Release Summary 25.07 | Apr 03, 2025
The following key features and improvements, along with bug fixes, have been released in Algonomy CXP products in the release version 25.07.
Find
Search Enhancements - Improved Product Discovery with User Engagement Ranking
We’ve introduced User Engagement Ranking (UER) to improve how products are surfaced in search results by factoring in real user interactions. The UER score is now used both at indexing and query time to deliver more relevant and engaging results.
Benefits:
-
More Relevant Results: Products with higher levels of engagement—purchases, views, and clicks—are ranked higher.
-
Improved Product Discovery: The system takes into account various factors, such as purchases, views, and clicks, to facilitate enhanced product discovery.
How It Works: The system automatically calculates a popularity score for each product in the background. This score is also applied in real-time during search queries using Solr’s additive boost functionality, enhancing relevance without impacting performance. Configuration is handled through the additiveBoost setting.
Recommend
New Layout Variables for Cleaner URL Access
Two new layout variables — PRODUCT_URL and CLICK_TRACKING_URL — have been added to the layout editor to simplify access to product and tracking URLs. These help developers implement SEO-friendly links without relying on parsing the clickURL, which may change over time.
-
PRODUCT_URL: Direct link to the product page (same as productURL in recsForPlacements).
-
CLICK_TRACKING_URL: Click tracking link without the ct parameter (same as clickTrackingURL in recsForPlacements).
These variables are now available in the Item Variables section and are included in p13n_generated.js.
Jira: ENG-29908
Enterprise Dashboard
Auto-Suggest Enabled for Attribute Names in DNR Rules – User Tab
The "Do Not Recommend" (DNR) rules under the User tab now support auto-suggest for attribute names when configuring rules that filter products based on viewed or purchased attributes.
While autosuggestions help quickly find existing attributes, users can also manually enter custom attribute names if needed.
Jira: ENG-29973
Autocomplete Now Enabled for Product IDs in Configurable Strategies Preview
The Configurable Strategies preview page now supports autocomplete suggestions when entering product IDs as personalization seeds. This update brings it in line with the functionality available in Recs Test Drive.
Jira: ENG-29887
Fixes to Product Comparison DynEx Template
The Product Comparison DynEx template now includes key styling and alignment improvements. Attribute names are bolded, product images and names are consistently sized and aligned, and scrolling keeps the top section and first column fixed. Additionally, the strategy message is used as the header, and the seed product column is visually distinguished for better clarity.
Jira: ENG-29911
Other Feature Enhancements
The following feature enhancements and upgrades have been made in the release version 25.07.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Social Proof: Social Proof MVT Reporting: Filter by Products with Messaging |
The MVT report for Social Proof tests on Item Pages now includes a filter to compare control and treatment visits only for products that displayed a social proof message. This allows a more accurate measurement of social proof impact. The filter is available via a new parameter in the MVT API and includes statistical confidence and lift metrics. |
03-Apr-25 |
|
|
Updated Column Header in Strategy Rules – Preferred Strategies Table |
To reduce confusion, the column header previously labeled “Message” in the preferred strategies section of Strategy Rules has been renamed to “Warnings.” This better reflects the intent of the column, which is to display any issues or alerts related to the strategy, not the strategy message itself. |
03-Apr-25 |
|
|
Find: Optimized Autocomplete Calls with findCallType=overlay |
A new query parameter findCallType=overlay has been introduced in the Find search API to support autocomplete use cases. This disables resource-intensive features like Hybrid Search and Query Understanding, reducing overhead from frequent character-level search calls. |
03-Apr-25 |
|
|
Ensemble AI: Sampling for Very Large Styles |
To handle styles with extremely large combinations, sampling has been introduced as a fallback when sub-styles are insufficient. When the total combinations exceed the system's processing capacity, the system samples a subset of products from each part, ensuring all seed products are still considered. Sampling ratios vary by scale (e.g., 1/5 for up to 500B, 1/15 for up to 100T, etc.). This approach improves processing efficiency while still delivering diverse ensemble recommendations. |
03-Apr-25 |
|
|
Improved Kong Container Startup Process |
The Kong container startup has been streamlined by bundling necessary files directly within the container. Configuration is now passed via environment variables, eliminating dependencies on the underlying host and enhancing portability and simplicity. |
03-Apr-25 |
|
|
Omnichannel Feed Process Migrated to Airflow |
The existing omnichannel feed process has been successfully converted and is now running as an Airflow DAG, improving automation, monitoring, and maintainability. |
03-Apr-25 |
Bug and Support Fixes
The following issues have been fixed in the release version 25.06.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Data Science Workbench: Fix for DSW Schedule Re-Creation Issue |
Resolved an issue where re-creating a schedule immediately after deletion (without refreshing the page) resulted in an error. Users can now delete and re-create a strategy schedule seamlessly without needing to refresh. |
03-Apr-25 |
|
|
Social Proof: API Behavior Improvements |
Addressed multiple issues related to Social Proof behavior. The "Track Experience" API is now correctly triggered, and the alignment of icons with text has been fixed. |
03-Apr-25 |
|
|
Find: Find Data Publisher: Exponential Backoff for 5XX Errors |
Fixed an issue where exponential backoff was not triggered for 5XX server errors in the Find Data Publisher. The retry mechanism now correctly handles transient failures, improving job resilience. |
03-Apr-25 |
|
|
Find: Find Data Publisher: Failed Metric Now Published for Incorrect Configurations |
Addressed an issue where the failure metric (find.datapublisher.failed) was not being published when the job failed due to incorrect database or Hadoop configurations. The metric is now correctly emitted to Datadog, improving visibility into configuration-related failures.
|
03-Apr-25 |
|
|
Streaming Catalog: Deleted Categories No Longer Appear in Streaming View API |
The Streaming View API now correctly omits deleted categories from the root section of the hierarchy. This fix ensures that removed categories no longer appear in responses, and all streaming functionalities continue to work as expected. |
03-Apr-25 |
|
|
Enterprise Dashboard : Restriction Rule Display Issue in Recs Test Drive Fixed |
The issue where restriction rules weren’t visible in the Recs Test Drive has been resolved. The applied rules now correctly match between the Experience Browser and Recs Test Drive. |
03-Apr-25 |
|
|
Enterprise Dashboard : Configurable Strategies: Lookback Setting Bug Resolved |
The issue with setting the lookback window for user history in Configurable Strategies has been fixed. Users can now configure lookback ranges beyond one day without encountering errors. |
03-Apr-25 |
|
|
Spellcheck Disabled for Product ID Searches |
Product ID searches were incorrectly triggering spellcheck, causing mismatched results. This has been fixed—spellcheck is no longer applied to product ID queries, ensuring accurate results. |
03-Apr-25 |