While true, I don't think this explains the disparity between regular CloudKit use and CloudKit Sharing use. For example, this snippet from Apple's CloudKit paper[1] particularly struck me: "We identified the top apps using CloudKit, based on their number of active users in the past month, and examined their use of private and public databases. We found that 20% and 49% of the apps use only the public or the private database, respectively, and 31% of apps use both databases (20 apps use the shared database)."

So few apps from their sample set were using CloudKit Sharing that they didn't even bother using a percentage! That's quite unusual for an Apple framework.

[1]: http://www.vldb.org/pvldb/vol11/p540-shraer.pdf