Moving to the cloud: a migration checklist for NZ businesses
A cloud migration is a change to how your business operates, not just where its servers run. Payroll deadlines, internet reliability, software contracts and staff access can all determine whether the move succeeds.
Start with the business outcome: replacing ageing equipment, enabling remote work, improving recovery or simplifying support. Then set measurable acceptance criteria, such as completing a payroll run, restoring a deleted file and accessing a shared application from each branch.
Work through the following decisions in order. For each stage, name a business owner who can approve the result. Your IT provider can manage the technical work, but business priorities and acceptable disruption need your sign-off.
1. Map dependencies and settle licensing before buying
Create an inventory of applications, servers, databases, file shares, devices and integrations. Record the owner, user count, current cost, renewal date and support status of each item. Include easily missed dependencies: scanners sending email, accounting exports, scheduled jobs and software tied to a server’s hardware identifier.
Decide what to retire, replace, retain or migrate. An old application may need upgrading before it can run on a cloud platform. Moving it unchanged can preserve its existing problems while adding hosting costs.
Ask each software vendor to confirm cloud hosting rights and support conditions in writing. Do not assume an existing licence transfers, or that a subscription includes every security and backup feature you need. Build a first-year budget that includes overlapping services during transition.
- Check user-based versus device-based licensing, shared computers, contractors and seasonal staff.
- Confirm database, operating system and remote-access licensing, including third-party hosting restrictions.
- Budget for migration labour, backups, monitoring, internet upgrades, data transfer charges and eventual data export.
- Record cancellation notice periods and keep essential existing services active until acceptance testing is complete.
2. Design identity and access before moving workloads
Choose the system that will control staff identities and how it will connect to existing directories. Decide whether local directory services must remain for legacy applications or devices. Resolve duplicate accounts and inconsistent usernames before synchronisation; otherwise, access problems can follow users into the new environment.
Require multi-factor authentication, prioritising phishing-resistant methods for administrators where supported. Separate administrator accounts from everyday email accounts, and use role-based access rather than making everyone an administrator to get through migration day.
Document how new starters receive access and how leavers lose it. Review shared accounts, service accounts and external users separately: automated jobs may depend on credentials that nobody remembers owning.
- Verify ownership of your domain, DNS service and cloud tenant; retain business-controlled administrative access.
- Set up and test tightly controlled emergency access accounts, with monitoring for their use.
- Confirm staff can enrol authentication methods before cutover, including people working remotely.
- Identify devices that need management, encryption or upgrades before they can access company data.
3. Agree on security, data location and recovery requirements
Classify the information you hold: ordinary business records, personal information, financial records and commercially sensitive material. Decide who may access each category and whether contracts or sector obligations impose additional controls. Confirm where primary data, backups and support access will be located; a selected hosting region may not cover every service component.
For personal information, assess your obligations under the Privacy Act 2020. Overseas cloud storage does not automatically breach the Act. Whether Information Privacy Principle 12 applies depends on the arrangement, including whether the provider acts solely as your agent or uses or discloses information for its own purposes. Get advice if the position is unclear.
Set recovery objectives in business terms: how much work could you afford to lose, and how long could each service be unavailable? Availability, replication and retention features are not automatically substitutes for backup. Specify backup coverage, retention, administrative separation and restore testing. Also agree who monitors alerts, handles incidents and contacts your business after hours.
4. Prepare the data and prove the migration method
Assign a business owner to each dataset. Remove duplicates and obsolete material only after checking retention obligations and legal holds. Map existing permissions to the destination, and check file names, path lengths, supported formats and application compatibility before transferring anything.
Estimate migration duration using a measured upload test and the chosen migration tool, not your internet plan’s advertised speed. Account for throttling, large numbers of small files and changes made while copying. For substantial datasets, plan an initial transfer followed by incremental synchronisation and a final write freeze.
Run a pilot with representative users and difficult workflows, not just a clean test account. Ask staff to complete real tasks and record the results.
- Test application integrations, printing, scanning, mobile access and access from every business location.
- Compare record counts, file counts and tool error reports; use checksums where practical.
- Validate permissions, key financial totals and a sample of important documents with their business owners.
- Test restoring migrated data and confirm the pilot meets your acceptance criteria before scheduling cutover.
5. Write the cutover and rollback runbook together
Choose a cutover window that avoids payroll, month-end and other critical deadlines. Write a timed runbook with named owners, dependencies, expected downtime and a clear go/no-go checkpoint. Include the final synchronisation, access changes, DNS updates where relevant and business validation. Check DNS settings beforehand; cached records can make service changes appear uneven.
Prepare rollback before approving cutover. State the conditions that trigger it, the latest safe decision time and who has authority to act. Preserve the old environment in a usable state, with verified backups and documented steps to restore access.
Rollback becomes harder once staff create records in the new system. Decide how you will capture and reconcile those changes; simply switching back can lose orders or duplicate transactions. Some changes may require a forward fix rather than reversal, so identify those points explicitly.
- Send staff a plain-language notice covering downtime, sign-in changes and the support contact.
- Freeze changes where required and confirm the final synchronisation has completed successfully.
- Obtain business-owner approval for critical workflows before opening access to everyone.
- Keep a decision log showing test results, faults and the reason for proceeding or rolling back.
6. Stabilise, hand over and retire the old environment
Reserve support capacity for the first working days after migration, including your earliest and latest shifts. Track failed logins, synchronisation errors, application performance and support requests. Compare actual cloud consumption with your budget, and assign an owner to investigate unexpected charges.
Confirm the operating model with your IT provider. Define what the help desk, network operations centre and security operations centre each cover, including escalation and after-hours response. Check that monitoring is actually connected to the new services rather than assuming existing coverage transfers.
Close the project only after agreed acceptance tests pass, backups have been restored successfully and operating documentation is handed over. Then revoke temporary access, remove migration tools and retire old systems according to an approved retention and disposal plan. Cancel unused subscriptions only when dependencies are cleared. Your final checklist should leave the business with clear ownership, recoverable data and a tested support path—not just a completed transfer.
Want this handled for you?
CorTech provides 24/7 monitoring, security operations and help desk support for businesses across Auckland and New Zealand.