Cloud migration assessments give us a view across a wide range of environments, and the same mistakes appear with enough regularity that they are worth documenting. These are not errors made by careless teams — they are the natural results of rational short-term decisions that accumulate into expensive long-term problems.
First: skipping the workload dependency mapping phase. Skipping it means the migration discovers dependencies in production, under time pressure, in ways that are much more expensive to resolve than they would have been during discovery. Every migration we have rescued at the dependency phase has involved a legacy application with undocumented dependencies that blocked a cut-over.
Second: treating lift-and-shift as a final state rather than a starting point. Moving a workload unchanged to a cloud VM is a useful first step — it gets the environment off expiring hardware or out of an ending co-location contract. It is not a useful final state because it carries all the operational overhead of self-managed infrastructure without the elasticity or managed-service economics of cloud.
Third: not testing rollback. Every production cut-over plan includes a rollback option in theory. The plans that work under pressure are the ones where the rollback path has been tested under controlled conditions. We run a dry-run including the rollback path as a standard step in every migration project.
Fourth: migrating at the wrong layer. Some workloads are best migrated as virtual machines; others are best re-platformed to managed services or re-architected to cloud-native patterns. We see too many workloads re-architected to microservices when a straightforward VM migration would have served the business purpose in a fraction of the time.
Fifth: not migrating cost governance alongside the workload. A workload migrated without resource tagging, cost-attribution configuration, and budget alerts carries no cost visibility from day one. Establishing tagging policy and alert thresholds as a prerequisite of migration produces materially better cost outcomes in the first year.