Resolution

To match reporting obtained from Google Analytics, structure the data connection (the dimensions and measures imported) to the same granularity used in the Google Analytics report. 

• To get correct results for Sessions, all the Event-related dimensions need to be removed from the query.
• To get correct results for Unique Events all the Event-related dimensions must be included in the query.
• To get correct results for Unique Users, remove all the session or event-related dimensions from the query. Also exclude the "Date" dimension and query on "Year" or another dimension that would not split the data.

Cause

Fields such as Unique Users are calculated per the set of dimensions in the connection which causes discrepancies. 

Google Analytics is different from other supported data sources in a number of ways. One major difference is that Google Analytics provides Tableau with data that has already been aggregated to some degree. Because of this, you may see issues with non-additive aggregates such as unique visits. Although Tableau will use the numbers based on unique visits per day, it cannot differentiate between the same visitor who visits multiple days and a visitor who only visits once. 
 

Additional Information

Mirroring reports is often difficult in Tableau Desktop due to the fact, that all the information and level of information detail used to make reports in Google Analytics is not always externally available. The following tools provided by Google allow for deeper exploration of available data:

• Google Dimensions & Metrics Explorer
• Query Explorer

If the results are still not the same, import data from Google Analytics in a raw form and compare the results with the data imported from Google Analytics in Tableau Desktop. 

More information from Google on the Users metric:

• How users are identified for users metrics.
• Google Analytics