# Search & Recommendations

## Introduction

Search & Recommendations extension enhances selected Sales Channels by optimizing Product Listing Pages (PLPs), refining search word predictions, and improving filters. The data flow involves transferring information from the platform to Algolia, where it is stored and powers storefront listing pages. While the data can be obtained through a catalog data loader, the Algolia extension elevates accuracy in results and search predictions, optimizing search-to-result time.

Key functionalities include indexing, sorting, and mapping for effective data organization. The Auto Complete Menu provides quick, intuitive suggestions, and Search Analytics offers insights into user behaviour. The Data Loader ensures seamless data transfer, while the File Uploader simplifies file uploads. In summary, the Algolia extension comprehensively enhance the search experience, optimizing the functionality and efficiency of Sales Channels in e-commerce.

## Steps to use Search & Recommendations Extension

1. Go to **Extensions** in Fynd Commerce.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/1.png)

center
em
Figure 1: Extension
br
1. Open **Search & Recommendations**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/2.png)

center
em
Figure 2: Opening Search & Recommendations
br
1. Choose from the available integrations:
![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/3.png)
Figure 3: Extension home page

  * **Common** – Manage global settings.
  * **Algolia** – Optimise search using Algolia APIs.
  * **Google Retail Search** – Sync and manage search with Google.


## Using Algolia in Search & Recommendations

The **Algolia** integration provides advanced tools to configure search ranking, sort orders, rules, relevance, analytics, and more. *Open Algolia to get started*.

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/4.png)

center
em
Figure 4: Opening Algolia
br
### Set up Algolia Credentials

To activate the integration, first set up your credentials:

1. Open **Credentials**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/5.png)

center
em
Figure 5: Opening Credentials
br
1. Enter the Algolia's API Keys in this section to complete the setup for the Algolia extension.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/6.png)

center
em
Figure 6: Setup 
br
1. Go to [algolia.com](https://www.algolia.com/).
2. Select **Dashboard** > **Overview**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/moonlight/Extensions/Algolia/algolia.com-dashboard.png)

center
em
Figure 7: Overview 
br
1. Select the company from the drop-down. Here, we have selected **Fynd** company.


![QG1](https://cdn.pixelbin.io/v2/doc/original/moonlight/Extensions/Algolia/algolia.com-indexdropdown.png)

center
em
Figure 8: Fynd Application 
br
1. Go to **API Keys**.
![QG1](https://cdn.pixelbin.io/v2/doc/original/moonlight/Extensions/Algolia/algolia.com-apikeys.png)
Figure 9: API Keys

  1. Copy the **Application ID** and paste it into the **Application ID** field on the Fynd Commerce.
![QG1](https://cdn.pixelbin.io/v2/doc/original/moonlight/Extensions/Algolia/algolia.com-applicationID.png)
Figure 10: Application ID

  1. Copy the **Search API Key** and paste it into the **Search-Only API Key** field on the Fynd Commerce.
![QG1](https://cdn.pixelbin.io/v2/doc/original/moonlight/Extensions/Algolia/algolia.com-searchAPIkey.png)
Figure 11: Search API Key

  1. Copy the **Write API Key** and paste it into the **Admin API Key (or Write Key)** field on the Fynd Commerce.
![QG1](https://cdn.pixelbin.io/v2/doc/original/moonlight/Extensions/Algolia/algolia.com-writeAPIkey.png)
Figure 12: Write API Key

2. Select the **Search** module from the left side panel and go to **Index** under **Configure**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/moonlight/Extensions/Algolia/image(4).png)

center
em
Figure 13: Index 
br
1. Go to the **Index** drop-down in the header and select the relevant index from the drop-down menu, and copy the **Index Name**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/moonlight/Extensions/Algolia/algolia.com-indexdropdown.png)

center
em
Figure 14: Index Name 
br
After entering the details in the aforementioned fields, other modules of the Algolia extension will be enabled for you to configure.

After setting credentials for an extension, only the index_name can be modified.

1. Add **Global** **Context**. This refers to universal settings that impact the overall behaviour of search functionalities across the entire application. This field is non-mandatory.
2. Add **Target Mart**. It is specific to Jio. Target Mart can be either Jio or Smart. This field is non-mandatory.


**Fill in the following fields with the copied details:**

| Field | Description | Example |
|  --- | --- | --- |
| **Application ID** | Unique ID to identify your Algolia project. | `3kb42k3v42u42jkjb32k4` |
| **Search-Only API Key** | Public key used on the frontend for search queries and analytics. | `23brkbjb23jb4j234j12n3b4l` |
| **Admin API Key** | Admin or Write API key. Keep this confidential. | `***********` (masked) |
| **Index Name** | The name of the index to store and retrieve data. | `Enter Index Name` |
| **Global Context** | Default search context. | `Global Context` |
| **Target Mart** | The mart type to target (choose either `JIO` or `SMART`). | `SMART` |


Click **Save** to store the credentials.

For more information on price internationalisation, click the *info link* above the form.

### Configure Sort & Replica Mapping

Control how your products are sorted and displayed in search results.

#### Steps to Configure Sort Options

1. Select **Sort & Replica Mapping** from the Algolia dashboard.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/7.png)

center
em
Figure 15: Opening Sort & Replica Mapping 
br
1. Review available sort options such as:
![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/8.png)
Figure 16: Sorting Options

  * Price Low to High
  * Popularity
  * Discount High to Low
  * Latest
  * Discount Low to High
  * Price High to Low
  * Rating
  * Size Depth (High to Low)
2. Select the sorting option and then in the Settings section:
![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/9.png)
Figure 17: Index Name

  * Update its **Index Name** (e.g. "AK_INDEX_PRICE_L_H").
3. Click **Save**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/10.png)

center
em
Figure 18: Clicking Save 
br
Clicking “click over here” in the info banner will take you directly to this settings screen.

### Enable Search Analytics

Use **Search Analytics** to track user behaviour and improve search performance.

1. Go to **Search Analytics** from the Algolia dashboard.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/11.png)

center
em
Figure 19: Opening Search Analytics
br
1. Enable **Click Analytics** to start sending `query_id` from Algolia to the front end, enabling click and conversion tracking.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/12.png)

center
em
Figure 20: Enabling Analytics
br
1. It enable **Search Analytics** to view:
  * Total searches
  * No result rate
  * Daily search activity
2. Click **Visit Algolia Dashboard** to access these insights directly on Algolia.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/13.png)

center
em
Figure 21: Clicking Visit Algolia Dashboard 
br
This link opens the official Algolia dashboard in a new tab.

### Upload Rules Using File Uploader

Create and sync bulk rules in Algolia using a structured Excel upload.

**Steps to Upload Rules**

1. Go to **File Uploader** under Algolia.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/14.png)

center
em
Figure 22: Opening File Uploader
br
1. Download the **Algolia Rules Demo Excel** file.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/15.png)

center
em
Figure 23: Downloading Algolia Rules Demo Excel
br
1. Fill in the spreadsheet using the format below:


| Column | Description | Example |
|  --- | --- | --- |
| rule_id | Identifier for the rule | `my_rule_1` |
| condition_type | Type of condition (`query`, `context`, `filter`) | `query` |
| query_type | Query operator (e.g. `is`) | `is` |
| query_value | Term to match | `juice` |
| context_name | Context grouping | `testing` |
| context_scope | Scope of the context (e.g. `all`) | `all` |
| filter_name | Filter to apply (e.g. `color`) | `color` |
| filter_value | Filter value (e.g. `red`) | `red` |
| from_date | Start date of rule validity (MM/DD/YYYY or DD/MM/YYYY format) | 6/9/2023 |
| to_date | End date of rule validity | 6/15/2023 |


1. Click **Upload File** and select the completed `.xls` or `.xlsx` file.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/16.png)

center
em
Figure 24: Clicking Upload File 
br
1. Monitor the upload status under **Uploaded Rules**:
  * Success
  * Failed
  * In Progress


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/17.png)

center
em
Figure 25: Uploaded Rules 
br
Example - A file named `sample_algolia_rule (4).xlsx` with status `SUCCESS` confirms that rules were synced.

### Configure Relevance Settings

Relevance settings allow you to fine-tune the way search results appear to customers. These settings define **how queries are interpreted**, **how results are ordered**, and **which terms are treated as equivalent**.

Go to **Relevance** under Algolia.

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/18.png)

center
em
Figure 26: Opening Relevance
br
To configure relevance settings, you will be redirected to the **Algolia Dashboard** by clicking the **Set up in Algolia Dashboard** button.

#### Searchable Attributes

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/19.png)

center
em
Figure 27: Searchable Attributes
br
Defines **which fields of your data are searchable**, and in what order.

| Purpose |
|  --- |
| You want to control how users find products (e.g., search by brand, category, or product name) |
| You want to prioritise certain fields over others in the search results |


**Example Configuration:**


```
[
  "product_name",
  "category",
  "brand",
  "description"
]
```

Put the most relevant field (e.g. `product_name`) first. This ensures better matches appear at the top.

#### Custom Ranking

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/20.png)

center
em
Figure 28: Custom Ranking 
br
Controls **how matched results are ordered** after applying the search logic.

| Use case | Example |
|  --- | --- |
| Show bestsellers first | `bestseller` |
| Prioritise highest rated items | `rating DESC` |
| Highlight most-viewed items | `views_count DESC` |


**Example Configuration:**


```
[
  "desc(bestseller)",
  "desc(rating)",
  "desc(views_count)"
]
```

Custom ranking helps display more business-relevant results, such as top-sellers, first.

#### Synonyms

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/21.png)

center
em
Figure 29: Synonyms
br
Synonyms expand or unify search intent by telling Algolia **which words are considered equivalent**.

| Use case | Example |
|  --- | --- |
| Search for "pants" should also return results for "trousers" | `pants ⇄ trousers` |
| "jacket" includes "blazer" | `jacket ⇄ blazer` |


All of these settings are configured directly within the **Algolia Dashboard** after redirection.

### Manage PromoRules Context

This section allows you to assign **context** to rules, so they can be triggered conditionally based on customer behaviour, device type, or other metadata.

#### Steps to Use PromoRules Context

1. Go to **PromoRules Context** from the Algolia dashboard.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/22.png)

center
em
Figure 30: Opening PromoRules Context
br
1. View previously synced rules, including their:
![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/23.png)
Figure 4: Uploaded Rules

  * **Rule ID**
  * **Context** (e.g. `testing`)
  * **Type** (e.g. `all`)
2. Use the search bar to find specific rule contexts.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/24.png)

center
em
Figure 31: Search Bar 
br
1. Click the **Bin Icon** to delete a context.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/25.png)

center
em
Figure 32: Bin Icon 
br
Example - Rule ID `my_rule_1` with context `testing` and type `all` ensures this rule only applies during testing campaigns across all scopes.

* Promote a rule only during a sale period.
* Apply different rules for desktop vs mobile experiences.


### Enable and Track User Events

User Events: help track and measure key customer actions on your storefront. These events can be used to optimise search relevance and enable personalisation.

#### Steps to use User Events

1. Go to **User Events** in the Algolia section.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/26.png)

center
em
Figure 33: Opening User Events
br
1. Toggle the switch at the top right to enable.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/27.png)

center
em
Figure 34: User Events
br
1. You'll now see a list of predefined events:


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/28.png)

center
em
Figure 35: Toggling Events
br
| Event Name | Description | Mapped Algolia Event |
|  --- | --- | --- |
| **Add to cart** | Captures when a user adds a product to the cart | `addedToCartObjectIDs` |
| **Product view** | Captures when a user clicks on a product | `clickedObjectIDs` |
| **Order complete** | Captures when a user places an order | `purchasedObjectIDs` |


1. Toggle each event to enable it based on your tracking requirements.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/29.png)

center
em
Figure 36: Toggling User Events
br
1. Click **Save**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/30.png)

center
em
Figure 37: Clicking Save
br
Why use this? These events improve search performance by feeding real-time behavioural data into Algolia, enabling better analytics and personalised ranking.

## Google Retail Search

Google Retail Search enables enhanced product discovery by leveraging Google’s search intelligence. With this integration, merchants can optimise cataloguing, indexing, and personalisation, improving customer experience and search conversion rates.

Go to **Google Retail Search** section.

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/31.png)

center
em
Figure 38: Opening Google Retail Search
br
### Set Up Google Retail Search Credentials

To connect your Google Retail Search account with Fynd, you must provide key credentials.

1. Go to **Credentials** in the Google Retail Search section.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/32.png)

center
em
Figure 39: Opening Credentials
br
1. Enter the following information:


| Field | Description | Example |
|  --- | --- | --- |
| **Project ID** | Google Cloud project ID used for Retail Search | `retail-store-123` |
| **Dataset ID (Optional)** | Identifier for a specific dataset (optional) | `product_dataset_v1` |
| **Service Account Key** | JSON key for service account with access to Retail Search | `{...}.json` |
| **GCS Error Path** | Path to your Google Cloud Storage bucket for error logging | `gs://fynd-retail-errors/` |


Use case - These settings allow Fynd to sync your product data and fetch search-related insights from Google’s platform.

1. Click **Save** to complete the integration setup.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/33.png)

center
em
Figure 40: Opening Save
br
Clicking the info link will guide you to more details about price internationalisation and its impact on Google Retail configurations.

### Track User Events for Google Retail

User Events: help you capture behavioural data such as searches, views, cart actions, and purchases. These insights are sent to Google Retail to improve search relevance and product recommendations.

1. Go to **User Events** in the Google Retail Search section.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/34.png)

center
em
Figure 41: Opening User Events
br
1. Toggle the switch in the top-right corner to enable.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/35.png)

center
em
Figure 42: Enabling User Events
br
1. A list of available events will appear:


| Event Name | Description | Mapped Event |
|  --- | --- | --- |
| Add to cart | When a user adds a product to the cart | `add-to-cart` |
| Remove from cart | When a user removes an item from the cart | `remove-from-cart` |
| Product view | When a user views a product detail page | `detail-page-view` |
| Product listing | When a user views the homepage or a PLP | `home-page-view` |
| Order complete | When a purchase is made | `purchase-complete` |
| Product search | When a user searches for a product | `search` |
| Order checkout | When a user visits the shopping cart | `shopping-cart-page-view` |


1. Use the toggle next to each event to enable tracking.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/36.png)

center
em
Figure 43: Toggling User Events
br
1. Click **Save** once configuration is complete.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/37.png)

center
em
Figure 44: Clicking Save
br
Use Case - Enabling these events allows Google to understand how users engage with your site, enhancing the personalisation and ranking of search results.

## Common Settings

The **Common** section allows you to configure and manage universal settings for search services across your store. These include pricing, catalogue sync, attributes, autocomplete, and more. This centralises configuration across both Algolia and Google Retail search engines.

Open **Common**.

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/38.png)

center
em
Figure 45: Opening Common
br
### Config – Price Internationalisation

Manage API-based currency conversion for product listings.

#### How to configure:

1. Go to **Config** in Common Settings.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/39.png)

center
em
Figure 46: Opening Config 
br
1. Select the Exchange Partner from the dropdown & enter Open Exchange API Key.
![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/40.png)
Figure 47: Config Settings

  * Exchange Partner Example: *Open Exchange*
  * Open Exchange API Key Example: *fynd-openex-12345678*
2. Click **Save** to apply the settings.


Use case - Display accurate prices for international customers by converting your base price using real-time exchange rates.

### Catalogue Sync

Synchronise products, collections, and currency data between Fynd Commerce and external search providers (Algolia or Google Retail).

#### Steps to Sync:

1. Go to **Catalogue Sync** in Common Settings.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/41.png)

center
em
Figure 48: Opening Catalogue Sync
br
1. Click **Sync**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/42.png)

center
em
Figure 49: Clicking Sync 
br
1. Choose options from the modal:
  * **Sync Type**: Select `Incremental` or `Full`.
  * **Select Provider**: Choose from `Algolia`, `Google Retail`, or `Both`.
2. Click **Sync**.


#### Sync History:

You’ll see a job list with the following details:

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/43.png)

center
em
Figure 50: Sync History 
br
* **Job Type**: Products, Collections, Currency
* **Status**: SUCCESS / IN_PROGRESS
* **Items Synced**: Auto-filled after completion


Use case - Ensures your search providers have up-to-date product and collection data from the Fynd backend.

### Auto Complete Menu

Customise autocomplete suggestions shown in the search bar.

Go to **Auto Complete** Menu in Common Settings.

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/44.png)

center
em
Figure 51: Opening Auto Complete Menu
br
#### Available Auto Complete Menu Options:

Turn on the toggle to enable auto complete menu:

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/45.png)

center
em
Figure 52: Auto Complete Menu Options
br
* Product Suggestion
* Brand Suggestion
* Category Suggestion
* Query Suggestion
* Departments Suggestion
* Fallback to Algolia (used if Vertex data is empty)


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/46.png)

center
em
Figure 53: Clicking Save
br
Example: Enable **Query Suggestion** and provide a Query Suggestion Index name like `fynd_query_suggestions`.

Use case - Improves search UX by suggesting products, categories, or brands as users type.

### Custom Attributes

Create and manage custom attribute mappings for your product data.

#### Steps to configure:

1. Go to **Custom Attributes** in Common Settings.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/47.png)

center
em
Figure 54: Opening Custom Attributes
br
1. Toggle **Enable Custom Attributes**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/48.png)

center
em
Figure 55: Enabling Custom Attributes 
br
1. Click **Add New**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/49.png)

center
em
Figure 56: Clicking Add New
br
1. Fill in:
![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/50.png)
Figure 57: Adding Custom Attributes

  * **Source**: e.g., `product_colour`
  * **Destination**: e.g., `custom_colour`


Optionally, toggle:

* Remain Original: Retains the original source field in the output.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/51.png)

center
em
Figure 58: Enabling Remain Original
br
![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/52.png)

center
em
Figure 59: Clicking Save
br
Use Case - Re-map fields for clarity or compatibility before syncing to search services.

### Attributes

Select which attributes from your product data should be passed to Algolia.

#### Example:

* `uid`
* `multi_size`
* `slug`
* `is_available`
* `algolia_facet`


1. Go to **Attributes** in Common Settings.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/53.png)

center
em
Figure 60: Opening Attributes
br
1. Toggle **Enable Attributes**.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/54.png)

center
em
Figure 61: Enabling Attributes
br
1. Enter or delete fields as needed.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/55.png)

center
em
Figure 62: Attributes List
br
1. Click **Save**.


Use case - Reduce payload size or enhance indexing performance by syncing only required fields.

### Data Loader

Switch search-related data services from Fynd default to external providers.

Go to **Data Loader** in Common Settings.

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/56.png)

center
em
Figure 63: Opening Data Loader
br
Options available for:

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/57.png)

center
em
Figure 64: Data Loader Features
br
* **Product Listing**
* **Collection Listing**
* **Autocomplete Menu**


Each row includes two dropdowns, select as per your preference:

* Mode:
  * `Fynd`
  * `Ext Search Provider`
* Search Provider:
  * `Algolia`
  * `Google Retail`
  * `Hybrid`


Use case - Enables flexible routing of search APIs to different engines depending on A/B testing or performance.

### A/B Testing Config

Split traffic between search providers for testing purposes.

Go to **A/B Testing** Config in Common Settings..

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/58.png)

center
em
Figure 65: Opening A/B Testing
br
#### Configuration:

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/59.png)

center
em
Figure 66: A/B Testing Configuration
br
* **Algolia Percentage**: Set share of traffic (e.g., `50`)
* **Vertex Percentage**: Remaining traffic (e.g., `50`)
* **Rotation Time**: Time (in hours) after which user session switches provider


This applies only when Hybrid mode is selected in the Data Loader section.

#### Analytics Stats:

![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/60.png)

center
em
Figure 67: Analytics Stats
br
Enter **Start Time** and **End Time** to filter A/B test metrics by time range.

### Search Tool Pinning

The **Search Tool Pinning** feature allows you to assign a specific search engine to a particular user. This is helpful for internal QA, performance analysis, debugging, and A/B testing use cases.

**Steps to Pin a Search Engine to a User ID:**

1. Go to Search Tool Pinning .


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/61.png)

center
em
Figure 68: Opening Search Tool Pinning
br
1. Fill in the required fields:


| Field | Description | Example |
|  --- | --- | --- |
| **User ID** | The unique identifier of the user to pin | `user_123456` |
| **Search Engine** | Select the engine to assign from the dropdown | `Algolia`, `Google Retail`, `Hybrid` |


1. Click **Save** to finalise the configuration.


![QG1](https://cdn.pixelbin.io/v2/doc/original/searchlight/extensions/search-recommendation/62.png)

center
em
Figure 69: Clicking Save
br
Use case - If you're running multiple search engines and need to fix one engine (e.g. Algolia) for a tester or a group of users, this ensures consistent and predictable behaviour for them—bypassing A/B testing or dynamic rotation logic.

This pinning overrides the A/B testing and hybrid settings for the specified user ID.