Release Summary 25.01 | Jan 09, 2025
The following key features and improvements, along with bug fixes, have been released in Algonomy CXP products in the release version 25.01.
Recommend
Advanced Merchandising as a Strategy or Fallback
We have introduced a new feature that allows Advanced Merchandising to be used as either a strategy or a fallback, providing greater flexibility for digital optimization managers. This enhancement enables Advanced Merchandising to compete with other strategies or be used specifically for long-tail products when behavioral cross-sell strategies have valid models.
A new site configuration, "use AM as strategy," has been added to control this functionality, which is disabled by default. When enabled, Advanced Merchandising will follow the normal strategy selection logic, applying filtering and boosting without backfilling slots. Predefined default behaviors include enabling system filters and removing duplicates.
Jira: ENG-29105
Top Views model Added to Configurable Strategies
Configurable Strategies now support the Top Views model, allowing optimization managers to create strategies based on CategoryTopViews and BrandTopViews for enhanced engagement.
This feature enables users to configure the Top Views model with options such as sitewide views, category or brand context, fixed seeds, and category or brand affinity with multiple items. Sorting options include User Affinity and Smart Shuffle, while filtering options allow for No Filter, Category Diversity, or Brand Diversity. Additionally, the preview breadcrumb dynamically updates based on these selections for accurate strategy previews.
Jira: ENG-29367
Update to 'Most Recent View History' Personalization Seed
The "Most Recent View History" personalization seed now selects products exclusively from a shopper's most recent browsing session, ensuring recommendations are highly relevant to their latest activity.
When this option is selected in Configurable Strategies, seed products are drawn from the most recent session, with the number of products determined by the relevant parameter. The defined time frame is respected—if the most recent session occurred outside this window, no recommendations will be returned.
Jira: ENG-29513
Reintroduction of 'User History Expire Days' for Personalized Strategies
The 'user history expire days' Site Configuration setting has been re-enabled to provide greater flexibility in selecting seed products for legacy Personalized strategies.
This global configuration now controls the timeframe for selecting products from a shopper's history for all legacy Personalized strategies. For Configurable Strategies, the strategy-specific time frame parameter takes precedence. If no time frame is configured for a particular strategy, the system defaults to the 'user history expire days' setting.
Jira: ENG-29514
Engage
Max Views Count in Content Catalog
The Content Catalog has been enhanced to help digital marketers monitor and manage performance for content with a max views setting.
The catalog now displays the current view count, total view count, and the last updated date and time for content with a max views threshold. These details are conveniently positioned between the max views setting and the tags section on the Content view page. Additionally, a new API has been introduced to the dashboard, providing max views count data. The API includes results from all jobs since the content_frequency_start_date and the date and time of each model build, ensuring comprehensive insights.
Enhanced Tag Update: Streamlined Dashboard Automation
Managing and applying enhanced tags has been made more efficient with new automation enhancements, reducing manual effort while ensuring accurate content tagging.
A bulk update feature now automates the process of re-saving all associated content whenever an enhanced tag is modified. This background job, running outside the dashboard UI, ensures consistent application of updated tag settings without requiring manual intervention. Additionally, a mechanism triggers a ps_catalog job during content publishing to ensure updates are synchronized across the system.
Jira: ENG-29439
Visual Styling for Enhanced Tags in Content Catalog
Enhanced Tags in the Content Catalog now feature distinct blue styling, making them easily identifiable from normal tags.
Tags with an enhancedTag value or matching the expected syntax (e.g., starting with brand:, category:, or attribute:) are displayed in blue. Additionally, hovering over an Enhanced Tag reveals its associated enhancedTag value, providing quick insights into its context.
Jira: ENG-29508
Ensemble AI
Ensemble AI: Transparent Image Generation for Catalog
We’ve enhanced Ensemble AI with support for transparent product images, enabling merchandisers to create visually realistic and engaging product combinations. Transparent images allow products to overlap seamlessly, offering shoppers a more immersive and accurate visualization of how different items look together.
The Client Facing Ensemble AI API and Portal API have been updated to include URLs for transparent images, either provided by clients or generated by us. If neither option is available, the APIs exclude the entry, requiring front-end and portal support for such cases. This dynamic handling ensures that transparent images are prioritized when available while maintaining flexibility.
Jira: ENG-28170
Ensemble AI: Configurable Partial Outfits
We’ve enhanced Ensemble AI with the ability to configure whether partial outfits are included in API responses, allowing merchandisers greater control over outfit generation and ensuring compatibility with Active Content requirements.
A new API parameter, ViewPartialOutfits, has been introduced. By default, partial outfits are included in the API response, but setting this parameter to false ensures only complete outfits are returned. This is particularly useful for Active Content, which requires a fixed number of parts in an ensemble. Additionally, another parameter, parts, enables merchandisers to specify the exact number of parts (3-6) required in an outfit. If the specified number of parts does not match, no outfits are returned, ensuring strict adherence to configuration.
Jira: ENG-27986
Ensemble AI: View/Click Tracking for Similar Products
Ensemble AI now tracks user interactions with the Similar Products feature, allowing merchandisers to monitor performance and attribute actions to specific outfits or styles.
Jira: ENG-28804
Ensemble AI: Customizable Layout Design for Styles
Merchandisers can now define and apply layout designs for styles in Ensemble AI, enabling visually appealing and tailored presentations for ensembles with transparent images. This feature allows the creation of Stitch Fix-style designs with overlapping images to enhance the shopper experience.
Layout designs can be defined at the style level, with parts corresponding to the style's definition. Users can adjust coordinates such as height, width, rotation, and order to control the visual hierarchy of parts. Blocks can be resized, moved, or layered, and changes can be previewed before finalizing. Undo and redo options are available for modifications until the design is saved.
These enhancements provide greater creative control over ensemble layouts, allowing merchandisers to deliver a more engaging and cohesive shopping experience.
Jira: ENG-29431
Ensemble AI: Privacy Mode Support
Ensemble AI now respects user consent by supporting a privacy mode parameter, ensuring that ensembles are not personalized for users who decline behavioral data tracking.
Clients can enable privacy mode by passing the parameter R3_COMMON.privateMode=true. This updates the Ensemble AI API calls (both client-facing and portal APIs) to include privm=true, ensuring that outfits generated through the Ensemble AI dynamic experience template align with user consent preferences.
Jira: ENG-28940
Social Proof
Social Proof A/B Testing with MVT
A new test type, Social Proof, has been added to the MVT testing framework, enabling digital optimization managers to evaluate the impact of social proof on revenue per visit.
This enhancement allows users to define test details, including the test name, traffic percentage, and test duration. Control and treatment options are preconfigured, with the control set to "Social Proof OFF" (API service disabled) and the treatment set to "Social Proof ON" (API service enabled). Reporting for these tests is consistent with existing MVT reporting standards, ensuring clear and actionable insights.
Jira: ENG-25838
Export and Download Social Proof Reports
The Social Proof Report now supports exporting and downloading both data and visualizations, providing enhanced flexibility for analysis and sharing.
Users can download the report table in CSV or Excel formats directly from the Social Proof Report tab within Social Proof Experiences. Additionally, visualizations such as graphs can be exported in PNG or JPEG formats, ensuring high-quality, shareable insights.
Jira: ENG-29487
Guided Selling
Guided Selling: Duplicate Screens and Screen Title Design Options
Guided Selling has been enhanced to improve customization and ease of use for digital marketers. Users can now duplicate existing screens with all configured values copied to a new screen, which is added at the end of the screen tabs. A new menu, accessed via a three-dot icon on the screen tabs, offers options such as Edit Screen Name and Duplicate Screen for streamlined management.
Additionally, new design options for the quiz step allow users to customize the Screen Title Font Size and Font Color under the Style section.
Jira: ENG-29292
Enterprise Dashboard
Ingredient Cross Sell: Category Selection Enhancement
We’ve introduced a new feature that allows digital optimization managers to configure the categories of products used by the Ingredient Cross Sell model. This enhancement ensures that only relevant product categories, such as food-related items, are considered as candidates for cross-sell recommendations, providing a more precise and personalized shopping experience.
On the Model Options page, a new option has been added specifically for the Ingredient Cross Sell model. Using the standard category picker, users can now easily select or exclude categories, tailoring the model's behavior to suit their needs.
Recs Test Drive: Enhanced Input Fields and Debugging Tools
We’ve enhanced the Recs Test Drive feature to improve efficiency in configuring recommendations and troubleshooting potential issues. These updates provide greater flexibility and clarity for optimization managers, enabling more precise adjustments and debugging.
Key improvements include:
-
Dynamic Input Fields: Based on selections, relevant input fields (Product, Category, Brand, and Search Term) now appear when the "Force Strategy" option or an email placement is selected.
-
Automated Debugging Assistance: If no recommendations are returned, the Rules tab for the selected placement or strategy is automatically displayed, aiding in quick troubleshooting.
-
Improved Category Recommendations: The request URL is now visible at the top of the Categories tab, similar to how it appears on the Products tab, for added transparency.
-
Additional Inputs: A Product input field has been added for wish list page placements.
Jira: ENG-29471
Sorting of Discover Attributes
We’ve introduced alphabetical sorting for attributes on the Browse Configuration page, making it quicker and easier for digital merchandisers to find the desired attributes.
Attributes in the dropdown menus for Personalized Weights, Global Weights, and Attributes for Browse sections are now displayed in alphabetical order. This improvement enhances usability and efficiency when configuring Discover, allowing merchandisers to navigate and select attributes with ease.
Jira: ENG-29204
Sorting of Strategies in Product Catalog
Strategies in the Product Catalog are now displayed in alphabetical order, making it easier for digital merchandisers to locate and preview specific strategies quickly.
This enhancement improves usability by streamlining the process of navigating through strategies, enabling more efficient configuration and previewing of results.
Jira: ENG-29202
Improved Search Experience on Strategy Configuration Page
The search functionality on the Strategy Configuration page has been updated to enhance responsiveness and usability. Instead of filtering results instantly as users type, the search is now triggered only when the user submits the query by pressing "Enter."
Jira: ENG-29059
Standardized Region Display in the Dashboard
Regions are now displayed consistently across the dashboard to reduce confusion for merchandisers creating rules.
For Do Not Recommend rules, Recommend Only rules, Strategy rules, and ConfigRR, the regions are displayed in the format: "region name (region id)". This enhancement ensures clarity and uniformity when managing rules and configurations.
Jira: ENG-29365
Reserved Names Restricted for API Client Keys
To prevent conflicts and improve functionality in Recs Test Drive and reporting, the system now restricts the use of default channel names as API client keys.
Users attempting to save an API client key with reserved names such as Desktop, Phone, Tablet, or Email will be blocked. A clear error message is displayed, informing users that these names are reserved and suggesting the use of alternative names.
Jira: ENG-29453
Reset Button Added to Recs Test Drive
A Reset button has been added to the input fields in Recs Test Drive, allowing digital merchandisers to easily clear seed fields (product, category, brand) and test different scenarios.
With a single click, all seed fields are cleared, streamlining the process of configuring and evaluating various test cases.
Jira: ENG-29024
Alert for Missing Strategy Messages in Strategy Rules
To prevent confusion, the dashboard now alerts users when a strategy lacks a strategy message during the creation or editing of Strategy Rules.
When a strategy is added to the Preferred Strategies list, or if existing strategies are present, the system checks for a strategy message corresponding to the selected page type. A new column in the Preferred Strategies table displays the message: "Don't forget to add a Strategy Message" for strategies missing this configuration. The message links directly to the Strategy Configuration page for quick updates.
Jira: ENG-29201
Category Details on Configurable Strategies in View Results
Category details are now displayed in the View Results section for strategies that recommend categories. Results include category images, names, and IDs, shown in a new Categories tab, alongside the existing Products tab for product recommendations.
Data Engineering
Zero Search Terms Report: Visualizations for Insights
The Zero Search Terms Report now includes visualizations to help Search Optimization Managers identify and address search terms yielding zero results. These insights enable managers to refine search queries and enhance search performance effectively.
The report includes a detailed table showing key data points such as Date, Region, Search Term, Facets/Filters, Number of Filters, and Number of Queries with Zero Results. Wildcard entries (e.g., *) are excluded to ensure meaningful data. Additionally, a bar chart visualizes the top 50 search terms with zero results, sorted by query count, offering a clear view of critical areas for optimization.
Jira: ENG-28526
Channel Dimension Added to Zero Results Reports
The Zero Results Reports for Find Search Terms now include Channel as a dimension, allowing merchandisers to analyze search terms resulting in zero results by channel.
This enhancement enables both aggregated and de-aggregated views of search terms by channel, helping identify channels with higher zero-result occurrences.
Jira: ENG-29141
Other Feature Enhancements
The following feature enhancements and upgrades have been made in the release version 25.01.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Enterprise Dashboard: Note Added to Primary Categories Page |
A note has been added to the Primary Categories page stating, "Updates to this configuration take effect after the next product feed file is processed or API update request received." This message is displayed in a save box, requiring the user to select "OK" to close it.
|
09-Jan-25 |
|
|
Find: Collection Count Matching Across Data Centers |
A script has been implemented to compare collection counts across data centers for each site, identifying any discrepancies. The script evaluates collections by site ID and data center, comparing counts from Solr 7 and Solr 9 to highlight differences. Detailed results, including discrepancies and collection details, are provided for further analysis. |
09-Jan-25 |
|
|
Find: Solr Upgrade to 9.7.0 |
The Solr version has been upgraded to 9.7.0 to address a bug in Solr 9.6.0 that caused replicas to remain in a downstate after node restarts, leading to system instability, particularly in the DAM data center.
|
09-Jan-25 |
|
|
Find: Exact Match for Search Synonyms with Hyphens/Dashes |
Enhancements have been made to ensure exactMatch behavior for search_syns and search_auto fields when handling search terms with hyphens or dashes. Now, for configurations like articleNumber set to exactMatch, products will not be returned for partial matches if the search term does not fully match the field value. |
09-Jan-25 |
|
|
Find: Optimization of Data Sending to Algonomy |
Improvements have been implemented to optimize data sending for regions using shared product IDs across multiple languages. Each country's product data is now efficiently distributed to corresponding language-specific Solr collections, reducing redundancy in the English Solr collection. This optimization significantly reduces data size from 160M to 62M by streamlining headers and language-specific collections. |
09-Jan-25 |
|
|
Find: Enhanced Validation and Retry Mechanism for Find Deployer |
The Find Deployer now includes enhanced validation and retry mechanisms for handling failed indexes and deployment failures. It validates failed indexes and retries in cases of 404 errors from the sync store, with alerts and retry information sent to Slack when a live dataset is missing, ensuring prompt action. Additionally, the system automatically triggers the retry API for all retryable exceptions, streamlining deployment processes and ensuring smoother operations. |
09-Jan-25 |
|
|
Find: Live Dataset Availability Check for Enrichment |
A validation process has been added to check live dataset availability for all sites using enrichment. Alerts are sent to Slack if any live dataset is missing, ensuring timely monitoring and resolution. |
09-Jan-25 |
|
|
Find: querySynonyms=false Support in Complementary Search Config |
Merchandisers can now configure querySynonyms:false in the complementary search config JSON to disable synonym usage for testing MVT with and without hybrid search. This feature is applicable only for streaming sites. |
09-Jan-25 |
|
|
Find: Optimization for Category Name Changes in SFI |
SFI has been optimized to reduce memory usage when updating product documents after a category name change. Previously, SFI retrieved all product fields from Solr, leading to high memory consumption and OOM (Out of Memory) issues during category name updates. The Solr query has now been refined to fetch only necessary fields, significantly reducing the data loaded into memory. After this optimization, SFI can efficiently handle category updates for large datasets. |
09-Jan-25 |
|
|
Find: Remove unnecessary metrics |
Unnecessary metrics in Find Deployer have been removed, retaining only app startup, deployments started, and deployments failed. A new metric for deployment success has also been added. These changes simplify monitoring and improve focus on critical metrics. |
09-Jan-25 |
|
|
Find: Automatic Retry for Failed Deployments in Find Deployer |
Find Deployer now includes automatic retry logic for failed deployments, with a defined retry limit. Retry counts are also added to the failed deployment metric for better tracking. |
09-Jan-25 |
|
|
Find: Daily Scheduling of Vector Jobs |
Vector jobs, including DataScience Airflow and Azkaban tasks, are now scheduled to run daily. This enhancement ensures consistent and timely execution of vector-related processes. |
09-Jan-25 |
|
|
Platform: API for Generating Signed Hash for Online Help |
A new API endpoint has been introduced to generate a signed hash for online help documentation. The API dynamically constructs the updatedHash using user IP, expiry time, OLH type (internal/external), and environment (QA/PROD), and then signs it with the HMAC SHA256 algorithm. This API supports GET and POST methods across modules and ensures secure, standardized access to online help documentation. |
09-Jan-25 |
|
|
Streaming Catalog: Scoped Action for Snapshots with Active Enrichment Subscriptions |
Scoped actions for snapshots with active subscriptions to enrichment dataset calculations have been updated to ensure that products and categories are correctly sent to engine out. |
09-Jan-25 |
|
|
Streaming Catalog: Enhanced View Store API to Fetch Headers |
The View Store Ingestable API now fetches headers ingested with each item in the streaming catalog. This update supports Atea data optimization by allowing clients to specify language in headers, ensuring data is indexed only in the relevant language-specific Solr collections. |
09-Jan-25 |
|
|
UPS: Wishlist Enhancements |
We’ve updated wishlist settings to streamline functionality. Default views are now set to 1, with up to 3 wishlists per profile. A new switch allows enabling or disabling wishlist event storage in the User Profile Store (disabled by default). Additionally, a configuration limits items per wishlist event to 100 by default. |
09-Jan-25 |
|
|
|
Data Engineering: Ensemble AI Reporting - Style ID to Name Mapping |
We have enhanced the reporting process for Ensemble AI by adding Style ID to Style Name mapping in the lookup process. This mapping is now updated daily in Redshift and Snowflake and is integrated into detailed Ensemble AI reports in ThoughtSpot, enabling users to view Style Names instead of IDs for improved clarity and usability. |
09-Jan-25 |
|
Data Engineering: MVT Reporting: Support for Social Proof Test Type |
The MVT API module has been updated to support Social Proof Test Types. MVT reports for Social Proof are now available, and events related to personalized API calls correctly include MVT details. |
09-Jan-25 |
|
|
Platform: French Language Support Added |
The system now supports French as a language option, enabling a localized interface for users. Under the user profile settings, employees can select French from the language dropdown, and all system labels will be displayed in French. Processes have been implemented to translate property files and import the translated versions into the system, ensuring seamless integration of the localized content.
|
09-Jan-25 |
|
|
Platform: Internal Help vs Customer Help |
Help documentation now separates internal and external content, ensuring internal resources are accessible only to Algonomy employees. Algonomy employees see two links: Help (customer-facing) and Internal Help (internal-facing). External users see only the Help link, offering documentation in respective language. Access requires login, and deployments are segregated across QA and PROD for secure content management. |
09-Jan-25 |
Bug and Support Fixes
The following issues have been fixed in the release version 25.01.
|
Jira # |
Module/Title |
Summary |
General Availability |
|---|---|---|---|
|
Recommend: Scheduled Query Issues in Data Science Workbench |
We have resolved a bug where the selected query was not visible in the UI when editing a scheduled DSW strategy. |
09-Jan-25 |
|
|
Recommend: User Affinity Configuration Update Issues |
We have resolved issues affecting the User Affinity Configuration, including a bug that prevented updates to the default configuration. |
09-Jan-25 |
|
|
Recommend: Default Category Tier Weights in User Affinity Configuration |
We have resolved an issue with the default User Affinity Configuration. Category tier weights have been updated from 1 to 0, ensuring optimization managers must opt-in to use the category tier options. |
09-Jan-25 |
|
|
Engage: Brand and Category Affinity Not Working in Engage User Affinity |
We have resolved an issue where brand and category affinity-based configurations were not functioning in Engage User Affinity. The catalog manager now retrieves brand and category properties correctly using the getProductById method. |
09-Jan-25 |
|
|
Ensemble AI: Duplicate Products in Ensemble AI Outfits |
We have resolved an issue in Ensemble AI where duplicate products were appearing in generated outfits. Outfits now display unique products as intended. |
09-Jan-25 |
|
|
Ensemble AI: Product Image Not Loading in Ensemble AI Layout Design |
We have resolved an issue in Ensemble AI where product images were not loading while designing the layout. The backgroundlessImageUrl issue has been addressed, and images now display correctly. Additionally, the design layout save functionality has been improved to prevent saving styles with layout issues. |
09-Jan-25 |
|
|
Enterprise Dashboard: Incorrect Rule End Dates Displayed
|
We have resolved an issue where rules displayed incorrect end dates after being saved. The correct dates now appear consistently across Advanced Merchandising, Recommend boosting/restriction, and strategy rules. |
09-Jan-25 |
|
|
Enterprise Dashboard: Targeting Non-Recommendable Products in Context Builder |
We have resolved an issue where "N/A" was incorrectly appended to product names when targeting non-recommendable products in the Context Builder. The "N/A" text is now removed, and the targeting API correctly matches on the product name. |
09-Jan-25 |
|
|
Enterprise Dashboard: Missing Alert When Switching Languages in Guided Selling |
We have resolved an issue where no alert message was displayed when attempting to switch languages after modifying a translatable field in a Guided Selling campaign. Users will now see an alert prompting them to save or discard changes before switching languages. |
09-Jan-25 |
|
|
Enterprise Dashboard: Intermittent UI Issue with Scheduled Query Status in DSW |
We have resolved an intermittent issue where scheduled queries in the Data Science Workbench appeared empty after being saved. The UI now reflects the correct status consistently. |
09-Jan-25 |
|
|
Find: Arabic Search Terms Not Honoring minQueryTerms
|
We have resolved an issue where Arabic search terms were not honoring the minQueryTerms setting due to tokenization issues. Vector search now works correctly with Arabic search terms, respecting the minQueryTerms configuration. |
09-Jan-25 |
|
|
Find: Exception During Solr Response Parsing in Find Deployer |
We have resolved an issue causing exceptions while parsing malformed Solr responses during index deployment. The updated find-deployer version now handles responses more robustly, allowing successful index builds for all cases. |
09-Jan-25 |
|
|
Find: Incorrect Category Tree in FIND Responses |
We have resolved an issue where FIND responses were returning an incorrect category tree, despite the correct hierarchy being present in the system. The category tree now reflects the correct structure, ensuring accurate results for category-based filtering. |
09-Jan-25 |
|
|
Find: Publishing Issue on Search Term Expansion Page |
We have resolved an issue where the publish button on the Search Term Expansion page did not function consistently. Changes now publish correctly to Solr across all data centers, ensuring consistent results. |
09-Jan-25 |
|
|
Find: Facetable Errors in Find Response on Staging
|
We have resolved an issue in the staging environment where Find responses incorrectly displayed errors for fields "product_saleprice_cents" and "product_pricecents" as not supporting facetable operations. The issue has been addressed, ensuring consistent responses between staging and production environments. |
09-Jan-25 |
|
|
Data Reporting: Missing Daily Event Date Filter in Sales by Channel Report |
We have resolved an issue where the daily_event date filter was not visible in the Sales by Channel Report, preventing users from adjusting the graph view. The filter now appears correctly, ensuring seamless functionality across all sites. |
09-Jan-25 |
|
|
Streaming Catalog: SetUp Issue for LSI VectorSearch Testing |
SetUp Issue for LSI VectorSearch Testing We have resolved an issue that prevented setting up the site for testing VectorSearch locally for Liquidity Services (1806). The data ingestion process now works correctly, addressing errors related to auction_start_date and auction_end_date. |
09-Jan-25 |
|
|
Streaming Catalog: Dataset Automatically Archived for WoC Calculation |
We have resolved an issue where datasets for WoC calculation were being automatically archived, causing missing fields in Solr for site 2105. The problem has been addressed, and datasets now remain active as expected, ensuring uninterrupted functionality for WoC and other Find features like vector search and global rank. |
09-Jan-25 |