Release Summary - Jul 16, 2024 (24.13)

ENG-28167

Enterprise Dashboard:

UI Configurable Strategies Fix for 'Multiple Items from History' Option

The "Use Multiple Items from a User's History" option has been refined for better usability. Now, it only activates when a Personalization Seed option in the "Based on User's History" section is selected. Additionally, this option is always inactive if a "Context" section option is chosen, ensuring the correct parameter is passed for multiple item usage without user intervention.

This update ensures that when multiple fixed category seeds are used, the preview and saved strategy accurately reflect recommendations from all fixed categories, providing a more consistent and expected user experience.

ENG-28576

Enterprise Dashboard:

User ID Input Field Added for User Browse History Model in Configurable Strategies

The Configurable Strategies UI now includes a User ID input field on the View Results tab for the User Browse History model. This enhancement allows digital merchandisers to preview the User Browse History model for any user by providing their user ID.

The User ID input field is required and is used in the payload for fetching recommendations.

ENG-28405

Enterprise Dashboard:

Recs Test Drive - Enhancements for Dev Mode and Custom Parameters

The Recs Test Drive feature has been enhanced to support dev mode, ensuring that test requests do not impact recommendation models and reporting. The ‘dev=true’ parameter has been added to all requests, and the ‘discoverDebug=false’ parameter has been removed.

Additionally, custom parameters are now properly included in requests. The custom parameters field has been moved for better visibility, widened for easier input, and includes placeholder text to guide users. Validation has also been added to ensure proper formatting of the parameters.

ENG-27098

Enterprise Dashboard:

Sales by Channel Report - Label Changes for Metrics

To maintain consistency across metrics, several labels in the Sales by Channel Report have been updated:

"Units per order from recs" has been changed to "Recs IPO (Items Per Order)"

"Non Recs IPO" has been changed to "Non Recs IPO (Items Per Order)"

"Orders from Recs" has been changed to "Rec Orders"

These changes ensure clear and uniform terminology throughout the report.

ENG-27627

Science:

Dashboard API for Compatibility Rules

Digital merchandisers can now create and manage compatibility lists for Advanced Merchandising, enhancing the recommendation of compatible products. This new feature allows the definition of Compatibility Rules, enabling merchandisers to specify seed and rec categories, attributes, and examples for compatibility data. These inputs are used by ChatGPT to extract compatibility strings, ensuring accurate recommendations.

The new API supports saving rule inputs to a database, managing multiple rules, and retrieving rule settings for display and editing in the dashboard. Merchandisers can create one-way or two-way compatibility rules and specify comparison types, providing flexibility and precision in setting up compatibility rules.

ENG-28246

Recommend:

Add Sale Price to Advanced Merchandising Dashboard API Responses

The Advanced Merchandising dashboard now includes sale prices in its preview, reducing confusion for digital merchandisers when defining contexts or recommendation groups based on price.

Both the /getSeeds and /getRecs API responses have been modified to include the sale price of products. This enhancement ensures that the UI displays the sale price when available, providing a more accurate and useful preview for merchandisers.

ENG-28064

Recommend:

Review of Slack Alert Logic and Notification Threshold for Email Recs

We have revised the logic and thresholds for Slack alerts related to email recommendations. This review ensures more accurate identification of issues within the email stack, reducing false positives and negatives, and enhancing the reliability of alerts.

ENG-27913

Recommend:

Use Primary Category as Seed for Strategies Using Category Affinity

When using a strategy where the seed is the user's category affinity, there is now an option to ensure the selected category is a primary category. This enhancement ensures that the selected recommendations and the strategy hint in the message are meaningful to the shopper, avoiding the use of non-primary marketing categories.

ENG-28116

Find:

User Affinity API for Find

A new User Affinity API has been introduced to enhance the Find feature with personalized facets and auto complete. This API provides access to user affinity data, which helps in delivering more relevant search results and suggestions based on individual user preferences.

The API allows all user identifiers (userId, sessionId, rcs) to be passed as parameters, ensuring flexible and accurate identification. Affinities are returned based on the userId, or if not provided, based on the provisional user ID. The affinity scores are derived from the specified affinity configuration, or the default configuration if none is specified, ensuring consistency in personalization.

ENG-28339

Find:

Visibility in Search Service to Find Filter

The search service now includes visibility support in the Find filter, ensuring more precise and context-aware search results. This enhancement allows filtering results based on visibility settings, user context, and region parameters, providing a more tailored search experience.

ENG-21637

Discover, Engage, Find, Recommend:

Enable Find Filter for Algonomy Products in Search Service

A new enhancement enables the Find Filter for other Algonomy products such as Micro Segment, Recommend, and Discover. Clients will now send a JSON object with candidates to the search service, which will be processed and converted to Find Filter format, integrating seamlessly with Solr queries.

The search service now includes a new endpoint (searchservice/findFilter) to handle these requests and generate appropriate responses. This update ensures enhanced filtering capabilities across various Algonomy products.

PLAT-3690

Find, Streaming Catalog:

Address Enrichment Attributes in NCA Before Storing and Processing

We have enhanced the handling of enrichment attributes in the NCA to ensure proper processing before storing enrichment-related data.

Previously, enrichment data was stored in the catalog database even if the "storable" flag in "catalogSettings" was set to false. This enhancement ensures that the NCA checks the "storable" flag before storing data in the catalog database, preventing unnecessary storage of non-storable attributes.

ENG-28535

Social Proof:

Badge Editing Issue in Social Proof Campaigns

We have resolved an issue in the social proof badging campaign where previously added languages were incorrectly displayed when editing badges. This issue caused newly added languages to show incorrectly across different badges after editing.

Now, when users edit a badge to add new languages, the correct languages will be displayed for each badge, ensuring that the language settings are accurate and specific to each badge.

ENG-28454

Recommend + Email:

Clear Page Contents Before Reusing the Page in Puppeteer

We have resolved an issue affecting the rendering of prices in email recommendations. Previously, the layout logic for displaying sale prices in bold and list prices in smaller fonts was inconsistent. This caused the list price to sometimes appear in smaller fonts, leaving the price section empty when a sale price was not available.

The logic has now been corrected, ensuring that products with only a list price display the list price in bold as intended.

ENG-28381

Find:

Reformulate Failure for Batch Client Resolved

In Recommend + Email, there were some edge cases where previously loaded elements were captured in the screenshot for a different recommendation. To prevent this from happening, we’ve modified Puppeteer to clear the page before taking the screenshot of the recommendation.

ENG-28366

Default Region ID -1 Issue in Autocomplete Requests

We have fixed an issue where the default region ID -1 was being passed in autocomplete requests, leading to no results when the region parameter was left empty by the client.

The fix ensures that autocomplete requests without a specified region will no longer default to region ID -1, resulting in accurate search results.

ENG-28181

Find:

SFI Downtime Due to Replica Count Issue

We have fixed an issue where SFI would go down if the replica count set in the portal was higher than the available replicas. This issue caused errors and interrupted the SFI process, leading to downtime.

With the fix, SFI now properly handles scenarios where the requested replica count exceeds the available nodes, ensuring smooth operation without interruptions.

ENG-27206

Find, Recommend:

Search Term Context Not Working for Recommend Restriction Rules

We have fixed an issue where "Do Not Recommend" rules were incorrectly using ruleTemplate 19 for Search Term context, making it applicable only for Find. This misled Recommend customers who expected it to work with Recommend placements.

Now, a new checkbox is available in merchandising rules for Search Term context to specify "Use for Find." By default, the rule will work for Recommend when the checkbox is unchecked. When checked, the rule will work for Find, with the Page Type set to Search Page. This option is available for both Rec Restriction and Rec Boosting rules.

ENG-28561

Enterprise Dashboard:

Category and Brand Not Displayed for Models in Configurable Strategies

We have fixed an issue where categories and brands were not being displayed correctly for certain scenarios in Configurable Strategies. This occurred when switching from a fixed category or brand seed to category or brand affinity in personalized seed selection.

Now, the categories and brands are displayed correctly when using any personalized seed options, ensuring a seamless experience for users.

ENG-28231

Enterprise Dashboard:

Configurable Strategy Models Region Seed Issue Fixed

We have resolved an issue in Configurable Strategies where the "Top Views Last Hour" model did not function correctly with the region as a seed. To prevent errors, the "Use Region as a Seed" option is now disabled for the Best Offers, Replenishment, Category-based Replenishment, Top Sellers Last Hour, and Top Views Last Hour models. This update ensures smoother configuration and prevents errors when using these models.

ENG-28575

Data Engineering, Social Message:

Social Proof Output Response API -Threshold Update Issue Fixed

We have resolved an issue where changes to thresholds for Cart Page experiences were not reflected in the API response. This issue affected Abercrombie & Fitch, where updates to the threshold were inconsistently applied across multiple experiences. The API now correctly shows the updated threshold for all experiences as expected.

ENG-28394

Data Engineering, Enterprise Dashboard:

MVT Multi-Dimensional Reporting (UI): Report Loading Issues Fixed

We have resolved issues with multi-dimensional reporting where reports failed to load properly when multiple dimensions were selected. Previously, reports would stall after selecting several strategies or changing date ranges, making the feature unusable.

To address this, the raw data API call is now triggered only when the user explicitly requests it by clicking the "Visit Data" option. Additionally, filters are disabled until the report fully loads to prevent multiple report requests from overloading the system.

ENG-28033

Advanced Merchandising, Recommend:

Adv Merch Recs Not Showing in Item Page for Specific Product

We have resolved an issue where Advanced Merchandising recommendations were not appearing in a placement when multiple placements were requested.

The problem was identified as excessive de-duplication of products. After adjustments, the recommendations are now correctly displayed for all products, including when multiple placements are requested simultaneously.