Marketing - All in One: Creating and Configuring Data Sources

A Data source is the core configuration unit in Marketing - All in One that controls where product recommendations come from and how they're generated for your storefront. Each data source applies to a specific placement — such as your homepage, a product page, or search results — using the recommendation strategy, trigger rules, and exclusion settings you define, and the same data source can be reused across multiple components. 

This guide explains how to create a data source, configure its recommendation strategy and rules, exclude specific products, and manage your data source list.

Prerequisites for Creating a Data Source

Before you begin, make sure:

• You have installed and authorized the Marketing - All in One app. If this is the first time you're installing or updating the app, you'll need to agree to the updated privacy policy before you can use this feature.

Creating a Data Source

To create a new data source, follow these steps:

1. Go to the Marketing - All in One app, then click Data source in the left navigation.

   

2. Go to the data source creation page:
   • If this is your first data source: click Get started on the page.
   • If your store already has data sources: click Create data source in the top-right corner of the page.
3. In the Name field, enter a name for the data source (up to 50 characters).

4. In the Placement section, select the page where the data source will be used. Available options include:

   

   • Home
   • Product details
   • Product listing
   • Search results
   • Cart
   • Account
   • 404 page
   • Blog
   • Custom page

   Note: The Cart drawer and Payment result placements are not yet supported and appear grayed out. Support for these will be added in a future release.

5. In the Recommendation strategy section, select the AI recommendation strategy the data source will use. The system automatically shows the strategies available for the placement you selected in the previous step, including:

   

   • AI recommendations: Analyzes the full customer profile to surface the products most likely to convert.
   • Trending now: Surfaces popular products based on views, clicks, and sales across your store.
   • Similar customers: Analyzes customer profiles and recommends products that similar customers have bought.
   • Related products: Analyzes product data across your store and recommends items similar to the product on the current page.
   • Frequently bought together: Recommends products that customers often buy together, based on past orders.

6. Click Create.

Once the data source is created, you're automatically redirected to its details page, where you can continue configuring the recommendation strategy and exclusion rules

Managing Basic Information for a Data Source

On the data source's details page, you can update the following basic information at any time:

• Name: Change the data source's name.
• Maximum number of products to recommend: Set the upper limit on the number of products the data source returns. You can enter an integer from 1 to 50; the default is 50.
• Placement: Shows the placement configured for the data source. This can't be changed after the data source is created.
• Connected components: Shows the components currently connected to the data source. Click a component name to open its settings page in a new tab for editing. If no component is connected, click Go to components to open the Components page and connect one. For details on creating components and applying data sources, see "Marketing - All in One: Creating and Configuring Components".

After making changes, click Update to save them.

Configuring the Recommendation Strategy for a Data Source

The recommendation strategy controls when products are recommended in a given scenario, and which products are recommended. Each rule contains one trigger condition and one or more actions.

Understanding What Makes Up a Recommendation Strategy Rule

Each rule contains the following settings:

• Trigger condition: Determines when the rule is applied. You can select Any condition, or set a specific condition based on the applicable scenario. For example:

  • Product details page: Current product
  • Product listing page: Current collection
  • Search results page: Keyword
  • Cart page: Products in cart

  Notes: Each trigger condition type supports up to 100 values. Keyword-based conditions use exact matching and are not case-sensitive.

  

• Action: Determines which products are recommended once the trigger condition is met. You can add up to 5 actions per rule, choosing from the following types:

  • AI strategy: Uses one of the AI recommendation strategies, such as AI recommendations, Trending now, Similar customers, Related products, or Frequently bought together. You can add one AI strategy action per rule.
  • Specific products: Manually select which products to recommend. You can add one Specific products action per rule, select up to 50 products, and set a sort order (for example, best-selling in the last 90 days or newest first).
  • Recommendation rules: Filters which products are recommended based on conditions such as collection, tag, product title, variant price, variant compare-at price, or inventory. You can set up to 5 filter conditions per action and combine them using AND logic. You can add multiple actions of this type, as long as the total number of actions in the rule doesn't exceed 5.

  

Understanding Rule and Action Priority

If a data source contains multiple rules, they're matched from top to bottom. Once a rule's trigger condition is met, the system runs that rule and stops evaluating the rules below it.

If a rule contains multiple actions, the system merges the recommended products in the order the actions are listed:

• Actions listed higher up are given priority when recommending products.
• Once the maximum number of products to recommend is reached, results from lower-priority actions are no longer shown.
• If multiple actions recommend the same product, it's shown only once, using the result from the highest-priority action.

For this reason, we recommend placing your more specific or higher-priority rules and actions first.

Adding and Managing Recommendation Rules

To add or adjust a data source's recommendation rules:

1. On the Data source details page, click the down arrow next to Recommendation strategy to expand the section.

   

2. Configure the rule as needed:

   • Click + Actions to add an action.
   • Click the up and down arrow icons next to an action to change its order.
   • Click the X next to an action to remove it.

   

3. To add another rule, click + Add rule at the bottom of the rule list. New rules default to the Any condition trigger and the AI recommendations action.

   

4. To reorder rules, drag the handle icon in the top-left corner of a rule card.

   

5. To remove a rule, click the X in the top-right corner of the rule.

   Note: Each data source must keep at least 1 rule, so you can't remove a rule if it's the only one left.

   

6. Once you've finished configuring the rules, click Update to save your changes.

Excluding Specific Products from a Data Source

If some products aren't a good fit to recommend to customers, you can use the Excluded products settings to leave them out of the recommendation results. This section is collapsed by default — click it to expand and configure it.

1. Click the down arrow next to Excluded products to expand the section.

   

2. Configure the products you want to exclude, as needed:

   

   • Exclude specific products: Click + Select products to choose the products to exclude. You can exclude up to 100 products per data source.
   • Exclude specific collections: Click + Add collections to choose the collections to exclude. Products that belong to the selected collections won't appear in the recommendation results. You can exclude up to 5 collections per data source.
   • Exclude out-of-stock products: Choose whether to exclude out-of-stock products. The default is Yes.
   • Exclude parent products: Choose whether to exclude parent products used in composite product listings. The default is Yes.
   • Exclude child products: Choose whether to exclude child products used in composite product listings. The default is No.
3. Once you've finished configuring these settings, click Update to save your changes.

Understanding Which Recommendation Strategies Each Placement Supports

The AI recommendation strategies available to you depend on the placement you choose, as shown below:

Managing the Data Source List

On the data sources list page, you can view, search, filter, and manage the data sources you've created for your store.

1. Click Data source in the left navigation to go to the data sources list.

2. Find the data source you need:

   • Use the search box to search by exact Data source ID or Data source name.
   • Click the Placement filter to filter by one or more placements.

   

3. To manage a data source, click ··· in the Actions column next to it, then choose:

   • Edit: Opens the data source's details page for editing.
   • Delete: Deletes the data source.

   

Important Notes for Data Sources

• Deleting a data source can't be undone, and will cause any component connected to it to stop displaying recommended products.
• The first time you install or update Marketing - All in One, you'll be asked to agree to the updated privacy policy. You won't be able to complete installation, authorization, or access the app until you do.
• You can find a link to the related privacy policy at the bottom of the data sources list page.

Frequently Asked Questions About Data Sources

Q1: What happens to a component connected to a data source after the data source is deleted?

The component itself isn't deleted, but since it no longer has a data source, it can't display recommended products on your storefront. The component list shows a "Data source deleted" message for that component.

Q2: Can one data source be connected to multiple components at the same time?

Yes. As long as the data source and the components share the same placement, the same data source can be connected to and reused across multiple components.

Q3: What happens if a rule needs more than one type of trigger condition?

Currently, each rule supports only 1 trigger condition — combining multiple condition types within a single rule isn't supported yet.