Recommendations - Exclude articles that are already v...

[Kim_VRT]

We have just started with the recommendation functionality in Target. We want to use the 'Most popular' algorithm on the page, but want to exclude the articles that are already viewed by the visitor. How can we set this up in Target for recommendations? With the exclusion functionality?

[Eric_Vidana · Accepted solution]

Adobe Target handles this out of the box as long as you have entity.id parameter for the Target call set up on the product pages. There is also a checkbox in the criteria creation screen to allow/disallow recommendations to be previously purchased (I think). In order for purchase to populate for the user profile, an order conversion will have to take place.

https://experienceleague.adobe.com/docs/dtm/implementing/target/configure-target/mboxes/order-confir...

[Kim_VRT]

Thanks for your feedback.

We have an entity.id parameter, so this means that no single configuration is needed in the recommendations? Already viewed pages are automatically filtered out? Is this real time?

[Eric_Vidana]

The entity.id param will assist with recently viewed algorithms. If you want to suppress a view, then you need to treat a view as a purchase. The reader needs to 'buy' the article if that makes sense. You can create an orderConfirmation Target request to achieve this. https://experienceleague.adobe.com/docs/target/using/implement-target/client-side/mbox-implement/ord... productPurchasedId

[Kim_VRT]

We have created an orderConfirmation Target request / mbox as indicated in the documentation you provided. We expected that the 'viewed' articles (purchased) are now automatically excluded from the most viewed algorithm. However, we still see them in the results. In the criteria creation, we do not see any checkbox that indicate that you can exclude 'purchased' articles. Do we miss something? Many thanks for your feedback!

[Eric_Vidana]

I think it would be good to look at the debug params that come along with Target. Go into Admin / Implementation and generate a new debug token and include that with your request. You should see a different sort of response that includes mboxTrace. It should output the profile of the user in question. What you want to look for is user.endpoint.lastPurchasedEntity. Let me know if that yields any insight..

[Kim_VRT]

Dear Eric, we see in the output of the mboxTrace that the "user.endpoint.lastPurchasedEntity" has indeed all the ID's of the already viewed "articles". So that's already good news.

In the recommendations on the website however, only the "article" that is currently viewed is excluded, not all the already viewed "articles". We cannot add the "user.endpoint.lastPurchasedEntity" in the script (we receive an error) in orde to be used in the "exclusions" for example. Do you have a solution for that?

Thanks!

Kim

[Eric_Vidana]

It sounds like you are experiencing a bug then. I am taking text from this page: https://experienceleague.adobe.com/docs/target/using/recommendations/criteria/create-new-algorithm.h... (Optional) Slide the Recommend Previously Purchased Items toggle to the “on” position. This setting is based on the productPurchasedId. The default behavior is to not recommend previously purchased items. In most cases you do not want to promote items a customer has recently purchased. It is useful if you sell items that people typically purchase only once, such as kayaks. If you sell items that people come back to purchase again on a repeated basis, such as shampoo or other personal items, you should enable this option. It might be worth opening a client care ticket and referencing this forum post.

[Kim_VRT]

Hi, we have finally found the issue. As we are media company, we have set the Industry vertical in the Criteria to Media/Publishing which is linked to set of Page Types. There you do not have the "Recommend Previously Purchased Items toggle", which is off course logic from this point of view. It would be good if there was a "Recommend Previously Viewed Items" toggle for Media/Publishing companies. Thanks for your help!