Issue

When you create an audience in Twilio Segment with the exact same name as an existing computed or custom trait, the audience can conflict with and overwrite the original trait mapping. This often results in the trait's data type unexpectedly changing (for example, from a number to a boolean value like false), causing the recorded trait volume to drop significantly. Deleting the audience after the conflict occurs does not automatically restore the original trait configuration or data values.

Product

Twilio Segment

Environment

Segment Console

Cause

The Segment audience builder currently permits naming configurations that match existing trait identifiers. When this naming conflict occurs, the system overwrites the original trait definition and remaps the data type. This configuration change and data loss persist even after the conflicting audience is deleted.

Resolution

To recover your original trait behavior and restore the historical data values, you must delete the audience with the duplicate name and re-import the correct trait data into Segment using one of the following methods:

Method 1: Ingestion of Identify Calls

You can repopulate the profile by creating identify calls with the correct trait information. 

Method 2: CSV Upload

If you have a backup or a local record of the correct user IDs and their corresponding trait values, you can manually upload them.

1. Prepare a CSV file containing a column for your user identifiers (such as user_id or anonymous_id) and a column for the specific trait_name containing its correct values.

2. Navigate to your Segment workspace and use the profile import features to upload the CSV file and update the user profiles.

Method 3: Reverse ETL

If the historical trait data is safely stored in your external data warehouse, you can sync it back to Segment.

1. Set up a Reverse ETL source in Segment connected to your data warehouse.

2. Configure a model to select the user identifiers and the correct historical values for the affected trait_name.

3. Map the warehouse column to the target trait in Segment and trigger a synchronization to overwrite the incorrect data.

Additional Information

To prevent this issue in the future, ensure that all new audiences use unique naming conventions that do not duplicate any existing custom or computed trait names.