There is a scenario where:
The company group names are the same (but arent the same business/organisation), eg:
Schools like "St Johns" or pubs like "The Red Lion".
In the data these have the same company name, but different addresses. So, the data importer is creating one company group with multiple subsidiary companies, i.e. one group is created as "The Red Lion" and is given multiple subsidiaries with different addresses, but in this scenario it would be better to create separate singe-companies.
Unfortunately, at the moment, we can't see any obvious way to distinguish between when this should happen (Subsidiaries of Tesco Stores, Texaco Garages, or indeed a pub chains like Weatherspoon's, nor a collection of private schools owned by the same group) vs when it shouldn't (several entirely independent pubs/schools with the same name).
To make this work both way's round, we'd have to figure out some rules/logic to identify when the two different scenarios should or shouldn't apply.
There are new options in the importer to allow the importer to "ignore potential matches and create duplicate Companies"