1.Is there any mention of offsite data in any doc ? As far as I know, your data can offline and online/realtime. Advertising campaign can be offsite targeted i.e. re targeting users who have visited your website on other websites via an ad exchange. Typically the publisher will segment their audience into targetable groups based on behaviours the audience exhibits on site.
1. First Party Data in Audience Manager : The definition of first party data is a company's own data. Now it can be online (realtime or onsite data )and offline (onboarded or historical data) both.
AA and AAM
If you have AA and AAM both and you implemented both solutions so that AA can share real time data with AAM. Then AA variables will come to AAM as realtime AAM signals.
Apart from this real time data you can also get historical data from AA, by creating segments in AA using a report suite and sharing those segments via Marketing Cloud to all Marketing Cloud solutions. That would come as onboarded data.
This is one way of collecting real time data in AAM.
CRM data or Onboarding your own data:
You know your customers who sign in on your site/app and you want to create segments of those customers. You can onboard or upload a data file with CRM ID and their corresponding signals in every row of that file : abc123 "color"="blue"
Before uploading this data you should have traits created in AAM which use those signals. Also, the ID sync should be implemented (this may require some coding) in prior to link your CRM ID with AAM's device ID.