to direct users elsewhere. For this, you can use the Redirects page in the Management Console.

Redirections can be configured on the Redirections page and are used to redirect users to a custom route or external page when a specific search keyword is entered.

Add new redirects

New redirects can be added by clicking on the “Add” button. Type the keyword you want users to be redirected from to another page and in the Redirects field, add the URL where you want the customer to be sent as in the below example.

A/B testing

Search A/B testing is your secret weapon for optimizing search results. By simultaneously testing different search configurations, you can identify what resonates best with your customers.

Choose the key elements to test, split your audience, and analyze results and increase your search performance. Gain valuable insights into user behavior and make data-driven decisions to improve conversions and customer satisfaction.

How does AB testing work?

In the LupaSearch Console or via an API, you can set up AB Tests.

Each AB Test includes:

• Name: The name assigned to the test.
• Triggering Query: The specific query that activates the test. (The test will be triggered if a user includes this query key in a public query request.)
• Queries: A list of queries used in the test. (Each query has a trigger chance percentage that indicates how frequently it will be used.)
• Active From / Active To: The time period during which the test is active.
• Enabled: A flag indicating whether the test is active. (The test will not be triggered if it is disabled.)

When an AB Test is active, LupaSearch monitors various statistics related to query engagement (such as clicks, CTR, and no-results rate) to help you assess which query performs better.

Benefits of on-site search AB Testing

You can use AB Testing to evaluate different search configurations, such as:

• Assessing whether additional query fields reduce the zero-results rate.
• Determining if enabling Personalization or AI Synonyms improves CTR.
• Comparing which default sorting or boosting rules lead to a higher click-through rate.
• Evaluating if different facets or facet ordering rules increase the estimated Add To Cart value.

What you should test

The possibilities of what you can test are many. For instance, you can test the following:

• Personalization enabled vs. disabled
• Different sorting rules
• Boosting bestsellers vs. trending search results

After conducting these tests, you can analyze detailed results, visual statistics, and graphs to have a better understanding of your on-site search configuration and make data-driven decisions for your e-commerce business.

AB Testing Recommendations

When setting up AB Tests, keep these recommendations in mind:

Use target queries only in A/B testing - ensure the target queries are used only for A/B testing purposes and not anywhere else in the search configuration. It avoids confusion and simplifies results. Always refer to the target query using the trigger query (searchQueryId).

Include sessionId and userId: make sure to include both sessionId and userId in /query and /event requests to track user behavior and analyze results. If either userId or sessionId is missing from these requests, the AB test may not be triggered. Ensure that at least one of these identifiers is present.

How to configure A/B testing in the dashboard

1. To configure AB Testing using Dashboard, go to the AB Testing section in LupaSearch Console:

2. Click the Add AB Test button to create a new AB Test:

3. Fill in the form, add trigger query and queries and click the Save Changes button. You can optionally change query trigger percentages manually (by default, they all will be equal):

4. Let the test run for a few days and then check the results using the "Eye" icon:

You will see a page with the AB Test results summary.

The page includes the following information:

• Main ab test details - name, active period
• Enabled flag
• Trigger query

Summary

Summary of the test results, including total number of each query executions, CTR, no results rate, and estimated add-to-cart value.

Query Charts - charts with daily stats for each query, including: number of executions, CTR, no results rate and click events.

High and Low CTR Terms - list of terms that have the highest and lowest (among TOP 100 most popular terms) CTR for each query.

Configure AB testing using API

To configure AB Testing using API, you can find more details here: A/B Testing

Cloning configurations

You can clone configurations. This functionality is especially useful for making minor adjustments. For instance, it is widely used in setting up A/B testing. You can create and set up one configuration (A), clone it, and then modify only the B element. This feature saves time and reduces the risk of accidental errors.

How to update AB testing

To update an AB testing, go to the Actions column and click on the first icon as in the below screenshot:

After you click on the edit icon in the Actions column, you will see the below pop-up from where you can update the AB testing:

Keywords

Keywords can be used to mark special words or phrases that will be excluded from grammatical stemming. This is often used for product brands which are not supposed to be stemmed. It can be especially useful in non-english languages where stemming can be more aggressive.

Important: to fully apply the keyword changes, you need to re-index your document index after each action