Moving four years of email into a CRM when the history might be worthless

Take an operator with 22 units, three spreadsheets, and four years of Gmail, moving to a real CRM with two properties under contract and a system that has been little more than a filter labeled units. The import is the hard part. A dedupe pass on a contact export might take 1,340 records down to 890, and of those maybe 400 are real ongoing relationships. The rest are past applicants who never leased, vendors used once, and a wholesaler who sent 60 emails in a single year. Every junk record still carries a phone number, which means every one of them can end up in a sequence by accident. The email history is often worse. Four years of threads can be piped into a CRM and attached to contacts, but the attachments are usually the only part anyone actually references, and those are typically already organized in cloud folders by address. There are two reasonable paths. Import everything, accept that the database starts dirty, and clean as you go. Or set a cutoff date, tag active contacts by hand over a weekend, and leave the archive as a mailbox that gets searched but never written to again. The cutoff approach tends to hold up well in practice, with one caveat: if old records have occasionally been needed to settle a tenant dispute and found only by searching a remembered phrase rather than by any filing system, that search capability should be preserved regardless of which path is chosen, since an archive that already works by search does not need to be migrated to keep working.