2026 Releases
Release 26.7.0
July 6, 2026
A/B Testing Experiment Limit
The number of A/B testing experiments that can run at the same time is now capped, with a default limit of 30. Once the limit is reached, creating a new experiment is rejected until a running experiment is ended or deleted. Experiments that have already finished do not count toward the limit.
Order Import Warnings
Two new warnings are now surfaced after order imports — on both dry runs and real imports — to help diagnose data quality issues before they affect user profiles and downstream attribution.
Heavy users: When a single user ID accounts for an unusually large share of orders in a file (≥2% of total orders), the import result lists those users as a warning. A disproportionate concentration of orders on one user often signals a mismatch in the user-identifier column — for example, a shared or placeholder value being used instead of a unique per-user identifier.
Repeated order IDs: When the same order ID appears in non-consecutive rows, the later rows are treated as an update to the already-closed order rather than part of the original order. This matters for attribution: only the first group of line items for an order is considered for attribution forwarding. When this affects ≥2% of the import's orders, a warning is shown with example order IDs.
See the Orders import documentation for details.
All Changes
Click here for the complete list of changes
Added
- Add a limit on the number of A/B testing experiments that can run concurrently.
- Show a warning when a user is assigned an unusually large share of orders in an import.
- Show a warning when the same order ID appears more than once in an import.
Release 26.6.1
June 22, 2026
Segment Imports
Audience now supports importing user segments directly from a file. Teams that maintain their audiences in a CDP, data warehouse, or other internal system can upload the resulting membership as a CSV and activate it in the Kevel Ad Server, without rebuilding the segmentation logic in Audience. Each file holds the complete membership of a segment, and every import fully replaces it, so imported segments stay in sync with their source — uploaded on demand, on a schedule, or whenever a new file is detected.
To set it up, create a new import and select the External segments type.
For full details, see the Segment imports documentation.
All Changes
Click here for the complete list of changes
Added
- Support importing user segments from files.
Release 26.5.1
May 25, 2026
Native Static Segments
Static segments are now out of beta, and can also be used in the Kevel Native Segments destination. Marking a segment as static means it will be sent directly to Forecast and immediately made available for predictions upon activation to UserDB.
Refer to the Kevel destinations documentation for further details.
All Changes
Click here for the complete list of changes
Added
- Stabilize static segments and make them available in the Kevel Native Segments destination.
Release 26.3.2
March 30, 2026
File Detection Imports
When users upload files to their own S3 bucket or SFTP server, imports no longer need to wait for the next scheduled run. Audience can now detect when a new file is available and trigger the import immediately, reducing the delay between a file being uploaded and its data being available in the system.
To set it up, create a new import, select the type "Automated ingestion" and set the frequency to "When a new file is available".
For full details, see the Import Guide.
All Changes
Click here for the complete list of changes
Added
- Support file detection imports.
Release 26.1.1
January 19, 2026
HTTP to HTTPS Redirection
To improve security and ensure data integrity, all HTTP interfaces are now automatically redirected to their safer counterpart HTTPS. This applies to all API endpoints and web interfaces. Existing HTTP clients should follow the redirect to the secure endpoint.
All Changes
Click here for the complete list of changes
Changed
- Public HTTP interfaces are now redirected to HTTPS.