Manage recommendation slots

This is for:

Merchandiser

Recommendation slots are specific locations within a webpage or application where personalized product recommendations are shown to visitors. These slots are strategically placed across Coveo-powered commerce storefronts to enhance the visitor experience, drive engagement, and increase sales by suggesting products relevant to the visitor’s preferences, context, and journey.

Merchandisers can create and manage recommendation slots using the Recommendations manager of the Coveo Merchandising Hub (CMH).

Note

It’s possible to create recommendation slots using the Commerce API, which can be a time-saving option for users managing multiple tracking IDs or storefronts.

Recommendation slot locations

In the Recommendations manager, recommendation slot locations (or slot types) refer to specific locations within a webpage or application where personalized recommendations are displayed. They help the merchandiser identifying the purpose of the slot within a storefront, ensuring they’re relevant and contextually appropriate.

Recommendation slots are found on:

  • The home page: Recommendations displayed on the home page, often in a scrolling carousel format.

  • A product listing page (PLP): Recommendations displayed alongside the list of products, helping users discover additional items that match their preferences.

  • A search page: Personalized recommendations embedded within or alongside the search page to help users find relevant products faster.

  • A product detail page (PDP): Recommendations appear on the page where users view detailed information about a product. These suggestions can include related items, frequently bought together products, or alternative options.

  • The Cart page: Recommendations that appear on the shopping cart page, often focusing on upselling or promoting complementary items to the ones already in the cart.

  • The Confirmation page: Recommendations presented during the checkout process, encouraging last-minute additions before the purchase is finalized.

  • Other pages: Recommendations that appear in other locations not explicitly defined in the above categories.

Product recommendation strategies

Merchandisers can use product recommendation strategies to define the logic behind the product recommendations displayed in a recommendation slot. Each strategy available in the Recommendations manager is designed to optimize the relevance of the recommendations for specific slot locations, they are the following:

  • Viewed together: This strategy recommends products that are frequently viewed together with the product(s) currently being viewed by a visitor. It’s optimal for product detail pages (PDPs), cart pages, and confirmation pages.

    This strategy is a seeded strategy, meaning that seed products must be passed to recommendation requests to the Commerce API to generate recommendations.

  • Purchased together: This strategy recommends products that are frequently purchased together with the product(s) currently being viewed by a visitor. It’s optimal for product detail pages (PDPs), cart pages, and confirmation pages.

    This strategy is a seeded strategy, meaning that seed products must be passed to recommendation requests to the Commerce API to generate recommendations.

  • Intent-aware: This strategy recommends products based on a visitor’s real-time actions and the product they are viewing. The intent-aware strategy is based on the Session-Based Product Recommendations (SBPR) ML model.

    This strategy is a seeded strategy, meaning that seed products must be passed to recommendation requests to the Commerce API to generate recommendations.

    Important

    Contact your Coveo representative to enable SBPR in your Coveo organization.

  • Most viewed: This strategy recommends the products that are the most frequently viewed by visitors of the storefront. It’s optimal for locations where you want to show generic recommendations, such as the home page.

    This strategy is popularity-based, meaning that recommendations are generated without the need for seed products.

  • Most purchased: This strategy recommends the products that are the most frequently purchased by visitors of the storefront. It’s optimal for locations where you want to show generic recommendations, such as the home page.

    This strategy is popularity-based, meaning that recommendations are generated without the need for seed products.

  • Recently viewed: This strategy recommends the products that have been recently viewed by visitors of the storefront. It’s optimal for locations where you want to show generic recommendations, such as the home page.

    This strategy is based on a visitor’s behavior, meaning that recommendations are generated without the need for seed products.

  • Recently purchased: This strategy recommends the latest products that a visitor has bought, analyzing the visitor’s recent purchase history to generate recommendations. It’s optimal for post-purchase areas of a site, such as order history pages or confirmation pages.

    This strategy is based on a visitor’s behavior, meaning that recommendations are generated without the need for seed products.

  • Purchased with recently purchased: This strategy recommends products that are most frequently purchased with the latest products a visitor has bought. It’s optimal for post-purchase recommendations, such as purchase confirmation pages.

    This strategy is based on a visitor’s behavior, meaning that recommendations are generated without the need for seed products.

  • Buy again: This strategy recommends products that a visitor has previously purchased and might want to buy again.

    This strategy is based on a visitor’s behavior, meaning that recommendations are generated without the need for seed products.

Refer to Supported strategies by seed type for more information.

Product-based strategies

Product-based strategies enable merchandisers to determine how seeded strategies influence the application of product recommendation models.

  • A seeded strategy uses one or more products as a reference point to generate product recommendations.

  • An unseeded strategy uses other references, such as recent events or product popularity, to generate product recommendations.

A merchandiser may enable a product-based strategy by selecting the product seed type in the slot creation process. While merchandisers don’t directly control the seed products, they can influence recommendations based on the product data provided.

The following recommendations strategies are available in the Recommendations manager based on the seeded or unseeded format:

Seeded strategies Unseeded strategies

Product seed types

Unseeded strategies don’t require a seed product to generate recommendations, instead they use other factors, such as a visitor’s history or popularity of products to recommend products, which is optimal for locations where you want to show generic recommendations, such as the home page. Seeded strategies rely on three seed product types, which influence the recommendation engine in a different way:

Note

When creating a recommendation slot, the slot location will impact which seed types are available.

  • Detail product: The recommendation engine uses the products currently being viewed by the user on a product detail page (PDP) to generate recommendations. This seed type is optimal for recommendation slots displayed on PDPs.

  • Cart product(s): The recommendation engine uses the product(s) currently in the user’s cart to generate recommendations. A visitor may have multiple products in their cart, which can be used as seeds to generate recommendations. This seed type is ideal for cart-related areas, such as cart pages or overlays, regardless of where they’re displayed on the site.

  • Purchased product(s): The recommendation engine uses the product(s) just purchased by the user to generate recommendations. A visitor may have purchased multiple products, which can be used as seeds to generate recommendations. This seed type is optimal for confirmation pages.

Important

Once a product-based strategy is selected during the recommendation slot creation process, you cannot change it. You’ll need to create a new slot to change these settings.

The system then leverages AI capabilities of Coveo to expand these seeds, identifying products with similar attributes, enhancing discoverability, and driving sales. Refer to Supported seed types by slot type for more information.

Additionally, merchandisers can apply Conditional filtering and ranking rules to further customize the recommendations based on the seed products.

Use case examples

The following examples illustrate how product recommendation strategies can be used to enhance the visitor experience and drive sales through both seeded and unseeded strategies.

Important

Creating a recommendation slot in the Coveo Merchandising Hub does not automatically display recommendations to the visitor. Depending on the seed type you choose, additional front-end development is required to configure how the recommendations are displayed.
Seeded strategy example

Scenario: On a Coveo powered ecommerce site, a visitor adds a Kayak to their shopping cart.

The merchandiser creates a slot The visitor sees the recommendations

  1. In the CMH Recommendations manager, a merchandiser creates a recommendation slot for the cart page.

  2. They enable the product-based strategies where the appropriate seed type for the slot location is pre-set to Cart product(s).

  3. They select the Purchased together strategy, which leverages the kayak in the visitor’s cart as the seed product and previews the recommended products.

  4. They set the display heading to You might also be interested in these to provide context for the recommendation slot.

  5. Optionally, the merchandiser might apply additional filter or ranking rules to fine-tune the recommendations by setting conditions based on product attributes.

  1. After having added a Kayak to their cart, the visitor reviews their cart before checkout and see a recommendation section beneath or alongside the cart summary.

  2. The visitor sees a list of Paddles, Life jackets, and Waterproof bags recommended based on the Purchased together strategy.

  3. The visitor can easily add these items to their cart by clicking on the Add to cart button next to each product.

  4. The process feels seamless and personalized, enhancing the shopping experience without requiring any additional searches.
Unseeded strategy example

Scenario: A visitor lands on the homepage of an e-commerce site and starts browsing for products.

The merchandiser creates a slot The visitor sees the recommendations

  1. In the CMH Recommendations manager, a merchandiser creates a recommendation slot for the home page.

    • By default, the product-based strategy isn’t applied, as the slot location is optimal for unseeded strategies.

  2. They select the Most purchased strategy, which displays the most purchased products, by giving more importance to recent events.

  3. They set the display heading to Hot items to provide context for the recommendation slot.

  4. Optionally, the merchandiser might apply additional filter or ranking rules to fine-tune the recommendations by setting conditions based on product attributes.

  1. A visitor navigates to the homepage of the Coveo powered ecommerce site.

  2. At the bottom of the page, they see a list of products trending under the Hot items section based on the Most purchased strategy.

  3. The visitor selects a item from the list and adds it to their cart.

Create a recommendation slot

Note

It’s possible to create recommendation slots using the Commerce API, which can be a time-saving option for users managing multiple tracking IDs or storefronts.

To create a recommendation slot:

  1. In the CMH, access the Recommendations manager.

  2. At the top right, from the Tracking ID dropdown, select the tracking ID for which you want to create a slot.

    • The slot you create will apply to all locales associated with the selected tracking ID.

  3. At the top right of the Recommendations manager page, click Create a slot.

  4. Enter the Name that will identify the slot in the Recommendations manager.

  5. From the Slot location dropdown, select the recommendation slot location where the slot will be displayed.

    Note

    When creating a recommendation slot, the slot location will impact which seed types are available.

  6. Click Next: Strategy settings.

  7. In the Strategy settings section, from the dropdown, choose whether you want to apply a product-based strategy.

    • The dropdown will display the product seed type based on the slot location you’ve selected in the previous step.

  8. Select the appropriate product recommendation strategy for the slot.

    • The preview will display a sample of the products that will be recommended based on the selected strategy (not available for behavior-based strategies). For seeded strategies, to preview the effect the strategy will have on the recommendations, you must enter products in the Find a product field.

  9. Click Next: Display settings.

  10. In the Display settings section, enter a display heading for the recommendation slot.

    • This heading will be displayed above the slot in the storefront.

  11. In the Display product thresholds section, enter the number of products you want to display in the slot, followed by the minimum.

  12. Click Create slot.

Edit a product recommendation strategy

Product recommendation strategies are used to define the logic behind the product recommendations displayed in a recommendation slot. Each strategy is designed to optimize the relevance of the recommendations based on specific user behaviors, product attributes, or business objectives.

To edit a product recommendation strategy:

  1. In the CMH, access the Recommendations manager.

  2. Select the slot for which you want to edit the strategy.

  3. From the navigation tabs, select Strategy & rules.

  4. Next to the applied strategy, click Modify.

  5. In the Edit strategy panel, choose one of the available strategies.

    • The preview displays a sample of the products that will be recommended based on the selected strategy (not available for behavior-based strategies). For seeded strategies, to preview the effect the strategy will have on the recommendations, you must enter products in the Find a product field.

  6. Click Next: Display header.

  7. Enter a new heading that aligns with the revised strategy in the Display heading field, and click Publish now.

Recommendation slot configuration

In this section of the CMH, merchandisers can manage slots by viewing their details, renaming them, retrieving the slot ID, editing display product thresholds, or deleting the slot.

To access the configuration page:

  1. In the CMH, access the Recommendation manager.

  2. Select View at the end of any recommendation slot listed.

  3. From the navigation tabs, select Configuration.