Cloud Migration Strategy: A Practical Plan for Growing Teams
Cloud Migration Strategy: A Practical Plan for Growing Teams

A cloud migration strategy is a decision made workload by workload: which systems move as they are, which get adjusted, which get replaced and which are switched off. The teams that finish on budget start with an inventory and a business case before they pick a cloud, then move in small waves. If you want an outside engineering team to run that assessment with you, our Discovery Audit is built for it. Below is the full method, written for small businesses, growing teams, healthcare providers and enterprises with tangled systems.
TL;DR
- A strategy is one decision per application, chosen from the “7 Rs”: retire, retain, rehost, relocate, repurchase, replatform, refactor.
- Work in four phases: assess, plan, migrate, optimize. Each phase repeats as you learn.
- Move in waves, starting with low-risk, high-value workloads, and define success metrics before each wave.
- The main risks are hidden dependencies, security gaps, downtime and cost surprises. Dependency maps and staging tests reduce all four.
- Healthcare providers need a business associate agreement with the cloud provider and a documented risk analysis before any patient data moves.
Table of Contents
- What a cloud migration strategy actually decides
- The 7 Rs, in plain language
- How to choose an approach for each workload
- The four phases of a cloud migration
- Writing the plan: what goes in the document
- Controlling cost on a smaller budget
- Cloud migration risks and how to reduce them
- Healthcare providers: protect patient data first
- Getting a plan you can act on
- FAQ
- Sources
What a cloud migration strategy actually decides
AWS describes a cloud migration strategy as a business case, a scope and an approach for moving from current IT resources to a new cloud environment. The practical meaning is simpler. You list what you run today, decide what each item should become, and set the order.
Two things follow from that:
- There is no single “best” strategy for a company. Microsoft’s Cloud Adoption Framework says that once you have a clear inventory, your plan must decide what to do with each workload. One company usually ends up using four or five of the approaches below.
- The strategy is a business document before it is a technical one. It states why you are moving (cost, scale, resilience, compliance, speed), what “done” looks like, and who owns each system.
The 7 Rs, in plain language
AWS’s prescriptive guidance lists seven approaches, known as the 7 Rs of migration: Retire, Retain, Rehost, Relocate, Repurchase, Replatform and Refactor (or re-architect). Microsoft’s list uses slightly different labels, including Rebuild and Replace, but the idea is the same.
| Approach | What it means | Typical fit |
|---|---|---|
| Retire | Switch the system off | Duplicate or unused tools found during inventory |
| Retain | Leave it where it is for now | Systems with a hard dependency, a contract, or no near-term benefit from moving |
| Rehost (“lift and shift”) | Move the application as is onto cloud servers | Fast moves, data center exits, stable apps |
| Relocate | Move virtual machines to a cloud service built for them without changing them | Teams already running virtualized environments |
| Replatform | Move with small changes, such as a managed database | Apps where a few tweaks cut maintenance |
| Repurchase | Replace with a SaaS product | Commodity functions such as email, HR or CRM |
| Refactor / re-architect | Redesign the application for the cloud | Core products that must scale or change fast |
One detail from the same AWS guidance matters for planning. For large migrations the common strategies are rehost, replatform, relocate and retire, and refactor is not recommended in the middle of a large migration because it modernizes the application while you are also moving it. Modernizing is better done as its own project afterward.
How to choose an approach for each workload
AWS notes that the right choice depends on the requirements of each resource, your IT environment and the business value you want. In practice, run each application through five questions:
- Does the business still need it? If not, retire it. This is the cheapest migration you will ever do.
- Is it a commodity function? If a SaaS product does the job, repurchase.
- Can it move unchanged without hurting performance or cost? Rehost or relocate.
- Would a small change deliver a large saving? Swapping a self-managed database for a managed one is the classic replatform.
- Is it the system that differentiates your business? Then it deserves a refactor, scheduled after the move so the two risks do not stack.
A worked example: a 40-person distribution company runs an old inventory app, an email server, a file server, a customer portal and an accounting package. The email server is repurchased as hosted email. The file server is rehosted. The accounting package is retained until its license renews. The inventory app is replatformed onto a managed database. The customer portal, which drives revenue, is queued for refactoring in a later wave. Five systems, five different decisions, and no single “strategy” label covers them.
Legacy systems are where this gets hardest, because the people who understood them have often left. If that sounds familiar, our piece on discovery-first legacy system modernization covers how to document what a system really does before deciding its fate.
The four phases of a cloud migration
Cloud providers each slice the process a little differently. Microsoft describes six stages: define strategy, plan, ready, migrate, govern and manage. Google describes four: assess, plan, migrate and optimize. The shared shape is the same, and four phases are easiest for a smaller team to hold in their head.
Phase 1: Assess
Build the inventory of servers, applications and dependencies, then build a data-driven business case. AWS’s guide for small and midsize businesses puts inventory and the business case first, and Google adds that this is where you identify your team and decide which applications move first.
For each workload, record the facts Microsoft’s migration plan template asks for: owner, business value, criticality, data sensitivity, compliance requirements, maintenance windows, business freeze periods and success metrics, as listed in its workload table. A freeze period matters more than it sounds: a retailer that migrates its order system in November has chosen its own outage.
Phase 2: Plan (mobilize)
This is where the strategy turns into a schedule. AWS’s SMB guide describes designing a landing zone, meaning the accounts, identity, logging and controls your workloads will live inside, defining success metrics and grouping applications into waves. It suggests treating each wave as a mini-project with its own metrics, such as cutover time, error budget and cost change.
Microsoft’s modernization guidance gives the sequencing rule: start with low-risk, high-value changes, order the rest by value and dependencies, define success criteria for every phase and adapt the plan from the results.
Phase 3: Migrate
Move in waves, and keep a way back. Google’s checklist recommends a phased, agile approach that lets you revert to the on-premises setup if necessary. Before each cutover, the people who use the system should confirm it works. Structured user acceptance testing, with clear gates before go-live, is covered in this UAT rollout guide for QA teams. Run the first wave on something a bad day will not hurt, such as an internal tool, so the team learns the process cheaply.
Phase 4: Optimize and operate
Migration is not finished at cutover. AWS describes its phases as iterative rather than discrete, so lessons from wave one change how wave two is planned. After cutover you right-size servers, set cost controls, review access and decide which retained or rehosted systems are ready to be modernized. Disaster recovery belongs here too. Our guide to disaster recovery in the cloud explains recovery time and recovery point objectives and how to set them.
Writing the plan: what goes in the document
A plan that survives contact with reality fits on a few pages and contains:
- Goals that can be measured. Tie technical targets (availability, performance) to business targets (cost reduction, operational efficiency), the way Airbyte’s planning guide suggests. “Move to the cloud” is not a goal.
- The workload register. One row per application with the fields above, and the chosen R.
- The wave schedule. Which systems move together, in what order, with freeze periods marked.
- Success metrics per wave. Cutover time, error budget, cost delta.
- Roles. A named business owner for every workload, not only a technical owner.
- Rollback criteria. The specific conditions under which you revert, agreed before cutover day.
- Budget. One-time migration cost, running cost and the skills you must hire or borrow.
Controlling cost on a smaller budget
Cost is the main worry for small teams, and it is manageable when you plan for it from day one.
- Expect savings only if you right-size. AWS reports that SMB customers see an average of 31% cost savings after migrating, especially when they right-size and automate their infrastructure. Moving an oversized server unchanged moves the waste with it.
- Tag resources and set budget alerts immediately. The same AWS guide recommends doing both on day one. Without tags you cannot tell which project is spending the money.
- Keep optimizing. The FinOps Foundation’s State of FinOps 2026 report says workload optimization and waste reduction remain the top priority for cloud cost teams. Cost control is a routine, not a one-time cleanup.
- Count the exit and transfer costs. Data transfer (egress) fees are a known surprise, and TechTarget reports some organizations spend up to 27% of their cloud budget on them.
For a sense of how project budgets scale once software work is involved, our custom software cost guide lists published ranges: $10,000 to $50,000 for an MVP or simple app and $50,000 to $150,000 for a mid-complexity platform. Smaller organizations planning the whole journey can also read our digital transformation roadmap for smaller organizations.
Cloud migration risks and how to reduce them
The risks are well known, and each has a specific countermeasure. TechTarget’s overview names egress fees, security and compliance gaps, hidden application dependencies, and disruptions such as downtime and incomplete data transfers, with thorough assessment, dependency maps, staging tests, encryption, access controls and ongoing audits as the mitigations.
| Risk | What it looks like | What reduces it |
|---|---|---|
| Hidden dependencies | An app breaks because it quietly relied on a server you moved | Dependency map in the assess phase; move linked systems in the same wave |
| Downtime | Cutover runs long, customers notice | Maintenance windows, freeze periods, rollback criteria, a pilot wave |
| Data loss or incomplete transfer | Records missing after the move | Staged test migrations, reconciliation checks before cutover |
| Security gaps | Open storage, over-broad access | Landing zone with identity, logging and controls designed first; encryption; audits |
| Cost overrun | Bill grows after go-live | Right-sizing, tagging, budget alerts, egress estimates in the business case |
| Integration problems | Cloud systems cannot talk to what stayed behind | Integration design per workload, including retained systems |
The last row is why integration deserves its own line in the plan. Retained systems do not disappear, and the connections between old and new are where many projects stall.
Healthcare providers: protect patient data first

If you handle patient information, the strategy has a compliance layer. HHS guidance says that when a covered entity uses a cloud provider to create, receive, maintain or transmit electronic protected health information, the provider is a HIPAA business associate and the two parties must sign a HIPAA-compliant business associate agreement. Both covered entities and business associates must conduct risk analyses of threats and vulnerabilities to that data, and a business associate must notify the covered entity of breaches of unsecured PHI.
In plan terms: mark data sensitivity and compliance requirements for every workload in the assess phase, sign the agreement before any patient data moves, and put the risk analysis on the schedule ahead of wave one. We go through this in more detail in our healthcare cloud migration guide and our HIPAA-compliant software guide. Questions about your specific obligations belong with your compliance officer or healthcare attorney.
How we approach migration
Most failed migrations we hear about went wrong before a single server moved. Nobody wrote down what the systems did, so the plan was built on guesses. We start with discovery for that reason: a transparent audit of what you run, what it costs, and what each system should become.
Our engineers come from big tech backgrounds, and we bring that rigor to businesses of every size, including the HIPAA and SOC 2 compliance that healthcare and finance teams need. Cloud architecture is only part of the answer. The harder work is the integration between old and new systems and the workflows around them, and that is where a tailored approach pays off.
Our advice to a small team is to resist the urge to migrate everything at once. Pick the right R for each system, move the low-risk workloads first, and let the results shape the next wave.
Cameron
Getting a plan you can act on
The assess phase is where most of the strategy gets decided, and it is where outside experience saves the most time. Our Discovery Audit is $999 one-off, and it is credited toward a build contract if you continue with us. It gives you a documented picture of your systems and a recommended path before you commit to a larger budget. Our Cloud & DevOps and System Integration work covers what comes after.

If you are ready to scope it, start from the Discovery Audit page, or see everything we build on the CoreWorx site.
FAQ
What are the 7 Rs of cloud migration?
They are retire, retain, rehost, relocate, repurchase, replatform and refactor (re-architect). Each describes what you do with one application: switch it off, leave it, move it as is, move it with small changes, replace it with SaaS, or redesign it. AWS’s guidance lists all seven.
How long does a cloud migration take?
It depends on the number of workloads, their dependencies and your compliance needs, so no honest single figure fits everyone. Plan in waves, each a mini-project with its own metrics, and the assessment will give you a realistic schedule. A pilot wave on a low-risk system shows how fast your team really moves.
What is the difference between a cloud migration strategy and a cloud migration plan?
The strategy says what happens to each workload and why. The plan turns those decisions into an ordered schedule with owners, budget, success metrics and rollback criteria. You need the strategy first, because the sequence of waves depends on which approach each system gets.
Should a small business refactor everything for the cloud?
No. Most small teams get the best return by retiring unused systems, repurchasing commodity software as SaaS and moving stable applications with few changes. Refactoring is worth it for the systems that differentiate your business, and it is safer as a separate step after the move.
What should healthcare organizations do before moving patient data?
Sign a HIPAA-compliant business associate agreement with the cloud provider and complete a risk analysis covering threats to the data, as HHS guidance describes. Record data sensitivity for each workload during assessment. Your compliance officer or healthcare attorney should confirm your specific obligations.
Sources
- What is a Cloud Migration Strategy? (AWS)
- Select your cloud migration strategies (Microsoft Learn)
- About the migration strategies (AWS Prescriptive Guidance)
- Strategies for Accelerating Migration to AWS
- How to create a cloud migration plan for SMBs on any budget (AWS)
- Cloud adoption plan template for migration (Microsoft Learn)
- Plan your cloud modernization (Microsoft Learn)
- What is Cloud Migration (Microsoft Azure)
- Migration Guide and Checklist for SMBs (Google Cloud)
- How to Build a Cloud Migration Plan in 7 Easy Steps (Airbyte)
- 6 cloud migration challenges to prepare for and overcome (TechTarget)
- State of FinOps 2026 Report
- Guidance on HIPAA & Cloud Computing (HHS)
- 90 Day UAT Tool Rollout: Operational Gates for QA Teams