Recommended Product

Requirements

• Mapp Engage, where Datastore is used

• Create a recommendation file in the correct format and upload it

Set up

1. Select your recommendations

2. It is very important that you are sure the data is properly formatted. Please see the formatting guide below to help you. 

   The Related Data format

   The name of the dataset will follow the established pattern: "mappdefaultrecommendedproduct".

   The "restricted" filed in the dataset's meta information should be set to "true".

   Field	Mandatory	Type	Description
   userId	yes	string	The internal Mapp Engage user (or contact) identifier. The same as for the rest of the restricted datasets.
   productSKU	yes	string	The product identifier. It should correspond to the one from the Product Catalog and Transactions dataset.
   accuracy	no	number	The prediction accuracy. Expressed in the format of Integers from the range of (0; 1000]. The lower the value, the lower the accuracy. For example, if the recommendation model produces accuracy in the range (0; 1.0], the accuracy 0.5843521 will be stored as an integer value of 584. If not provided, the default value of 1000 should be stored (physically, it could be an empty value that will be interpreted as the default).
   model	no	number	The Identifier of the model. This is an Integer number without special requirements from the storage perspective. If not provided, the default value of 0 should be stored (physically, it could be an empty value that will be interpreted as the default).

   Import File Format

   To import the recommended product records a CSV file should be created. It should have a standard structure of header with column names and the number of rows, each representing one record. The separator is a comma (",").

   Field	Mandatory	Type	Description
   key	yes	string	In this field, the contact's email or the Mapp Engage userId should be provided. If the value is a number, the import module assumes this is the userId, otherwise - an email. During the import process, emails will be translated to the userId (the same as for the rest of the restricted datasets).
   productSKU	yes	string	The product identifier. It should correspond to the one from the Product Catalog and Transactions dataset.
   accuracy	no	number	The prediction accuracy. Expressed in the format of Integers from the range of (0; 1000]. The lower value the lower accuracy. For example, if the recommendation model produces the accuracy in the range (0; 1.0] the accuracy 0.5843521 will be stored as an integer value of 584. If not provided, the default value of 1000 will be stored.
   model	no	number	The Identifier of the model. This is an Integer number without special requirements from the storage perspective. If not provided, the default value of 0 will be stored.

3. In Mapp Engage, go to Administration> E-commerce > Recommended Products. To find products for a particular user, enter the Engage Contact ID.
    

4. Upload the data as a recommendation file, similar to how you would to your Product Catalog.

5. They can then be used as personalization blocks in your email send-out or with Segmentation Builder.

Example of a Personalization Block

Your recommendation block will be called ecx:recommendedProducts and needs to include a source (PRECALC), your UserID and Model ID, and a minimum accuracy between 0 and 1,000. 

<%ForEach var="recommendedProduct" items="${ecx:recommendedProducts('PRECALC', user.pk, '3', 500)}"%>
    <%${recommendedProduct.productName}%> - <%${recommendedProduct.productPrice}%>
<%/ForEach%>

Firstly, you must meet a number of preconditions:

• You will need to activate the Mapp Intelligence Recommendation API and Marketing Automation. If this is not already activated, please speak to your customer success manager or account manager.

• For Mapp Intelligence solutions, your Product Catalog must be uploaded, and the same on Mapp Intelligence and Mapp Engage

• User matching (with the email redirect) is configured and in use

• Intelligence tracking (with product tracking) is active

• You want to send out messages individually

Set up and Use

1. You will need a recommendations API. See how to set that up here. 

2. Make sure you have the Placement ID for your recommendations. This can be found by going to Marketing > Placements > Reco API, finding the correct placement, and clicking on the API Info action

3. You will need to use a personalization block in Message Creation or Segmentation Builder. The source will be MAPPINT, and for your modelID, you will need the PlacementID from your Mapp Intelligence Recommendation API.

Example of a Personalization Block:

<%ForEach var="recommendedProduct" items="${ecx:recommendedProducts('MAPPINT', user.pk, 'plcmnt5', 0)}"%>
    <%${recommendedProduct.productName}%> - <%${recommendedProduct.productPrice}%>
<%/ForEach%>