Preference Tags

更新时间:
复制 MD 格式

Preference tags derive values from imported or reported user behavior tables or order detail tables. The attribute with the most occurrences or the highest value becomes the tag value, such as a user's preferred shopping time period, price range, channel, theme, category, or highest single-item purchase price.

Example 1:

Create a tag named "Preferred category" from the top three categories each user prefers. The tag value is the specific category.

The following table shows how many times four users purchased items in five categories.

User

Category 1

Category 2

Category 3

Category 4

Category 5

User A

2

3

1

0

3

User B

1

2

2

0

1

User C

0

0

0

1

0

User D

0

0

0

0

0

The resulting "Preferred category" tag values are:

User

Preferred category

User A

Category 2, Category 5, Category 1

User B

Category 2, Category 3, Category 1

Note: Category 1 and Category 5 are both ranked 3rd by count and are not null. Category 1 is selected because it has the highest serial number.

User C

Category 4

User D

NULL

  • The "Preferred Category" tag covers 3 users (User A, User B, and User C), with a coverage rate of 75%.

  • For a single tag value, such as "Category 3", the number of covered users is 1 (User B), with a coverage rate of 25%.

Example 2:

Order placement times are divided into four periods: 0:00 to 6:00, 6:00 to 12:00, 12:00 to 18:00, and 18:00 to 24:00. The period in which a user places orders most frequently becomes the value of the "preferred shopping time" tag.

Create a preference tag

Preference tags are created from user behavior tables and order detail tables. For data requirements, see Data Table Requirements. For information about importing user behavior and order detail data, see User Behavior and Order Details. For information about reporting and storing order event data, see Event Hub.

Note

To create a time-based preference tag, the time field in the user behavior table and order detail table must have at least hourly granularity.

Steps:

  1. Choose Workspace> User Insight > User Tags > Tag Management. image

  2. Click New Tag in the upper-right corner.

  3. In the dialog box that appears, select Preference Tag and click OK. 222

  4. Configure the parameters in the Basic Information section.image

    1. Enter a label name. Click the 3人icon next to the tag field to customize the tag field name. The prefix is fixed to "user_tag_". If you do not customize the tag field name, the system will generate a field name based on the tag name.

    2. Set whether tags are manually updated or automatically updated on a periodic basis. The periodic basis supports daily, weekly, and monthly.

    3. Select a label category. The default category is Unclassified.

    4. Enter a label description.

  5. Configure tag rules. image

    1. Select User Behavior. If you create a tag based on user behavior, select a specific behavior type under the Interaction Behavior category. If you create a tag based on order details, select a specific behavior type under the Order Behavior category.

      If the selected behavior has multiple data sources, select a source table. If only one data source exists, no manual selection is required.

    2. The behavior time. You can use multiple methods to set the behavior time. For more information, see Time-based.

      When you create a tag based on order details, the table may contain multiple time fields. Before you set the behavior time, you must select the corresponding order time field.

      Note

      If the behavior time is set to a non-specific (relative) time, the time range is recalculated based on the update time when the tag is refreshed. If set to a specific time, the result remains unchanged unless the underlying behavior or order records change.

    3. You can also set additional filter conditions. For example, in the preceding figure, the data is filtered by mini program.

      Click Add Attribute to add a filter condition. You can specify up to 10 conditions. Multiple conditions are combined with AND logic.

    4. Set preference conditions. Preferences are divided into two types:

      • Feature: The attribute with the most occurrences or the highest value becomes the tag value. For example, the category a user purchases most often, or the highest price of a single item. You can select the top 1, 3, or 5 results.

      • Time-based preference: User behavior is grouped into time periods, and the period with the highest frequency becomes the tag value, as shown in the following figure. You can modify, add, or remove time periods. Time periods cannot overlap. 4

  6. After the configuration is complete, click Estimate Number of Users at the bottom of the page to preview the number of users and the coverage rate of each tag value. 5

  7. Click Save. The custom tag is created.

The new tag appears in the tag list in the Executing state. You can use it after its status changes to Executed. For more information, see Tag Management.