What’s new in 16.10
The RichRelevance 16.10 release on August 4, 2016, includes the following client-facing updates:
- Advanced Merchandising Rules can now be authored to target a cNet Attribute based on equality/inequality.
- Dashboard - 1) Updates to Browse Boosting pages for the Find initial Beta and 2) Initial updates to add Region data to the Strategy Rules page.
- Discover - Initial updates for the Find Beta processing.
- Recommend - Version 1.2 of p13n.js will send added to cart items.
- Science - 1) Work continues to complete on the new machine-learning based King of the Hill (KOTH), 2) Initial steps to transition to the DAPS system (new system using new methods to create models/strategies data) and 3) Improvements to Engage optimization.
- Advanced Merchandising Rules can now be authored to target a cNet Attribute based on equality/inequality, existence of the attribute or regular expression rules.
- Browse Boost UI Changes for Discover for the Find initial Beta.
- Adding notch value on the Boost Amount when selecting a particular level
- "Importance" value being displayed on the rules list page
- When selecting an attribute we will now provide a dropdown of values to select
- Initial updates to add Regions context to Strategy Rules Page. These updates will be finalized in a future release
- Show Regions that the rule is enabled in
- Allow Region as a context when creating/editing a rule
- Initial updates for the Phase 1 deployment of new Find functionality. This is a limited Internal Beta.
- Sites instrumented with version 1.2 of p13n would not send the atcid parameter when the instrumentation correctly set the values, causing issues when trying to filter cart items from cart page recommendations.
- Updates to KOTH (King of the Hill) processing to improve the quality of product recommendations.
- Sophisticated machine learning approach to optimizing strategy selection, now includes data on how much we know about a given product as part of the "context" for decision making
- DAPS is a new system to improve timelines to generate data for various strategies (ViewedPurchased, ...)
- One advantage is processing of data via map reduce jobs.
- Small improvements to Engage optimization.