Release Summary 25.05 | Mar 06, 2025
The following key features and improvements, along with bug fixes, have been released in Algonomy CXP products in the release version 25.05.
Ensemble AI
Accurate Product Listing in Ensemble Style Preview
The preview for Ensemble Style definitions now accurately lists seed products based on selected filters and criteria. Previously, products that did not match the style's filter settings were included in the preview. This update ensures that only the correct seed products are displayed, allowing merchandisers to review outfits effectively.
Jira: ENG-29708
Use User History to Specify Seed Products in Ensemble
Merchandisers can now preview ensembles based on a user's view or purchase history instead of selecting seeds only from a predefined style definition. This update allows email marketers to test and validate how outfits will be generated for specific users. Users can enter a test user ID and choose to generate outfits based on recently viewed or purchased items, with the option to specify up to three seed products. This functionality is available within both style definition previews and across multiple.
Jira: ENG-29672
Merchandiser Preview of Ensembles Unaffected by MVT
Merchandisers can now preview ensembles in the portal without being impacted by MVT tests. Previously, when an MVT test (On/Off) was active, some ensembles were not displayed due to API enable/disable settings. This update ensures that MVT configurations do not interfere with the merchandiser review process, and such requests are excluded from test scenarios and logging, preventing any impact on visits and reporting.
Jira: ENG-29746
Ensemble AI Reporting: Style Performance
A new Style Performance report is now available for Merchandisers to track the effectiveness of different styles based on shopper engagement. The report provides key performance metrics such as clicks, orders, attributable sales, conversion rates, and ATC counts. It includes graph visualizations (line and bar charts) for trends and comparisons, along with a detailed table view with filters for date range, channel, region, and currency.
Note: This report is enabled only when Ensemble AI is enabled.
Jira: ENG-29397
Enterprise Dashboard
Co-occurrence Report – Filter by Co-purchased Categories
Merchandisers can now refine co-occurrence reports by selecting specific co-purchased categories to analyze alongside a chosen category. Users can select one or multiple co-purchased categories at any hierarchy level, with the option to include all by default. An additional filter allows users to view only primary categories. These selections will apply to both graph and table visualizations, providing a more focused and relevant analysis of co-purchase trends.
Jira: ENG-29703
Recommend
Boost Rules Now Influence Product Ordering in Manual Recommendations
Manual Recommendations now support Boosting and Recommendation Restriction rules, giving merchandisers greater control over product display. Boost rules will now influence the ordering of products within Manual Recommendations, ensuring prioritized items appear at the top, including both manually selected products and backfill recommendations.
Jira: ENG-29516
Find
Conversion Rate in Find Search Terms Report
The Find Search Terms Report now includes a conversion rate metric to help search optimization managers analyze search term performance. Conversion rate is calculated as Total Visits with Search Conversions / Total Visits with the Search Term, providing insights into which searches drive user engagement.
Jira: ENG-29114
Other Feature Enhancements
The following feature enhancements and upgrades have been made in the release version 25.05.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Social Proof: Privacy Mode Support in Social Proof API |
The Social Proof API now respects the privacy mode (privm) parameter, ensuring that when privm=true, personalized messaging (e.g., "Since last visit") is excluded from the response. Other qualifying messages that meet the threshold will still be displayed. This enhancement applies to both client-side and server-side integrations, allowing greater control over user privacy preferences. |
06-Mar-25 |
|
|
Social Proof: Migration of Social Proof Templates to Folder Structure |
Social Proof messages and badges will now load default templates from static HTML, CSS, JS, and JSON files instead of relying on change-scripts and the Templates API. Custom templates saved with an experience or variation remain unchanged. |
06-Mar-25 |
|
|
Social Proof: Social Proof Optimization: Model Data Available in FDC |
The Social Proof Optimization model file, generated from training data, is now moved from the backend to front-end data centers (FDCs). This ensures the prediction API can access the model data efficiently. The model files are stored in a structured format under designated directories, enabling seamless integration with the API while maintaining system health checks and accessibility. |
06-Mar-25 |
|
|
Ensemble AI: Save and Retrieve Design Layout via API |
The Ensemble AI API now supports saving and retrieving layout designs, enabling merchandisers to define and apply custom layouts for outfits. This enhancement ensures that design configurations—including image positioning, dimensions, and layering—are stored and accessible via both portal and client-facing APIs. These layouts can be used for rendering ensembles in dynamic experiences and active content, enhancing the visual appeal of displayed outfits. |
06-Mar-25 |
|
|
Streaming Catalog: Parallel Scoped Actions for Improved Efficiency |
Scoped actions can now run in parallel across partitions/sites, preventing smaller sites from being blocked by long-running actions on larger catalogs. This enhancement improves processing efficiency and reduces wait times. |
06-Mar-25 |
|
|
Streaming Catalog: Active Snapshot Subscription for Enrichment Calculation |
Active snapshots can now be subscribed to an enrichment calculation dataset, eliminating the need to create a new snapshot for hybrid search. This enhancement ensures catalog vectors are automatically reconciled with ingested products, streamlining the process with minimal customer involvement. |
06-Mar-25 |
|
|
Streaming Catalog: Add Property Definition Collection to Existing Snapshots |
Existing snapshots can now be updated with new property definition collections without requiring a new snapshot. This enhancement reduces friction for customers by enabling seamless rollout of new search features like query tags and hybrid search. |
06-Mar-25 |
|
|
Streaming Catalog: Add Item Type to Subscription Events in Engine.out |
Subscription events in engine.out now include the itemType property, ensuring that enrichment calculations correctly update relevant item properties. This enhancement allows seamless handling of calculations for different item types while maintaining compatibility with existing enrichment, streaming, and item-store-consumer functionalities. |
06-Mar-25 |
|
|
Recommend: Track Boosted Product Clicks in Avro Logs |
Boosted product clicks are now logged in Avro logs, providing insights into the performance of Recommendation Boosting rules. The logs capture which products were boosted and by which rule, enabling better analysis and reporting on the impact of product boosting. |
06-Mar-25 |
|
|
Science: Use User Profile as a Seed in Ensemble AI |
Ensemble AI now supports using user history as the seed instead of requiring a product ID. Users can specify View History or Purchase History as the seed option, selecting 1 to 3 recent products. This enhancement enables personalized recommendations in email campaigns without manually providing product context. |
06-Mar-25 |
|
|
Additional Logging for Sync Model Listener |
Enhanced logging for model sync listeners in rrserver to help diagnose model sync issues across data centers. Similar to cache sync logs, these logs will track model sync activities, making it easier to troubleshoot inconsistencies. |
06-Mar-25 |
|
|
Ensemble AI: Style Performance: ThoughtSpot Visualizations |
ThoughtSpot visualizations are now integrated into Ensemble AI Style Performance Reporting, providing enhanced data analysis capabilities. The table visualization includes key dimensions such as date, styles, channel, region, and currency, along with performance metrics like views, clicks, orders, CTR, conversion rate, and RPV. |
06-Mar-25 |
|
|
UPS: Current Cart and Wishlist State in UPS Response |
To ensure accurate cart and wishlist data, UPS now derives the current state of the cart and wishlist for each user, addressing cases where view events are missing or outdated. The updated UPS response now includes real-time cart views, added items, and removed items. |
06-Mar-25 |
Bug and Support Fixes
The following issues have been fixed in the release version 25.05.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Enterprise Dashboard: UI Fixes for Product Comparison Rules |
Resolved an issue where Product Comparison rules were sending an array of all previous global configurations instead of just the current one, aligning with backend expectations. Additionally, the Shared Attributes Count field has been updated from a slider to a simple input field for improved usability. |
06-Mar-25 |
|
|
Enterprise Dashboard: Fixed Product Catalog Links for IDs with Backslashes |
Product links in the dashboard now correctly encode backslashes, ensuring they load properly in the Product Catalog. This fix applies to links from various sections, including Strategies, User Profile, Configurable Strategies preview, Advanced Merchandising, Real-Time Report, and Recs Test Drive. |
06-Mar-25 |
|
|
Social Message: Fixed Priority Order for Social Proof Badging |
Social Proof Badging now correctly prioritizes the top badge when multiple badges are disabled. Previously, the system incorrectly assigned the highest priority to the bottom badge. |
06-Mar-25 |
|
|
Streaming Catalog: Fix for Incorrect Conversion of Custom Product Attributes |
Custom product attributes of type long_list and int_list were incorrectly converted to string_list. This issue, seen in streaming-view and recommendation responses, has been resolved by fixing the attribute extraction logic to correctly handle long_list and int_list types. |
06-Mar-25 |
|
|
Streaming Catalog: Fix for SFI Downtime in RDN and FRA Data Centers
|
SFI experienced downtime due to a delete request for the catalogtagger collection, leading to increased Kafka lag. The issue has been resolved by ensuring snapshot activation only proceeds when no active scoped actions are in progress, preventing conflicts that could cause failures. |
06-Mar-25 |
|
|
Recommend: Fix for Manual Recommendations Not Playing |
We have fixed an issue in Recommendations in Email, where the response did not correctly apply the minimum item count when multiple layouts were configured for a placement. With this fix, the response now consistently uses the minimum item count specified in the request, ensuring expected product recommendations. |
06-Mar-25 |