Three decisions before a cybersecurity audit
Define a cybersecurity audit's question, criteria, evidence, independence and follow-up before arranging access to systems or relying on its report.

A cybersecurity audit needs an agreed question, criteria and scope. It can examine evidence of how controls operate, but it can’t promise that an organization will never have an incident.
Before arranging the work, be clear about what you’re buying. A vulnerability scan looks for known weaknesses automatically, a penetration test tries to exploit them, and an audit checks evidence against agreed criteria. They support one another, but each produces something different.
1. What will the audit assess?
Identify the systems, information and business processes in scope. Record exclusions and the period the evidence should cover.
Agree on the criteria: an applicable contractual requirement, policy or selected control framework. Don’t assume that a supplier’s standard package matches the purpose of your audit.
The Cyber Centre’s baseline controls can help identify practical security areas for a smaller organization. Where a specific assurance or certification is required, confirm that requirement and the auditor’s qualifications separately.
Include supported software, access management, recovery and staff procedures. Software past its end of support is a finding in itself, but a newer product isn’t automatically an effective control.
2. How will evidence be collected safely?
Agree on the records, interviews and tests needed to answer the question. Include who gives the auditor access, how sensitive information is handled and which live-service activities are permitted.
NIST’s control-assessment methodology describes procedures that can be tailored to a defined assessment. It’s a US technical reference, useful for method. Canadian businesses aren’t required to follow it.
Confirm the required independence of the reviewer. If your IT provider would audit controls it also runs, ask how it keeps the review objective. A client or insurer may require an independent auditor.
Prepare an inventory and relevant records with their owners. Identify missing logs or incomplete history before the work starts, since those gaps limit what can be concluded.
Set rules for any disruptive test and a route to stop it. Put any permission to test live systems in writing, with its limits.
3. What will happen to the findings?
Agree on a report that states the scope, method, evidence, findings and limitations. It should separate confirmed gaps from items that couldn’t be assessed.
Have the business owner and technical owner review the practical impact of each finding. Assign corrective work, who signs off each fix and a follow-up date.
Leave the original findings as they were. When a control is fixed, verify the fix and record the new result separately.
Include staff preparation where a revised procedure affects ordinary work. Give people usable instructions and a way to report problems. A signed policy acknowledgment on its own changes little.
Decide how the report will be stored, who may share it and when the next review is due.


