Vulnerability management programmes rarely collapse in a single dramatic moment. They weaken through a series of small errors that accumulate. A scanner is left pointing at an outdated list of hosts. A finding is assigned to a team that does not own the system. An exception is approved and then forgotten. A fix is closed without anyone checking it. Each error seems minor on its own. Together they produce a programme that looks active and provides little assurance.
This guide describes the mistakes that appear most often, explains why they happen and suggests how to recognise and correct them. It is written as a diagnostic aid, so it can be used on an existing programme as well as during design.
Mistake one: scanning an inventory nobody maintains
The most common error is building a programme on an asset list that is out of date. The list may have been correct when it was created, but the estate has since changed. New cloud accounts have been opened, systems have been decommissioned without removal from the list, and third-party services have appeared without anyone recording them.
The result is a programme that reports on the past. Coverage figures look healthy because they measure the list, not the estate. The fix is to reconcile scan results against independent sources, such as network records, cloud account listings and configuration data, and to review the inventory on a regular cycle.
Mistake two: findings without owners
A finding that has no owner will not be fixed. It may be visible in a report, but visibility is not action. Many programmes send findings to a central team, which then forwards them to system owners who are unaware of the work or unable to act on it.
The correction is to link every asset to a named owner before scanning begins, and to route findings to that person automatically where possible. Owners should know what they are responsible for and what response is expected. An owner without authority to make the change should be identified, so that the request reaches the person who can.
Mistake three: treating severity as the only priority
Severity ratings are useful, but they are not a complete guide to priority. A high-severity finding on an isolated system with no external path may matter less than a medium-severity finding on an internet-facing service that holds sensitive data. Programmes that sort strictly by severity can spend effort on the wrong items while more important exposures remain open.
Prioritisation should consider exposure, asset criticality and credible evidence of exploitation. The rules should be written down and reviewed, and the people applying them should know when they can depart from the default.
Mistake four: uncontrolled exceptions
Accepting risk is a legitimate decision, but it must be controlled. The mistake is to approve exceptions without recording a rationale, an owner or a review date. Those exceptions then become permanent. Over time, the register of accepted risks grows, and no one remembers why each item was accepted.
Each exception should have a named approver with the authority to accept that level of risk, a clear reason, any compensating control and an expiry date. Expired exceptions should return to the queue for review. The register should be examined regularly, because the reasons for accepting a risk can change.
Mistake five: closing findings without verifying fixes
A finding is often closed when the ticket is resolved, not when the weakness is confirmed as removed. That practice hides two problems. The first is that the fix may not have worked. The second is that the weakness may return when a system is rebuilt or a configuration is reverted.
The correction is to verify fixes, either through a rescan, a configuration check or a sample review. Where verification is impractical for every item, a sampling approach can give reasonable assurance. Track the rate at which closed findings reappear, because it is a direct signal of how well verification is working.
Mistake six: scanning without planning for impact
Active scanning can disrupt fragile systems, and authenticated scanning can take actions the operator did not intend. Programmes that run scans without agreeing timing and intensity with system owners risk causing the outages they are meant to prevent. The reputational cost can then undermine support for the whole programme.
Agree scan windows, intensity and methods with owners, particularly for sensitive systems. For fragile systems, consider configuration-based assessment as an alternative to active probing.
Mistake seven: relying on the scanner for everything
Automated scanning is valuable for breadth, but it does not replace skilled testing. Scanners struggle with business logic, chained weaknesses and some configurations. A programme that relies only on scanning may miss the issues an attacker would find most useful.
Penetration testing should complement scanning, directed at areas where automation is weakest. The two activities answer different questions and should be scoped accordingly.
Mistake eight: reporting activity instead of exposure
Reports that count scans run, tools deployed and findings closed describe activity. They do not show whether exposure is falling. A report that is full of activity figures may give leaders a sense of progress that the underlying data does not support.
Reports should include measures that connect to decisions, such as the number of high-priority findings beyond their target period, the status of critical systems and the number of exceptions past review. Each measure should have a clear definition.
Mistake nine: ignoring the change process
Vulnerability management that operates outside the change process creates friction. Patches arrive during freezes. Configuration changes collide with releases. Operations teams feel that security is working against them, and they respond by delaying work.
Plan scanning, patching and configuration changes together. Use the existing change calendar to schedule fixes, and make sure emergency changes follow a defined route.
How to use this guide
Choose three of the mistakes that seem most relevant to the programme. For each, find evidence in the current process. Ask whether the evidence shows the mistake is present, partly present or absent. Then decide which correction would have the largest effect and who would own it. The exercise often surfaces problems that have been accepted as normal.
Common mistakes about mistakes
Some teams respond to this list by trying to fix everything at once. That usually fails. Choose the one or two errors that most weaken the programme and correct them first. A second mistake is blaming individuals for errors that come from unclear process or ownership. Most of the errors described here are structural, and they are best addressed through design rather than exhortation.
Next steps
Review the current programme against the nine mistakes and record where each one appears. Agree the top two to correct, name an owner for each and set a date for review. Revisit the list after each review cycle, because new mistakes appear as the estate and the team change.
Frequently asked questions
What is the most common mistake in vulnerability management?
Working from an inventory that does not match the estate. Many other errors follow from that, including incomplete coverage and findings sent to the wrong owners.
Is closing a finding in the ticketing system enough?
No. The fix should be verified, through rescanning, configuration checking or sampling, before the finding is treated as resolved.
How can a team tell whether its exceptions have become a problem?
Check the register for items past their review date, items without a clear owner and items whose reasons no longer apply. Any of these signals a need for review.
Related reading
If you are working through this topic, these related articles cover neighbouring questions:
