14.02 Release Notes (01/21/2014)
Regions in Merchandising Rules
Requirement: Regions must be passed via feeds.
Region context (Where section) has been added to recommendation restrictions and boosting rules. Example: Apply a rule if the shopper browsing is in the region Germany.
The following applies to sites that have multiple regions:
The Where sections include a new tab called Region. This tab has a region drop-down menu that contains regions of the current site. Click Add to select a region. Users can also exclude the selected regions.
New Strategy Page Updates
- Stats checkbox added
- Warning icon added to inform users that a particular (enabled) strategy is not eligible to run unless it has a strategy message set
- Filter searches by strategy name, strategy family, strategy message, and alt message
Universal Customer Mapping in RichRecs and PCS
Important note: This feature augments RichRecs and PCS. A DataMesh license is not required to use this feature.
The universal customer mapping feature in RichRecs and PCS allows customers to improve the quality of personalization and increase ROI by mapping previously disparate online user identities together into a single universal customer identity. This allows RichRelevance to account for a consumer’s views and purchases across mobile devices and multiple computers when generating personalized content in RichRecs and PCS.
The customer FTPs user ID mapping files to the RichRelevance platform. The customer can send updated user ID mapping files at any frequency. New files are total replacements, not updates. Files are expected to be 10-15 million rows in size.
For the initial version, we support CSV files with the following required fields. The column headers should be provided in the file:
|User ID||The ID currently instrumented as the RichRelevance external user ID in Recs and PCS|
|Universal Customer ID||The universal ID used to map multiple user IDs to a single identity|
- RichRelevance automatically processes the file within minutes of its transfer to the platform and publishes the user ID mapping data as a model.
- RichRecs returns improved product recommendations by consolidating previously disparate views and purchases.
- Recently viewed products – When recommendations are requested, the user’s recently viewed product history will be constructed in real-time by combining the current user ID’s view history with all mapped user IDs’ view histories
- Recently purchased products – When recommendations are requested, the user’s recently purchased product history will be constructed in real-time by combining the current user ID’s purchase history with all mapped user IDs’ purchase histories
- The following strategies which take recently viewed products and/or recently purchased products as input will now return recommendations that incorporate the online behavioral data mapped to a single customer:
- PCS returns improved sort order by consolidating previously disparate views and purchases. This is done in the same method as RichRecs above, except that PCS only has a single algorithm/strategy for producing results.v