- There was an intermittent problem that showed inactive categories in the catalog browser. With a change in the background logic, this issue has been resolved.
- Layout issues on the Manual Recommendations page have been fixed.
There were several different issues affecting the use of CNET attributes in Advanced Merchandising rules. These issues have been resolved.
In order to make better use of offline data and have offline purchases weighted properly, some changes were made to the models used to sort top sellers.
With the release of Launch Without Rules, there was an issue that caused the logic to not consider any "recommendable=true" if the customer had more than fifty items marked as "recommendable=false," thus leading to a failure to produce effective results. The logic was modified to consider three things: it checks to see that the content is recommendable ("recommendable=true"), that the content is active (today > start_date AND today < end_date) and the SQL query is sorted to favor newly created content.
A new "Abandoned Browse" trigger has been added, allowing retailers to segment the users who viewed the products, but added nothing to cart and bought nothing. This new trigger can be configured by the retailer in a number of ways:
- Amount of time required for browse to be considered abandoned?
- Maximum lookback for browse abandonment - this determines how long a user can be eligible to receive an abandon browse email after the last item was browsed?
- A number of times
- More than X times;
- Less than X times;
- Exactly X times;
- Between X and Y times.
- Price of the product(s):
- is more than X
- is less than X
- Is between X and Y
The Runtime Annotation Dashboard (RAD), which is a dashboard available to RichRelevance customers, has been added.
Previously, when creating a Tag Rule and specifying more than one tag as the target, they could only be joined using "OR." With this release, both "AND" and "OR" can be used to join multiple tags.
A new added Search Keyword trigger now allows you to search for any search term.