Data migration that keeps the history, not just the records

Longer than most plans allow, because the effort is in profiling and cleansing rather than in the transfer itself. The move might run in hours; agreeing what to do with twenty years of inconsistent records is what takes the time.

It gets a decision rather than a default. Options are usually an extension field, an archive that remains searchable, or an accepted loss — but it should be a choice someone made knowingly, not something discovered afterwards.

You can, but it is worth costing honestly. Licences, hosting and the security patching of a system nobody maintains often exceed the cost of migrating the history properly in the first place.

By agreeing the criteria in advance and reconciling against them — record counts, financial totals, and sampled checks on the awkward cases. Without criteria set beforehand, sign-off becomes a matter of nerve rather than evidence.

Sometimes, using parallel running and a synchronised switch, though it costs more and adds complexity. For many organisations a planned weekend outage is the cheaper and safer answer, and we will say which we think applies.

You roll back, which is only possible if the rollback has been tested rather than merely written down. That is why rehearsal runs include rehearsing the way back.