# Background Pinned recommendation allows you to manually fix specific items at the top of the recommendation feed to meet strong exposure requirements for promotional campaigns or commercial promotions. By focusing traffic on core items during specific periods and scenarios, this feature effectively helps increase conversion rates. This document explains how to create and manage pinned recommendations. # Feature Introduction Pinned recommendation is an operational intervention feature that allows operation team to manually select one or more items and display them fixed at the top of the recommendation feed. This feature bypasses the standard personalized recommendation algorithm by defining specific items and order, ensuring that specified items receive priority exposure under specific conditions to meet commercial promotion or campaign operation needs. ## Key Features * **Manual Pinning**: allows you to fix one or more items at the top of the recommendation feed. * **Priority Management**: when multiple pinned items exist, you can set their order to decide the display sequence. ## Usage Restrictions * Pinned items occupy regular recommendation slots and take priority over all personalized recommendation algorithms. * The platform automatically ignores pinned items that have been deleted. ## Additional Notes * If the recommendation engine returns **zero** results (empty list), the platform **will not** return the pinned items either. * If the number of total items returned by the recommendation engine is less than the number of items you have pinned, the platform will only display a portion of the pinned items to match the returned count. # Configuration Steps 1. In the left navigation of the recommendation experience, under **“Recommendation Strategy” \- “Item Sorting” \- “Pinned Items”** , click **“Add Items”**

In the pop\-up dialog, find the target item by entering the item ID or keyword in the search bar at the top, then check one or more items you want to pin from the result list. The platform automatically searches by the following methods: item ID, exact keyword search, and fuzzy keyword search.

2. Click **“Confirm”** , and the recommended items will be added to the **“Pinned Items”** list. 3. In the **“Pinned Items”** list, you can reorder items by dragging and dropping, or remove them by clicking the delete icon. You can also toggle the switch to the left of “Return pinned items in each request” to enable or disable this feature. For details, see: [Repeat recommendation on each refresh](/docs/85296/1863361#c05becd3).

4. Before applying the configuration online, it is recommended to run a test to ensure the pinning works as intended and avoid unexpected impacts on real users. In the experience area on the right, click the “Recommend again” button to simulate multiple refresh requests by the same user and check the pinning effect.

5. If the experience meets expectations, click “Sync Changes to API” to activate the settings. ## Repeat recommendation on every refresh When you enable “Return pinned items in each request”, the product returns pinned recommended items on every request. In certain business scenarios (such as waterfall streams), each request returns top pinned recommended items that conflict with the APP interaction, causing the pinned recommended items to be exposed multiple times within a single user session. You can avoid this issue by disabling "Return pinned items in each request". |**Function status** |**Effect** | |---|---| |Enable |**Return pinned recommendations** for each request | |Disable |For each request, the platform determines whether to return pinned items based on the provided `session_id` parameter:|\ | ||\ | |* If the `session_id` has **already returned** pinned items within 24 hours, the platform automatically ignores the pinned recommendation item settings.|\ | |* If the `session_id` has **not returned** pinned items within 24 hours, the pinned recommendation item settings take effect and pinned recommendation items are returned. | :::tip \*Description If you disable "Return pinned items in each request", you must include the `session_id` parameter in your request. For more details, see [Request Body](/docs/393/92213#4fa39169). ::: # Frequently Asked Questions **Q: What should I do if I get the error "Item ID does not exist" when creating a pinned recommendation?** A: Please verify whether the item ID you entered is correct. **Q: I cannot find my item?** A: Please check the keywords you entered or try changing to new keywords. The platform can only use item dataset fields that are configured as "filterable" and "Searchable (text)".

Viking AI Search