Moving to the cloud goes wrong in predictable ways: a migration scheduled in the middle of month-end, a critical system made dependent on a single internet line, a cutover with no rehearsed way back. None of these are technology failures — they are planning failures. This checklist is the sequence we follow with clients to avoid them.
Work through it in order. The local realities — connectivity, billing in dollars, where data lives — are built in, not bolted on.
1. Plan and assess
- List every system, application and data store you currently run.
- For each, record what breaks when it is unavailable (the business impact).
- Set a target RPO (how much data you can afford to lose) and RTO (how fast you must recover) per system.
- Decide what should move to the cloud, what stays on-premise, and what becomes hybrid.
- Confirm any data-residency or compliance constraints before choosing a provider or region.
- Estimate monthly cost in USD and convert to Pula at a conservative exchange rate; add headroom for growth.
2. Get the connectivity right first
- Confirm your primary internet line can carry the new cloud workload.
- Arrange a second internet line from a different provider for automatic failover.
- Test failover actually works before you depend on it.
- Check latency to the target cloud region for any latency-sensitive system.
3. Prepare security and access
- Enable multi-factor authentication (MFA) on every account, especially administrators.
- Define roles with least-privilege access — no blanket admin rights.
- Confirm encryption is enabled for data in transit and at rest.
- Set up monitoring and alerting for unusual activity.
- Document who has access to what, and how access is removed when someone leaves.
4. Protect your data before you move it
- Take a full, verified backup of everything being migrated.
- Confirm you can actually restore that backup (test it).
- Keep the source systems intact and available as a rollback until cutover is confirmed.
5. Migrate in the right order
- Start with low-risk services: email and file storage.
- Migrate, then validate data integrity before moving on.
- Move business-critical systems (accounting, ERP) only after the low-risk moves succeed.
- Schedule each cutover outside busy periods — never during month-end or a peak trading day.
- Rehearse the cutover steps so the live run holds no surprises.
6. Cut over and verify
- Run the cutover in the planned window.
- Verify each migrated system works end to end, with real user tasks.
- Confirm backups are running against the new cloud systems.
- Keep the old environment available until the new one has run cleanly for an agreed period.
7. Optimise and hand over
- Review actual cloud spend after the first full month and right-size what is over-provisioned.
- Switch off anything migrated but no longer used.
- Set a recurring review of cost, backups and security configuration.
- Document the new setup and who supports it.
- Schedule the first restore test for the new environment.
A note on doing this well
The copy-the-data part of a migration is the easy part. The value — and the risk — is in the planning around it: the connectivity groundwork, the rehearsed cutover, the tested rollback, the timing around your operations. If you would like a second pair of hands on any of that, a cloud assessment is the place to start, and our wider thinking is collected in the Cloud Computing pillar.
Frequently asked questions
How long does a cloud migration take?
It depends entirely on what you are moving. Email and file storage can move in days. A full ERP or line-of-business system, migrated properly and tested, is usually a matter of weeks because the planning, data validation and cutover rehearsal matter more than the copy itself. We always plan the cutover around your operational calendar — never in the middle of month-end.
Will we have downtime during migration?
A well-planned migration minimises downtime, often to a single scheduled window outside business hours. The key is preparation: validating data, rehearsing the cutover, and keeping the old system available to roll back to until the new one is confirmed working.
What is the most common cloud migration mistake?
Moving a critical system to the cloud before sorting out backup connectivity and a tested rollback plan. The second most common is migrating during a busy period. Both are entirely avoidable with planning, which is what this checklist is for.