You could import the excel file into a helper entity that mirrors the columns of the spreadsheet. Once you do that, you can build a microflow to parse through the Reference objects in Column B and create those objects if they don't already exist.
This can be done using the default excel import module without having to create a helper entity and custom microflow.
According to the FAQ documentation from the XLS Importer:
How do I map reference sets?
This depends on how they are represented in you Excel file. If each element of your reference set has its own row in your sheet, the straightforward mapping should do the right thing. Set the mapping type of the reference columns to “reference”. Set the key type of the columns that form the key values to “Yes, only for the associated object”. The mapping will now automatically make sure no duplicate objects are created. Example of this situation (User N – 0 Group):
Person (key for Person object) Group (key for Group reference)
If the reference set is represented by multiple column groups (object entries) on each Excel row, you can select the above mapping for each of the column groups. So, you should have multiple columns mapped to the same object reference and attributes. In this case you should make sure that the key attributes are in the first columns for each reference, otherwise the import will not be correct. Example of this situation (User N – 0 Group):
Person Group1 (key for Group) Group2 (key for Group) Group3 (key for Group)
John Users Admin Manager
Jack Users Manager