Personalised recommendations on a customer basis, not product based

[lisa_w]

Hello ,

the recommendations in the Marketing Cloud are product-related. Is it possible to create recommendations at customer level, i.e. customer-related, in order to expand personalisation? The personalisation should only be displayed for customers who are logged in to the shop and are identifiable.

For example:

- Top products from a product category from a special customer

- Top products based on the purchases of a specific customer

And if so, how does one proceed?

Thanks for help.

Best regards.

[KunalBansal]

Dear lwe,

Please note that the recommendation model is available for Products, Offers, Item-of-Interests only.

• You cannot use what are you expecting in-terms like a special-customer/ specific-customer in Recommendation directly as available for Products.
• That type of scenario you need to handle in Segmentation based on customer Interaction.

I hope this clarifies. Thank you.

Best Regards,

Kunal Bansal

Senior Marketing Consultant

[lisa_w]

Thanks for Your Answers.

[deanmg1]

Hi lwe

We use Offer Recommendations to provide customers with list of available offers based on interactions in SMC

There are a few master data and transactional data dependencies here....

1. Offers that have content and are in released status
2. Interactions (eg. Sales Order) that link to the customer ID
3. Offer Recommendation Model (we have defined custom HANA views for our requirements but there are standard algorithms to be used, as per Kunals screenshots)

Those are the core elements you need. Within the model you can link your interaction type as well as the period to consider for those interactions. As an example we use Sales Order for the last 90 days

The following documentation should get you going:

https://help.sap.com/viewer/b88f770e4b7c4ecead5477e7a6c7b8f7/2105.500/en-US/9eb788ded90a46d2a73b935a...

I see you are wanting to surface these to your webshop - depending on the Commerce solution there are very short time outs for this data (7s generally). Reason being is that the loading of this on the page for a customer any longer would lead to a bad shopping experience. For this you need to consider the type of algorithm used.... runtime more than liekly wont work depending on the data set. We use generation algorithms - this means we have the models generating early in the morning to prep the data

Alignment of data is key in this process

Enjoy