You are correct, the data source is vCenter for both Veeam ONE and Veeam MP.
However Veeam ONE has a different approach, and it scans the vCenter much more often, getting performance and properties for objects every 20 seconds.
This scanning by Veeam ONE allows them to make faster mapping of event-to-property-to-user. So they can make a pretty safe assumption in their report, for which user changed an object.
Veeam MP has different methods to collect vCenter data. For Performance data and Property data, MP scans vCenter every 5 minutes. And then there is latency when injecting property data into SCOM - this is just a factor of how SCOM works.
The MP approach is more scalable - this is the only way to support environments with 1,000s or even 10,000s of VMs integrated into SCOM.
So - we can report on the changing properties of any vSphere object, we have this report out-of-box. And we can report on which users are changing objects, we capture all those vCenter events.
But we cannot safely make assumptions about which user changed which object and exactly when.
hope that answers your question!