Hello Brian , interesting geometries there! Just to be on the same page, I believe the purpose of these simulations is to assess the drag increment due to positioning of the bikes, is that correct?
I would recommend you to create an additional region refinement box around the bikes. This should give you more accurate results since the bikes are already in the wake region of the car where the flow is recirculating and highly unsteady. I would also go with a higher fineness value for the global mesh settings, cause it seems like the elements near the rear far field are relatively big.
Regarding the result control, allowing more iterations to each simulation may be a good idea. As I can see in the result control, aerodynamic coefficients are still converging. Drag, lift coefficients are calculated based on the reference values you provide so I would also create direct Force and Moment monitors to observe the results. Pressure drag is the important one in your case so creating a probe point behind the bikes and observing the convergence of flow quantities may also give you some insights.
Splitting the Car and Bike faces will also allow you to understand the contribution of each component into the drag coefficient and interference drag (Car-Bike interaction) as well.
Here is a knowledge base article that may help you to check the convergence of your simulation:
Hope these are helpful,
Kaan.