My Piano Content widget shows duplicated recommendations that should not show up. How can I remove them?
Sometimes the same content is served at different URLs.
Answer: Add a unique cXenseParse:articleid <meta> tag to your HTML markup. Read more about Document parsing
My recommendations widget shows pages that should not show up. How can I remove them?
This can happen for content that is deleted or no longer relevant.
Answer: Read about excluding content.
My site requires SSL. How can I deliver content recommendations on my site?
Answer: When loading recommendations on SSL sites, loading differs from non-SSL requests.
-
cx.js is loaded from https://scdn.cxense.com/cx.js instead of http://cdn.cxense.com/cx.js.
-
Use the https:// prefix (instead of http://) for the API call to recs.cxense.com.
-
Serve images from https://dibtijsj75kwx.cloudfront.net/analytics/ImageService/ instead of http://analytics.cdn.cxense.com/analytics/ImageService.
I want to deliver recommendations from multiple sites. Is it possible?
Answer: You can request recommendations across multiple sites by listing site IDs in the "recs" nodes of your content settings object.
If any listed site has no content, the request will fail.
A more flexible option is "combined" content settings, with one branch per site.
The site ID of each recommendation appears in the result field "siteid".
How do I set up "dynamic" recommendations using the current API?
Answer: Create different content settings based on the current section, time of day, etc., using "conditional" content settings.
How to analyze clicks in Piano Insight?
In Piano Insight, page views from recommendation clicks include the custom parameter cx_source=cxrecs.
For an example with more detail see: "Tagging" recs clicks for Piano Insight
How to use Piano Audience segment ids for Piano Content configurations?
Integrate Piano Audience and Piano Content using the "conditional" branch feature, as described on this wiki page:
-
Content settings conditions >> "userSegment" (requires Piano Audience)
-
which belongs to the recommendation type "conditional"