Corrected an issue that caused special characters in Engine API requests to result in 500 errors instead of the 200 status.
Corrected an issue that caused the Recommendations Overview dashboard to report active recommendations experiences along with recommendations experiences that ran during the selected lookback period when in fact the account had no live recommendations experiences and vice versa.
Corrected an issue in the Product Finder wizard that was preventing the appearance of the pop-up message warning a user about unsaved changes.
Corrected an issue that prevented a user from deleting an ID Collector.
Corrected an issue in which the Engine AP was inaccurately identifying the country of origin for some IP addresses received from a client in data sessions.
Corrected an issue with filters used in Product Finder questions that was causing some customers to see the same products regardless of how they answered the questions.
Corrected an issue that was causing variant label responses to be truncated in Engine API responses.
Users can now create a filter for a recommendation strategy that uses the value from a specified custom variable that is passed at run time.
A Social Proof action can now be configured so that no message appears when the product's inventory reaches a designated level.
General filters and recommendation filters are now applied when calculating precompute.
The default message about previewing a recommendation strategy has been modified to inform users that a preview is not available for strategies configured with recommendation algorithms that rely on data calculated in real time, such as Recently Viewed and Replenishment.
The On product page category of action conditions now includes options to target or exclude pages for products based on a product attribute-based filter.
The Update History table for a Behavioral Triggers Extract dataset is now limited to the previous 90 days. The table had previously displayed all results.
Product Finder now notifies a customer with a "No Results Found" message when it cannot return recommendations based on the customer's answers to the questions in the Product Finder questionnaire.
Implemented a new Product Finder action type for Omnichannel experiences that can be returned by the Engine API.
The Viewed and Also Viewed and the Purchased and Also Purchased recommendation algorithms have been added to precompute so that they can make use of market-level data.
The button users click to edit the name of a target now works so that users can edit a target's name and save that change.
The product catalog specification now supports the quantity field so that clients can pass inventory counts to the platform using their product catalogs.
Corrected an issue that was preventing the Product viewed WHO target from working with product IDs that contain special characters.
Users can now use the Social Proof Messaging feature on product listing pages, search results pages, and cart pages.
The Maximum Threshold setting on the Social Proof Messaging action template enables the user to manage shoppers' perception of products becoming too popular by optionally configuring the maximum number of social media messages that can appear and what happens when the threshold is met.
Clients who have purchased the Recommendations add-on now can access the Merchandise tab in Action Builder.
Corrected an issue that was preventing recommendation strategies based on the Last Item Viewed recommendation algorithm from returning products to populate the recommendation slider.
Corrected an issue of data getting stuck in the pending state, so now recommendation strategies process consistently.
Corrected an issue that caused data retention failure when using the Partial Upload Only option to update Locations Manager.
The new Preview modal allows users to see up to the first 10 recommended products returned by a recommendation strategy. Users access the modal by clicking a button on the strategy's configuration page.
The UI of the recommendation strategy configuration page has changed so that the Base Recommendation on setting for collaborative recommendation algorithms has been replaced with two settings: the Base Recommendation on Items setting for selecting the type of customer behavior on which to base the recommendations, and the From setting for selecting the session scope from which to draw the customer behavior data. These changes give the user a better experience when configuring a recommendation strategy.
Improved the UI for the Datasets list page that appears for clients that have the one-click DATASETS option in the top navigation bar.
Corrected an issue that was throwing an internal server error when trying to delete a Product Finder questionnaire.
Users can now edit a market name and save it. If the same name is already in use for another market, then the change is not saved and a warning appears.
An issue was fixed that was preventing a user to duplicate an experience.
Products can now be badged using custom lists.
These return recommendations tokens that are returned with recommendations are used for reporting.
These affinity scores allow onboarded Recommendations datasets to work with the randomization setting and with Boost and Bury in recommendation strategies.
An experience is now activated only if the user confirms the activation on the Experience list page. This step prevents a user from accidentally turning on an experience just by clicking ACTIVATE on the Experience Editor page.
The Contains filter in recommendation strategies now supports multiple values.
With the delete functionality, retailers can now delete unused markets as well as edit and create them.
Recommendation were throwing an error due to special characters in spite of the catalog loading properly. This error was fixed such that recommendations can evaluate the data properly.
Relocated the Randomize Recommendations toggle to a better location below the lookback period configuration.
Subsequently Purchased premium recommendation type implemented to be used for cross-selling and re-engagement purposes.
With the removal of the feature flag, clients can now access the Markets feature without having to request access to it for the account.
Retailers can now delete a recommendation strategy.
Personalization users who have enabled the option to receive the catalog email notifications now receive the emails for catalog changes across the retailer.
New recommendation strategy report provides marketers a view into the performance of different strategies.
Corrected a bug that occurred when duplicating an Automated Personalization experience with a single variant holdout that caused it to change to random assignment
Expanded search capabilities to include a type ahead and increases the maximum number of results returned.
Added the ability to archive custom lists in recommendation strategies, action conditions, and WHO targets.
Extended session length to capture all views and clicks for endcaps.
Contact your Client Success Manager (CSM) for more information about premium capabilities.
Boost (push products to the top) or Bury (push products to the bottom) based on the product attribute of your choice in a product recommendations strategy.
No error should appear when creating an Automated Personalization experience.
The Custom Target Name should now be shown to the end user in the analytics report for an Automated Personalization experience.
Improvements to speed while searching for or switching between accounts in the Kibo Personalization UI.
Deduplicate products based on PIDs instead of SKUs.
Allows user to set a recommendation filter with product_type to match the exact entry in the product catalog.
Uploading a CSV with non-Unicode characters now returns an appropriate error message
Duplicating an experience with a Product API or product ID now has the correct homepage URL
Current Cart Value is now a valid choice when creating an Insert Dynamic HTML action
Description: On the dataset detail page for product catalogs, the "Products: ###" display showing the total count of products in the catalog was unintentionally removed in a previous update. It has been restored.
Description: When creating a custom report in the platform and the WHEN is configured to pull data for "all days available" we display the following message: This time range includes data from before and after Monetate's session update on May 1, 2016. All Analytics remain actionable.
Description: When creating an HTML Countdown Timer and select the next day as the end time for the countdown timer, the HTML renders an incorrect end time. For example, creating the HTML Countdown Timer today 2/9/2021 and selecting tomorrow same time for the end time, the HTML renders: There are 51 years, 7 weeks, 3 days, 22 hours, 47 minutes.
Description: Engine Insights tab is rendering an Internal 500 Error and not loading.
Description: The Trending strategy identifies products that have recently started converting at a higher rate. It captures the period of time in which products are becoming popular, before they become top sellers. - This algorithm looks at the sales volume (quantity of items sold) over the past 7 days (short period) compared to the 30 days (long period) before that and returns results ordered by largest increase. This is available under Premium Models feature flag and clients that have access to advanced recommendations module will be able to use it.
Description: The Dynamic Test summary on the Dashboard currently includes text like this: Dynamically allocated 1.3M sessions to higher performing variants. However, we observed that the Druid query that calculates that number is not performant, so we updated the dashboard text in this summary: Dynamically allocated 1.3M sessions
Description: After a product catalog is updated and the data is ingested, it takes a few minutes before that data is made available for search in the UI. On the search tab, rather than showing a blank page during this time, we have added a message that indicates that the data will be available soon.