Start with the business question, not the backup dashboard
If an important shared folder disappeared tomorrow, who would restore it and how would you know it was complete? That is the useful question for a small business owner. A green status indicator says a backup job completed; it does not tell you whether the right data was included or whether staff can use it after recovery.
This checklist is a recovery exercise, not a recommendation to delete production data. Have your Microsoft 365 administrator or IT provider approve the test, choose a quiet period and use a dedicated test account or folder. Never test ransomware or deliberately disable protection on your live systems.
1. Write down what is protected
List the workloads your business relies on: Exchange Online email, OneDrive files, SharePoint sites and the information staff access through Teams. Do not assume that a product covering one workload covers everything else. Ask for the actual protected account and site inventory, not just a licence count.
Microsoft 365 includes recovery and retention features, and Microsoft also offers Microsoft 365 Backup. Those are not interchangeable promises. Your available recovery options depend on the workload, configuration, licences and backup product. Establish what is enabled in your own environment before buying another service.
- Identify shared mailboxes, shared sites and new staff accounts that may fall outside automatic enrolment.
- Record the oldest recoverable point and how frequently each workload is protected.
- Check who can change retention, delete backups and authorise a restore.
- Confirm what happens when an account is removed or a backup subscription ends.
2. Agree what successful recovery means
Choose a recovery point objective: how much recent work could the business afford to lose? Then choose a recovery time objective: how long could the affected team work without the data? These are business decisions, not figures to copy from a vendor brochure.
For example, your finance team may need a shared workbook sooner than an old project archive. Record your own priorities and compare them with the service you actually have. Restoring one small file does not establish the recovery time for an entire site or mailbox.
3. Prepare a small, safe test set
Use invented sample information, not customer records, employee details or financial data. Where the product supports it, restore into a separate test location rather than overwriting live content. If it only supports an in-place restore, ask your administrator to design a non-disruptive exercise.
- Create a test email with a harmless attachment and record its subject, sender and attachment name.
- Create a test document, save a second version and note the difference between versions.
- Create a folder with a few sample files and known access permissions.
- Wait until your administrator confirms the test data exists in a recoverable backup point.
- Record the test scope, approver and intended restore destination before making any changes.
4. Recover the data and check it as a user
Have the authorised person perform the recovery through the documented process. Note when the request started, when data became available and when a staff member confirmed it was usable. Include the time spent obtaining access or approval; those delays matter during a real incident.
Open the recovered attachment and document. Compare the expected version and folder contents. Check that the intended user can access the recovery location and that an unauthorised user cannot. Restored data in the wrong place, with the wrong access, is not a successful result.
- Was the correct recovery point available?
- Were attachments and file contents intact?
- Were the expected files and versions recovered?
- Were permissions checked rather than assumed?
- Did recovery meet the business target for this specific test?
5. Distinguish native recovery from backup recovery
A recycle-bin recovery proves that the recycle bin worked for that item. A version-history recovery proves that the required version was still available. Neither, on its own, proves that the backup service could recover the same data.
Write down which mechanism you used. If you are testing a separate backup product, recover through that product. Also ask how recovery would be authorised if the usual administrator account were unavailable. Keep emergency access procedures secure and separate from ordinary staff documentation.
6. Keep evidence and close the gaps
Keep a short record containing the test date, workload, recovery point, result, elapsed time and person responsible for any follow-up. Store evidence securely; screenshots of real mailboxes or customer files can expose information unnecessarily.
Assign an owner and due date to each gap, then repeat the failed part after it is corrected. Repeat your drill after significant changes to accounts, sites, backup configuration or providers, and on a regular schedule chosen for your business risk. Testing helps validate a recovery plan; it is not a guarantee that every future incident will be recoverable.
Questions to send your IT provider
Use the answers to agree a realistic recovery process rather than buying on storage capacity alone. For New Zealand businesses, also ask where backup data is stored, who can access it and whether that fits your privacy and contractual obligations. Do not assume every service keeps data in New Zealand.
- Which Microsoft 365 workloads, accounts and sites are protected today?
- Can you demonstrate a restore from our configured backup, using safe test data?
- What is excluded, and how are new accounts added to protection?
- Who monitors failed jobs and who owns the recovery request?
- How would we regain access if our normal administrator could not sign in?
- What evidence will we receive, and what follow-up work is outside our agreement?
Sources and related guidance
Want this handled for you?
CorTech provides 24/7 monitoring, security operations and help desk support for businesses across Auckland and New Zealand.