So the challenge is clear, but how do we solve for it? You might consider data blending, or taking the minimum quota figure instead of the sum. Both of these options have caveats, however. Data blending can perform poorly on larger data sets. And taking the minimum quota figure relies on having the salesperson dimension in the view.
This brings us to our solution: using a Level of Detail Expression.
This approach is scalable, dynamic, and reusable. Since the quota is always replicated across the salesperson dimension, we can fix the level of detail to the salesperson and always take the minimum quota:
This calculation will give the correct value, whether we are looking at the view by salesperson or by total quota.
While this is one example, the application of this simple LOD Expression is vast. At Tableau, we use this across many of our data sources not only to improve performance, but also to ensure that others don’t fall victim to using incorrect numbers in their own analysis. I recommend changing the name of the duplicated metric to “incorrect value” and naming the LOD version “correct value” to ensure people use the right one.