Propensity score matching is the technique used in A/B testing analysis. This matching consists of assigning to each test record in the treated group a valid, comparable record in the control group. The matching algorithm relies on propensity scores. It is therefore important to ensure that the distribution of propensity scores for the treated group shows a good overlap with the propensity score distribution of the control group. The closer they are the better the matching and the more reliable the average treatment estimates.
Every propensity score bin that has treated group records should also have records from the control group. If not, you may need to include additional independent variables to improve the quality of your propensity scoring.