The challenge
A farm advisory platform was moving its clients onto a new platform database. The move covered 42 client groups and 101 legal entities: about 4.29 million rows of history, from milk production to financials. Advisers and their clients rely on that history, so the migration had to be complete and accurate, and everyone needed to know exactly what had moved.
What we did
We started by defining exactly what would move. Each legal entity was assigned to a cohort, so the migration could be planned and checked entity by entity rather than as one large, opaque transfer. Working in cohorts also meant a problem with one entity could be dealt with on its own, without holding up the rest.
Before moving anything, we ran a collision analysis to find records that would clash with data already in the new database, such as duplicate identifiers or overlapping entities. Resolving those up front kept surprises out of the cutover.
We then wrote a full runbook: every step in order, how to check it worked, and what to do if it didn’t. The aim was a migration that could be rehearsed, repeated and verified.
The result
All 42 client groups and 101 legal entities were moved onto the new platform database following the runbook. Today that database holds about 11.9 million records across all its tables, from milk production to financials.
The runbook and cohort definitions remain useful beyond the cutover, giving the platform a documented, repeatable way to bring further data across.
Migrations like this are mostly about preparation. The data work matters, but the planning, the collision checks and the runbook are what make a large move predictable. If you’re facing a platform change of your own, our data platforms work covers migration from planning to cutover.