Migrating a workload from one cloud to another sounds straightforward until you're in it. Then it touches infrastructure, identity, data, operations, and every team that depends on the workload. The complexity isn't usually in the technology. It's in the approach.
Thomas Maurer and Raffaele Garofalo get into this in a new episode of the Azure Essentials Show. Raffaele, a Partner Solution Architect at Microsoft, has seen these migrations up close, and he walks through the three lessons that separate smooth migrations from the ones that stall or fail outright.
I wanted to highlight this episode because these lessons match what we keep seeing in the field. If you're planning a migration (or stuck in one that's not going well), this is a good place to start.
The traps teams fall into
Before getting to the lessons, Thomas and Raffaele spend time on why migrations go sideways. Three patterns come up:
Overengineering. Teams try to rebuild and optimize the workload as part of the migration instead of moving what already works. They want to modernize everything at once, and that ambition creates scope creep that buries the project.
Misaligned stakeholders. Migration projects involve users, operators, developers, compliance teams, and business stakeholders. When those groups aren't aligned before work begins, scope creep is almost inevitable. And if the project stalls on complex workloads instead of showing early wins, leadership loses confidence. That loss of momentum can threaten the whole initiative.
The "flip the switch" myth. There's a belief that you can shut down one cloud environment and turn on another overnight. In practice, that risks outages, data loss, and situations where rollback is no longer possible.
Sound familiar? Here's what to do instead.
Lesson 1: Get stakeholder alignment, and keep it
Alignment isn't a one-time kickoff meeting. Microsoft recommends starting every migration with a readiness assessment scored across 10 dimensions, followed by a workshop where all stakeholders share their requirements, constraints, and timelines.
One thing Raffaele emphasizes: the team that runs the workload today should be the team that drives the migration. Outsourcing it to external talent risks unexpected discoveries late in the process and less ownership. Partners can help you plan, but your workload team should execute.
Alignment doesn't end once you start, either. Formal stakeholder approval is required before any AWS resource is decommissioned, documented with clear success criteria. That prevents shutting down something that's still needed.
Lesson 2: Migrate like-for-like, optimize later
This one surprises people. Why go through a migration if you're not going to modernize?
Because modernizing during migration adds scope, increases complexity, and erodes confidence in the workload. The recommended sequence: migrate first, stabilize, then optimize from a position of stability.
In practice, that means matching AWS services to their closest Azure equivalents and not switching platforms during the move. You don't want to change your container orchestration tool at the same time you're changing your cloud provider.
Get the easy wins out the door. Build momentum with your stakeholders. Tackle the complex workloads after you've proven the foundation is solid.
Lesson 3: Use a blue-green cutover to preserve rollback
Blue is your current AWS environment. Green is your new Azure environment. Both run in parallel throughout the migration, so you can shift traffic back to AWS if something goes wrong.
Compare that to a big-bang cutover: high downtime, high risk, and hard rollback. Blue-green gives you low downtime, low risk, and easy rollback. The trade-off is cost, since you're paying for two environments during the transition.
Raf addresses the pushback head-on: running parallel environments is expensive, and stakeholders will question it. But for most teams, the extra cost is worth it because blue-green reduces risk and operational burden. The parallel period isn't indefinite. The old environment stays available until the new one has been verified, signed off on, and proven stable.
The key is defining upfront what "stable" looks like and what triggers a rollback. That clarity protects both the team and the business.
Watch the episode
Watch: Tips for Migrating a Workload to Azure, Azure Essentials Show
Thomas and Raffaele cover all three lessons in about six minutes. Worth a watch if you're evaluating or planning a migration.
Go deeper: the full article series
The video covers the highlights, but there's a complete article series on Microsoft Learn that walks through the full five-phase migration process (Plan, Prepare, Execute, Evaluate, Decommission) with checklists for each phase.
Migrate a workload from Amazon Web Services (AWS) to Azure
If you're just starting to assess or you're mid-migration and need to course-correct, the series gives you a structured path forward.