Engineer cost into the move.
Before the new bill arrives.

Migration cost optimization puts cost decisions into the move itself, so the new estate does not inherit the old bill. We record a dated baseline of current spend, retire unused reports and workloads before they are converted, size the target on the workloads that will actually run, and set a cost case beside each migration wave. You know what the new estate should cost before it is built.

This work happens during a move; Data Estate Optimization tunes an estate that already runs. Here the decisions are what to retire, how large the target should be and what each wave should cost. Once the target is live, the costed wave plan and baseline pass to the operating team, and later tuning of the running Databricks or Fabric bill belongs to Data Estate Optimization.

We size the target from the workload map, volume and serving needs, with every assumption written down. The cost of the move and the cost of the run are considered together, against a dated baseline of what the source costs today. After cutover, the Lakehouse Health & Cost Check reads the new Databricks or Fabric bill and benchmarks your own workloads, so the estimate is checked against real numbers.

It is scoped in the first working session and delivered at a fixed price, because its length depends on the size of the move. It can run on a planned move or on one already under way. You keep the dated baseline, the target sizing assumptions, the cost model and the costed wave plan, with the point in each wave where source spend retires.