Skip to Content
0

Product Recommendation scenarios in content studio

May 24, 2017 at 11:16 AM

142

avatar image
Former Member

Hi,

I have maintained a new recommendation scenario (including all necessary steps arround, like model types etc). This recommendation scenario can already be used in hybris commerce without problems. Now I want to use this scenario as well in Content studio, but there only the standard recommendation scenarios will be shown and not mine.

Any hints what I could do?

BR

Stefanie

10 |10000 characters needed characters left characters exceeded
* Please Login or Register to Answer, Follow or Comment.

2 Answers

Saurabh Kabra
May 24, 2017 at 05:45 PM
0

Hi,

You need to make sure that "COMMUNICATION MEDIUM" is maintained as "EMAIL" in recommendation scenario.

Snapshot attached.

cusersi336320desktopprodreco.png

Thanks

Saurabh


Show 1 Share
10 |10000 characters needed characters left characters exceeded
Former Member

Thank you!

0
Matthias Reiner
May 31, 2017 at 06:56 AM
0

Additional Info:

https://help.sap.com/viewer/b88f770e4b7c4ecead5477e7a6c7b8f7/1705%20500/en-US/b6e0e555fb26727fe10000000a44538d.html

  • Note

    Recommendation scenarios that you intend to use in email campaigns to elicit recommendations must use an Optimized model type.

...

https://help.sap.com/viewer/b88f770e4b7c4ecead5477e7a6c7b8f7/1705%20500/en-US/92e0e555fb26727fe10000000a44538d.html

Optimized Algorithms

Optimized algorithms enable recommendations to be cached and used for similar consumers. As a result, future requests for the same recommendations can be provided much faster with less resource consumption. Recommendation model types can be configured to use optimized algorithms exclusively. All Association and Item Mapping algorithms, along with some Query algorithms, are optimized.

Show 4 Share
10 |10000 characters needed characters left characters exceeded
Former Member

Thank you!

0

Hi Matthias,

So if mail only support optimized algorithms... then e.g. collaborative filtering couldn't be used?

Which recommendation models are suitable to be used in email then? Only TOPN?

Kr,

Joyca

0
Former Member

Hi Joyca,

Once the model type is set to optimized then any model you create in Manage Recommendations or Recommendation Models apps with that model type you will see algorithms that are applicable for optimized models only. Any algorithms that don't apply to having an optimized model will be hidden. This should help you identify all applicable algorithms. This approach provides a better user experience since the user doesn't have to know which algorithms have to be used to have an optimized model rather let the system guide you.

Best regards,

Angelo

0

Hello Angelo,

OK Technically...

But Practically: in an email, I want personalized recommendations of course.

But when using association algorithms (which are optimized), like Often Bought Together Across Transactions (Interactions), you always need a Leading Item, right? In the Email Content Studio, where you set your Recommendation Scenario, you can add such a Leading item. But that is fixed? Then you will send the same recommendations to everyone in your email?

I tried sending an email with and without a Leading Item. But both mails come up empty... While the contacts that I send to do get results in their factsheet on the Commerce tab in the "Recommendations based on Buying History"... How does that work actually, those recommendations based on buying history, as it uses the buying history as leading items? And how do I use this like that in my emails?

Also the Collaborative Filtering algorithms are not optimized, while these just sound exactly like something you'd want to use in an email: if I understood correctly, it determines people similar to you, and recommends products that those similar people bought. Perfect for emails! But can't be used?

Thanks for some clarification,

Joyca

0