---
title: "Importing Recommendations for Personalization"
slug: "creating-recommended-products-using-stored-attributes"
updated: 2026-01-26T16:20:15Z
published: 2026-01-26T16:20:15Z
canonical: "docs.mapp.com/creating-recommended-products-using-stored-attributes"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mapp.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Importing Recommendations for Personalization

## Goal

Upload recommended product records to Mapp Engage so you can use them in personalization blocks and Segmentation Builder.

---

## Prerequisites

To complete this task, you need:

- Mapp Engage with Datastore enabled.
- A recommendation file that is correctly formatted and ready to upload.

---

## Procedure

1. Select the recommendations you want to upload.
2. Create your recommendation file in the required format.

Use the formatting details in [/v1/docs/creating-recommended-products-using-stored-attributes#the-related-data-format](/v1/docs/creating-recommended-products-using-stored-attributes#the-related-data-format) and [/v1/docs/creating-recommended-products-using-stored-attributes#import-file-format](/v1/docs/creating-recommended-products-using-stored-attributes#import-file-format) to ensure the file is valid.
3. In Mapp Engage, go to *Administration > E-commerce > Recommended Products*.
4. Enter the Engage Contact ID to find products for a particular user.
5. Upload the data as a recommendation file, just as you do for your Product Catalog.
6. Use the uploaded recommendations as personalization blocks in your email send-out or in Segmentation Builder.

---

## Related Data Format

- The dataset name follows this pattern: `mappdefaultrecommendedproduct`.
- In the dataset metadata, set the “restricted” field to “true”.

| **Field** | Description |
| --- | --- |
| `userId` | The internal Mapp Engage user (or contact) identifier, the same as for the rest of the restricted datasets. |
| `productSKU` | The product identifier should match the one in the Product Catalog and Transactions dataset. |
| `accuracy` | The prediction accuracy. Expressed as integers in the range [0, 1000]. Lower values mean lower accuracy. Example: if the accuracy is 0.5843521 (range (0; 1.0]), store it as 584. If not provided, the default value 1000 should be stored (or left empty and interpreted as the default). |
| `model` | The identifier of the model. This is an integer without special storage requirements. If not provided, the default value 0 should be stored (or left empty and interpreted as the default). |

---

## Import File Format

Create a CSV file with:

- A header row with column names
- One row per record
- A comma (`,`) as the separator

| **Field** | Description |
| --- | --- |
| key | The contact’s email or the Mapp Engage `userId`. If the value is a number, the import assumes it is the `userId`. Otherwise, it is treated as an email. During import, emails are translated into the `userId` (the same as for the rest of the restricted datasets). |
| productSKU | The product identifier should match the one in the Product Catalog and Transactions dataset. |
| accuracy | The prediction accuracy. Expressed as integers in the range [0, 1000]. Lower values mean lower accuracy. Example: if the accuracy is 0.5843521 (range (0; 1.0]), store it as 584. If not provided, the default value 1000 will be stored. |
| model | The identifier of the model. This is an integer without special storage requirements. If not provided, the default value 0 will be stored. |

---

## Example of a personalization block

Your recommendation block is called `ecx:recommendedProducts`. It must include:

- a source (`PRECALC`)
- User ID
- Model ID
- a minimum accuracy between 0 and 1,000

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

## Related

- [Recommended Product](/recommended-product.md)
- [eCommerce: Recommended Products](/ecommerce-recommended-products.md)