1 of 1 people found this helpful
Better think of GA as a 'cube' (and it actually is).
Every Measure value you've got from GA is an aggregate
calculated for each combination of the unique values
of the Dimensions taken.
And aggregates could be additive or not.
Distinct count (of Users for example) is non-additive.
The results calculated for each Day doesn't sum up nicely
at the Week / Month / Year level, since there are repeated Users.
So to get 'true' User counts from GA, one should do it separately
for each level of granularity (daily / weekly / monthly etc).
In Tableau it could lead to having different datasources for each.
Hope it could help understanding.
I am having a similar problem, only for me the unique users aren't even aggregating at the day level correctly. I'm getting values nearly twice what they should be.
The dimensions I am using are: Country, Date, Device Category, Page, Source/Medium.
Is some combination of these the problem?