All Products
Search

This Product

    Quick Audience:Data Import Overview

    Document Center

    Quick Audience:Data Import Overview

    Last Updated:May 26, 2025

    After Quick Audience is connected to a data source, you can store the data table to be imported in a computing engine. Then, you can declare the table structure of the data table in Quick Audience and run scheduling tasks to import data from the table to Quick Audience for subsequent analysis and application.

    Data import process

    1. ID Type Management. Set the user ID types supported by Quick Audience and the ID mapping policy.

    2. User Attribute Management and build a user attribute information system.

    3. Configuration table structure:

      Note

      For more information about the format requirements of the following tables, see Data Table Requirements.

      • User tag table: The user tag table records a series of characteristic attributes of users and is used for Perspective Analysis and Audience Filtering.

        When you configure the schema of a user tag table, you must specify the tags that carry user attributes.

      • Statistical Table: The statistical table records the statistical data of user operations during the statistical period. You can perform Audience Filtering based on the data in the statistics table.

      • User Behavior Table: The user behavior table records the browsing, purchase, collection, and purchase behaviors of users. Based on user behavior data, you can perform Audience Filtering. You can also generate custom tags such as AIPL Model and Preference Tags. AIPL models can be used for AIPL User Analysis, AIPL Flow Analysis, and Audience Filtering.

      • Order Details Table: The order details table records order information at the sub-order granularity. Based on user behavior data, you can perform Audience Filtering and generate custom tags such as RFM Model and Preference Tags. RFM models can be used for RFM Analysis and Audience Filtering.

      • Order Summary Table: The order summary table records the order data of the last N days. The imported data can be used to generate an RFM Model. The RFM model can be used for RFM Analysis and Audience Filtering.

      Note

      You can also import all the preceding data tables through API operations. You can obtain the API description from the user.

      The following figure shows the links for generating AIPL/RFM models, custom tags, and groups from various tables.

      V4表&模型&人群&自定义标签

    4. Execute the scheduling task:

      You can use Scheduling Task to manually schedule or periodically schedule data for the configured table structure to import data and integrate computing data.

      An import task must be scheduled and executed at least once. Otherwise, no data is available on the Quick Audience side. If the data on the source side is updated, schedule the import task again to use the latest data in subsequent analysis.

      During import, the system identifies users based on IDs (identifies the same ID as the same user and performs deduplication), and assigns a QAID to the deduplication user. This calculation process is called ID Mapping. For more computational logic, see ID mapping and User Unique Identifier QAID.

      Note

      Each time a task is scheduled, the ID of the users involved in all data tables must be re-calculated. To avoid frequent consumption of computing resources, we recommend that you create the same task for scheduling all data tables.

      In subsequent operations such as user analysis and crowd filtering, the QAID will be used as the unique identifier of the user. Regardless of the table from which users or groups come, all data such as ID types, tags, and behaviors are queried based on their QAIDs to integrate cross-channel data and build an omni-channel tagging system. image

      FAQ

      What are the precautions when importing data tables?

      When you configure the data table structure and scheduling tasks, take note of the following:

      • The names of data tables that are stored in MaxCompute are all in English. You can specify a data table alias to identify the imported table. You can add the data source, brand, table type, and purpose to the alias of the data table.

      • When configuring the table structure, select the type and encryption status of the ID field. You can select MD5, SHA256, AES, or Unencrypted.

        If the ID field is encrypted by AES, you must enter the key in Organization System Configuration. After the ciphertext is imported, the system decrypts the ciphertext and uses the original text.

        In subsequent scenarios such as push and marketing, some channels only support specific ID types and encryption types.

      • Each time a task is scheduled, the ID of the users involved in all data tables must be re-calculated. To avoid frequent consumption of computing resources, we recommend that you create the same task for scheduling all data tables.

        You cannot configure multiple data table structures for the same data table. You cannot add the same data table to multiple scheduling tasks.

      • If Executed is displayed for a scheduling task, all tables included in the task are imported. If Executed Failed is displayed, move the pointer over Executed Failed to view the error message of each table. Then, correct the error and click the 432icon to manually try again.