Picture the report that lands on an IT manager's desk after the first vulnerability scan of the year. It lists every known CVE found across the managed fleet, sorted by severity, and it runs to several hundred rows. Everyone agrees it matters. Nobody agrees on what to do with it on Monday morning.
That gap between a list and a decision is where most vulnerability management efforts stall. The scan worked. The data is accurate. But a spreadsheet of identifiers is not a work queue, and without a way to rank it, assign it and prove it is finished, the list gets longer every month while the team keeps patching whatever is easiest.
This article is about closing that gap. It is not a patching tutorial; we covered the mechanics of automation in our guide to automated patch management and the exploit window. Here the question is different: once you know which known CVEs exist on your endpoints, how do you turn that knowledge into prioritized, verifiable action that a small team can sustain?
A CVE is an identifier for a publicly known flaw. It tells you that a weakness exists in a piece of software, and it carries a severity score. It does not tell you whether the affected machine holds payroll data or is a spare laptop in a drawer, whether anyone can reach it from the internet, or who in your organization is responsible for fixing it. Those answers live in your business, not in the scanner.
Severity scores describe how bad a flaw is in general. They do not describe how bad it is for you. A critical flaw on a device that is rarely used and holds nothing sensitive can wait behind a moderate flaw on the machine your finance team uses every day. Teams that sort purely by score end up spending their best hours on the wrong devices, and they find out only when something goes wrong on the right one.
The same CVE can appear on three machines or on three hundred. A count of affected devices is the first honest signal of how much work a finding really represents, and it is also a signal of how much good a single fix will do. Without that count next to each finding, two flaws with the same score look identical, even when one is a ten minute task and the other is a month of coordination.
Ownership is the other half. A finding that belongs to no one is a finding that never closes. In small IT departments the same three people patch servers, answer tickets and run onboarding, so a CVE without a named owner and a date quietly joins the backlog.
Leaving known flaws open has a cost that rarely shows up in a budget line. It shows up as the incident that consumes a week of the whole team, the audit that finds the same gap twice, and the outage that arrives at the worst possible moment. When you compare the effort of a disciplined weekly review with the total cost of ownership of one avoidable incident, the review is cheap. Ask GB Advisors for a current quote if you want to model that comparison for your own environment.
The visibility problem sits underneath all of this. If you cannot see which devices are in scope, you cannot see which vulnerabilities apply. Our piece on endpoint visibility for remote teams explains why that foundation comes first.
The method below works with any tool that can produce a list of known CVEs per device. It is deliberately simple, because a method that needs a workshop to run will not survive the second month.
Before you rank vulnerabilities, rank the things they live on. Sort your endpoints into a few groups that reflect business impact: devices that handle money or personal data, devices that run something the business stops without, devices that belong to executives or administrators with wide access, and everything else. Three or four groups are enough. You are not building a perfect model, you are creating a vocabulary that lets the team say "this finding is on a group one device" and have everyone understand the urgency.
This is the same logic behind prioritizing incidents by impact and urgency. Service desks already know that a ticket's priority comes from who is affected and how fast the damage grows. A vulnerability is simply an incident that has not happened yet.
With groups defined, rank each finding using three questions in order. How severe is it? How many devices carry it? Are any of those devices in your most valuable group? A moderate flaw present on every machine, including the critical group, usually deserves attention before a critical flaw on one isolated device. A count of affected devices makes this ranking possible, and a view of the devices with the most findings tells you where a single maintenance window clears the most risk.
Write the ranking rule down in one paragraph. When the rule is written, the weekly review becomes a five minute check instead of a debate.
Some findings have a straightforward remedy: a patch exists, it can be approved, and it can be deployed. Others need a different kind of work, such as removing software nobody uses, replacing an application that is no longer maintained, or scheduling a change with the business owner. Split the list into these two lanes early. The first lane is routine and can be largely automated. The second lane needs a person, a decision and a date.
Software that is no longer needed deserves special attention. Every application on a device is a surface that can carry a flaw, so the cheapest fix for some findings is to remove the application altogether. Ask the device owner whether the software is still used. A short conversation often closes a finding that no patch could.
Mixing the lanes is a common reason lists stall. The patchable items sit behind the hard ones, and the easy wins never get delivered.
Set deadlines by group and severity, and keep them short enough to mean something. A critical finding on a group one device might carry a few days. A low finding on a group three device might carry a quarter. The exact numbers are yours to choose, and they should reflect what your team can actually deliver, because a deadline that is missed every time teaches everyone to ignore deadlines.
Then name a person. Not a team, a person. Shared ownership is the fastest way to guarantee that nothing happens.
A finding is closed when it stops appearing on the device, not when a ticket is marked done. Rescan, confirm, and only then close. When you decide to accept a risk instead of fixing it, record who accepted it, why, and when it will be reviewed. An undocumented exception is indistinguishable from an oversight, and auditors treat it that way.
A last point on scope. The method above ranks work you can already see. It assumes that your inventory is complete and that every endpoint reports in. If part of the fleet is invisible to your tooling, the list understates your exposure, and every ranking built on it inherits the blind spot. Fixing coverage is therefore step zero, and it is worth revisiting every quarter.
Verification is the step that turns a patching routine into a vulnerability management practice. Most teams do the first four steps with reasonable discipline and then lose the thread at the end, because confirming closure feels like duplicate work.
A patch can be approved, queued and still not installed. Devices are offline, users postpone reboots, and a pending reboot can stop a patch cycle until it completes. Third-party software adds its own friction: if an application is open, the update can fail. Each of these leaves a device in a state where the dashboard says one thing and the machine says another. The only reliable evidence is a fresh look at the device after the fix.
The faster your assessment data refreshes after a change, the shorter the gap between acting and knowing. If a software update shows up as resolved within minutes, an engineer can confirm the result the same afternoon and move on. If confirmation takes a week, the work is already cold and failures are discovered long after the context is gone. When you choose tooling, ask how quickly it reflects an install, an update or a removal, because that interval decides how tight your loop can be.
A total number of open vulnerabilities is a poor headline. It rises when a new batch of CVEs is published even if your team did everything right. Report three things instead: how many findings were closed this period, how long the oldest critical finding has been open, and how many exceptions are active. Those numbers show whether the practice is working, and they hold steady when the outside world gets noisier.
Another useful companion is the discussion in our article on why backup alone is not enough for ransomware recovery. Closing known flaws reduces the chance you will need to restore, and backups limit the damage if you do. The two practices answer different questions.
Method is only half of the answer. The other half is cadence. A practice that depends on heroic effort once a quarter will fail, and a practice that fits into a standing hour each week will survive vacations, incidents and staff changes. The rhythm below assumes a team of two to five people sharing endpoint duties.
Open the overview and read only the top of it: the findings with the widest reach, the devices with the most open findings and the device roles most affected. Apply the ranking rule from step 2 and pick the handful of findings that will be worked this week. The goal is not to clear the list. The goal is to choose, in ten minutes, what the team will finish by Friday and who owns each item.
Resist the urge to widen the scope in the same meeting. A short list that gets finished builds trust in the process. A long list that gets half finished teaches the team that the weekly review is optional.
Approve and deploy the fixes for the patchable lane, and schedule the harder items with their business owners. After each deployment, go back to the same finding and check whether it still appears. If it does, find out why before the end of the day: the device may be offline, a reboot may be pending, an application may have been open, or the patch may not have been offered to that machine at all. Each cause has a different remedy, and the sooner you learn which one applies, the cheaper it is to fix.
End the week with a three line report: findings closed, oldest open critical finding and active exceptions. Send it to the person who funds the IT team, in plain language, with no screenshots of dashboards. This is how a technical routine earns executive trust. It also gives you a natural place to raise the exceptions that need a business decision, such as an application nobody wants to retire.
Over a few cycles the pattern becomes visible. Closed findings per week stabilize, the oldest critical finding gets younger, and the conversation with leadership shifts from "are we exposed?" to "what is the next decision?". That shift is the real return on the practice, and it does not require a larger team to achieve.
NinjaOne Vulnerability Management covers the identification and patching parts of this method, inside the same console that manages the endpoints. Ownership, deadlines and exceptions stay with your team.
NinjaOne Vulnerability Management identifies known CVEs on managed endpoints. Once activated under Administration, Apps, Installed, it covers two sections: NinjaOne Patch for Windows and macOS, and Real-Time Software Assessment for Windows only. It reports known CVEs, which means it finds published flaws and does not claim to find unknown ones. That scope matches the first lane of the method: things the industry has already catalogued and that your team can act on today. Because only known CVEs are in scope, a clean result means no catalogued flaws were found, not that a device is free of every possible weakness.
Real-Time Software Assessment, which is Windows only, works from inventory NinjaOne has already collected. Software CVEs update within a few minutes of an install, an update or a removal. OS-related CVEs update about one hour after a patch scan. That timing is what makes step 5 practical for software findings: an engineer can deploy a fix and see the result reflected within minutes. For OS findings the result follows the next patch scan by about an hour. The Vulnerabilities tab shows the affected device count for each CVE, so you can confirm the count falls after the fix.
The Vulnerability Overview shows totals by severity, plus three views that map directly to step 2: Top 10 Vulnerabilities by Device Count, Top 10 Impacted Device Roles and Top 10 Devices by Vulnerability Count. These lists show the top 10 entries only, so they are a starting point for prioritization and not a full inventory. A Vulnerability access permission controls who can see the data. Read together, the views answer three practical questions: which flaw touches the most devices, which kind of device is most exposed, and which single machine deserves the first maintenance window.
The Vulnerabilities tab lists severity, CVE ID, operating system, description, remediation patch (KB) and affected device count for each finding. The remediation patch column connects a finding to the fix, and the affected device count gives you the reach measure for ranking. Each row works as a decision aid you can read in a minute: how severe the flaw is, which patch addresses it and how many devices still need it. Start with the rows that have the largest device count, and the rest of the list becomes easier to schedule.
NinjaOne patching lets you approve security updates by severity (Critical, Important, Moderate, Low), and each category can be set to Approve, Manual or Reject. CVE and CVSS data in the patch views come from the NIST database, and a Patch CVSS Score policy condition is available. Apply now deploys a patch on demand, bypassing the patching policy, and the Patch Management Dashboard tracks Pending, Approved, Rejected, Installed and Failed statuses so you can see whether the fix landed.
For the wider platform picture, see the NinjaOne page on GB Advisors and our cybersecurity solutions overview.
Many IT teams in the region run lean. One small department often supports several offices or several countries, with a mix of newer laptops, older desktops and machines that are rarely on the corporate network. In that setting, a method that depends on a dedicated security analyst reading raw scanner output is unrealistic. The five steps above are designed for the opposite situation: a handful of people, a weekly rhythm and a clear rule for what comes first.
Connectivity matters too. Devices that sit offline for days, or connect over variable links from home or from regional branches, make verification harder, because the evidence arrives late. That is a practical reason to favor assessment data that refreshes quickly after a change, and to treat an offline device as its own category in the weekly review instead of an invisible one.
Managed service providers serving several clients face a version of the same challenge at larger scale. Ranking by reach and by device group is what keeps a multi-client practice from drowning in findings. Our article on how MSPs manage multi-tenant IT covers that operating model in more detail.
GB Advisors serves 1,600 clients across Latin America, the Caribbean, the United States and Canada. In our experience as advisors, the missing piece is rarely another scanner; it is a simple rule that turns findings into owned, dated work.
Banks, retailers, logistics firms and manufacturers across the region often operate large numbers of endpoints in branches, warehouses and plants, where a single device can be a point of sale, a scanner or a production workstation. For those environments the device group in step 1 is not an abstract idea. A point of sale terminal belongs in the most valuable group because the business stops when it stops, and its findings should be ranked accordingly even when the severity score looks modest.
Use this as a standing agenda. If you can answer yes to each line, your vulnerability management practice is moving from reports to results.
Monthly patching is the foundation, and it handles the routine lane well. What it does not tell you is which flaws remain open after the cycle, which software sits outside the patch routine, or whether failed installs were noticed. A vulnerability view answers those questions and gives the patch routine a way to prove its own results.
It is a defined scope, and defined scope is useful. Known CVEs are the flaws that have been published and that attackers can read about as easily as you can. Closing them removes the most accessible entry points. It does not replace other controls, and it does not claim to find flaws nobody has catalogued yet. Treat it as one layer in a broader security approach.
The method here fits into a weekly review of under an hour once the grouping and the ranking rule exist. The alternative is not "no process", it is an unranked list that grows and a team that reacts to whichever finding is loudest. A light, regular practice costs less than recurring emergencies.
Top 10 views are a prioritization aid, designed to point you at the findings with the most reach. They are the right place to start each review, and the Vulnerabilities tab lists each finding with its severity, CVE ID, remediation patch and affected device count for the deeper work. Start with what the top 10 shows, clear it, and the next review begins with a fresh ranking.
Treat them as a category of their own. A device that is offline cannot be patched, and its findings may be out of date, so it appears in your weekly review as a separate line with its own owner. Chasing offline devices is dull work, but ignoring them means the fleet you believe you protect is smaller than the fleet you own.
In large organizations it may be. In most mid-sized companies the people who run endpoints are the same people who patch them, and the security conversation happens through them. Giving that group a clear ranking rule and a way to verify closure is how security outcomes improve without hiring a separate team.
A list of known CVEs is information. Prioritized action is the result you actually want: findings ranked by what matters to the business, assigned to a named person with a date, fixed through a routine patch lane where a patch exists, and verified on the device before anyone calls them closed. Tools help by showing severity, reach and the remediation patch in one place, and by refreshing software findings within minutes, so you can confirm many fixes soon after they land. The discipline around them is what turns data into protection.
If your team has a list and no ranking rule, start with the endpoint groups from step 1 and write the ranking paragraph from step 2. It takes an afternoon, and it changes the next review.
Talk to GB Advisors about building a vulnerability management practice that fits your team, your devices and your business priorities. We will start with your process and recommend the platform second.