
Vulnerability management: from scan report to fixed finding
Vulnerability management is not the scan but everything that happens after it: who gets the finding, what goes first, when it is done and how you show that it is done. In most organisations the scanner runs fine and the report sits there. This is the order that turns a list of vulnerabilities into a programme you can keep up.
A scanner produces findings; vulnerability management makes sure every finding gets an owner, a deadline and a check afterwards.
Do not prioritise on CVSS alone. Whether a flaw is being exploited, whether the system is reachable from the internet and how much it matters decide the order.
An exception is a decision by the system owner, with a reason, a temporary control and an expiry date. Without an expiry date it is a forgotten finding.
Vulnerability management starts after the scan report
A vulnerability scan tells you what is open. Vulnerability management is the question of what you do about it, and that question comes back every week, because new vulnerabilities are published every week and your environment changes too. A two-hundred-line report that ends up in a folder is not security; it is an archive.
The process has six steps, and it only works once all six are written down:
- Know what you have: an inventory in which every system has an owner.
- Collect findings: from the scanner, but also from other sources, in one list.
- Prioritise: on severity, exploitation, exposure and how much the system matters.
- Assign: an owner and a deadline per finding.
- Fix, mitigate or accept: and the last one only with a decision.
- Verify and report: scan again and show how it is going.
The step where most organisations stall is the fourth. The scanner runs, everyone sees the list, and nobody feels they own line 47.
Without an inventory you do not know what you are missing
A scanner only sees what it can reach. A laptop that has been at someone’s home for three weeks, a cloud service a department signed up to on its own, or a server in a network segment that was not in scope will not appear in any report. The first thing a mature programme does is make visible the gap between what you think you have and what the scan finds.
Record at least four things per system:
- An owner: who decides about the system, not just who looks after it technically.
- How much it matters: what stops if it fails or leaks. A simple high, medium and low is enough.
- Exposure: is it on the internet, only on the internal network, or in a segregated part?
- Who may update it: you, your IT partner or the vendor. For printers, cameras, tills and medical or industrial equipment it is often the manufacturer, and that decides how quickly anything can happen.
Combine a network scan with agents on endpoints. A network scan finds unknown devices; an agent still sees a laptop when it is not in the office.
Where findings come from
The scanner is the largest source but not the only one. A workable process puts everything in one list, so that the same vulnerability does not turn up three times with three different owners:
- The vulnerability scanner, internal and external.
- Endpoint security software, because many EDR products also report missing updates and vulnerable software.
- Security advisories from vendors and from national bodies. In the UK, the NCSC publishes alerts about vulnerabilities that are being actively exploited.
- Penetration test reports and audit findings. They find what a scanner cannot, such as excessive permissions or flaws in your own application.
- Reports from outside, from a customer or a researcher who has found something. Make sure there is an address where that lands and someone who reads it.
The goal is not the longest list but a single list. Deduplicate on system and vulnerability, not on source.
Prioritising: a CVSS score is a start, not an order
Every vulnerability gets a CVSS score from 0 to 10, and it is tempting to start at the top. That works badly: a first scan often returns dozens of critical lines, and the score says how bad a flaw can be technically, not how likely it is that someone will use it against you. Four questions sharpen the order:
- Is it being exploited? The US agency CISA keeps a public catalogue of known exploited vulnerabilities (KEV), and EPSS from FIRST estimates the probability of exploitation per vulnerability. If a flaw is on that catalogue, it goes first.
- Can the system be reached from outside? A critical flaw on a VPN gateway or a web server is a different matter from the same flaw on a machine that only runs internally.
- How important is the system? Finance or the domain goes before a test server.
- Is there already another control? A service that is switched off or shielded lowers the risk, even if the software has not yet been updated.
Translate that into a small number of classes, four for example, and attach a deadline to each. Then nobody has to argue line by line: the class decides the order. For UK organisations with Cyber Essentials, the scheme already sets one of those deadlines: high and critical updates within 14 days of release.
Every finding has an owner and a deadline
A finding without a name next to it does not get fixed. So agree who is responsible for each kind of system: endpoints with IT operations, a business application with its application owner and the software vendor, a network device with whoever runs the network. Raise a ticket per finding or per group of findings, so progress sits in the same system as the rest of the work.
Patching is one way to fix something, not the only one. Sometimes a configuration change is enough; sometimes you switch a service off, add a firewall rule or replace the device. How updates themselves are rolled out, with rings and maintenance windows, is covered in our article on patch management; vulnerability management is the layer above it that decides what goes first and checks whether it worked.
Agree the deadlines per class once, with senior management involved, so they are not reopened at every incident. Add what happens when a deadline is missed: who is told and who then decides.
Exceptions: accepting risk with an expiry date
Not everything can be done on time. A production machine runs an operating system that no longer receives updates, a vendor has no fix yet, or an update breaks an integration the business needs. That is normal. What should not be normal is a finding like that sitting quietly on the list.
An exception that works has five parts:
- A decision by the system owner, not by the administrator. Whoever carries the risk accepts it.
- A reason that still makes sense a year from now.
- A temporary control, such as isolating the system, restricting access or extra monitoring.
- An expiry date on which the decision is reviewed.
- A place in the reporting, so the number of exceptions stays visible.
Showing that the programme is working
A finding is only closed once a new scan confirms it has gone. Marking a ticket as resolved is a claim; the rescan is the proof. Schedule it automatically, so nobody has to remember.
Do not report the total number of vulnerabilities: that figure says little and moves with every new publication. What does say something:
- How many findings have passed their deadline, per class.
- How long remediation takes on average, also per class.
- How many exceptions are open and how many have reached their expiry date.
- How many systems the scan found that were not in the inventory.
That is also what an auditor or a customer wants to see. ISO/IEC 27001 has a dedicated control for it (8.8, management of technical vulnerabilities), and NIS2 explicitly lists vulnerability handling among the measures an organisation must take.
Where our work stops
What we do. We scan, prioritise and fix findings in the environment we manage: endpoints, servers, network and Microsoft 365. We keep the list with owners and deadlines, schedule the rescan and report in the form above. We are ISO/IEC 27001 certified, and that is how we work ourselves.
What we do not do. We do not change code in another vendor’s software; for a vulnerability in your business application that vendor has to act, and we follow it up. We do not decide whether a risk is accepted: that belongs to the system owner. And we are not an auditor: we certify nobody and do not declare that your programme meets ISO 27001, NIS2 or Cyber Essentials. That judgement comes from an independent auditor, a certification body or your regulator.
Questions we get about this
The five that come up most often.
What is vulnerability management?
Vulnerability management is the ongoing process of finding, prioritising, fixing and verifying vulnerabilities in your systems. The scan is only the start; the process makes sure every finding gets an owner, a deadline and a rescan.
What is the difference between vulnerability management and patch management?
Patch management is about rolling out updates: which ones, on which devices and when. Vulnerability management is broader: it decides which vulnerability goes first, whether an update is the right fix or another control will do, and whether the problem has really gone. An update is one of several ways to close a finding.
How do you decide which vulnerability to fix first?
Look beyond the CVSS score. A flaw that is being exploited, on a system reachable from the internet that matters to the business, goes first. Public sources such as the CISA KEV catalogue and the EPSS score help you estimate that likelihood of exploitation.
Can you leave a vulnerability unfixed?
Yes, if the system owner consciously decides to. Record the reason, put a temporary control in place such as isolation or extra monitoring, and set an expiry date on which the decision is reviewed. An exception without an expiry date is a finding that gets forgotten.
Can you confirm our vulnerability management meets NIS2 or ISO 27001?
No. We set up the process and run it in the environment we manage, and we provide the reporting an auditor wants to see. Whether it complies is assessed by an independent auditor or the regulator; we certify nobody.
Further reading
What this process connects to.
Shall we put your scan report in order?
Send us the latest report you have. We sort the findings into classes with an owner and a deadline, so it is clear what has to happen this week and what can wait.
Practical IT knowledge in your inbox
New guides on management, security and the workplace, written by the people doing the work. No sales talk, and you can unsubscribe in one click.