I’ll reply to my own message in case it helps…
We followed the advice to spoof the data load domains - and are now paying temporary staff to ‘despoof’.
Since you don’t have access to the DB tables, you cannot script out the spoofing, nor are Hubspot able to.
They did confirm that the importer tool was the only ‘de-duplicator’ which was useful.
But we’re still in the position that after our final data load, rather than being super excited to use a brand new system, our first activity is a forced data cleanse!