DIVIDE yields incorrect results with numerator and denomenator in separate tables

[PeakBizOpt]

I have the below data model that utilizes the one-to-many relationship of the field 'Date" in red. I need the blue square to be the numerator and the green square to be the denominator. It's also worth noting that both fact tables have five entries for each date (one entry for each building). I'm using the following DAX: 

Avg Allocated Utilization % =

    DIVIDE(

        SUM('Utilization Report'[Average Person Count]),

        SUM('Capacity and Allocation'[Allocated]),

    0

    )

Ultimately I need to show this and three other measures that are the same (different data such as capacity figures instead of allocation figures) in a clustered column chart. The chart is separated by the 'Building' field as shown below (I'm clipping off the building names at the bottom purposely.) My expectation is that this will show the entire aggregation of data, then I can use date filters such as month and/or day of the week to filter down to targeted data as needed.

My problem is that these values are incorrect by quite a bit. The Peaks should be showing anywhere from 50%-80% and the Averages should be 20%-40%. 

The source data is clean and I've tried many different angles for the DAX with no luck. Thank you in advance for the guidance!

[lbendlin · Accepted solution]

Your data model is missing the building dimension.