Overview
The pinned item management feature allows you to pin specified items to the top of a recommendation page. This ensures that these items can be first exposed each time a user accesses the recommendation page. The pinned item management feature allows you to intervene in recommendation results based on personalized recommendations. The feature is often used to pin news about state leaders, exclusive news, items for increasing traffic, and featured content to the top of a page.
Instructions
Configure pinned items by using the console
The pinned item management feature supports only the recommendation service type of You May Also Like. By default, the feature is disabled when you create a scene and select You May Also Like for this scene. You need to select Single Item Operation from the Operations Assistant drop-down list, click the Pinned Item Management tab, select a scene, and then click Edit to enable the pinned item management feature in the scene. Then, you can configure pinned items.
Add a pinned item
You can add a pinned item by clicking + Add Pinned Item, selecting the item that you want to pin to the top, and then clicking OK.
You can filter items by entering a title or the combination of an item ID and an item type in the search box.
A maximum of 10 items are supported in the list of pinned items.
Unpin an item
You can click Unpin in the Actions column to unpin an item.
Adjust the order of pinned items
You can click Up or Down in the Actions column to adjust the order of pinned items. The pinned items in the recommendation results are arranged based on the order.
Publish configurations
After you click Publish, the pinned item management feature of the scene is in the publishing state. You cannot modify the feature when the feature is being published. After the pinned item management feature is published, the recommendation results are returned based on the latest list of pinned items.
If you pin an item to the top, you cannot recommend the item in a fixed position.
Obtain recommendation results by using server SDKs
When a user accesses a recommendation page, the system sends a request to Artificial Intelligence Recommendation (AIRec) to obtain recommendation results. If the request is the first request for this access, the value of the strategy field in request parameters needs to be set to fixedSlot. This way, the system returns pinned items. You do not need to specify the strategy field in subsequent requests. The system returns items according to the algorithm-based recommendation.
For more information about how to obtain recommendation results, see Obtain recommendation results.
1. In the first request, the value of the returnCount field must be greater than or equal to the number of pinned items. Otherwise, only some pinned items can be returned. The pinned items cannot be displayed in full on clients.
2. If the pinned items are invalid because they are unpublished or expired, the system automatically adds items following the pinned items according to the algorithm-based recommendation results.