IT management

Patching by real-world risk: CVE, CVSS and CISA KEV explained

Thousands of new vulnerabilities appear each month. How to decide which updates can't wait, using CVSS, CISA's KEV list and EPSS.

On the second Tuesday of the month, Microsoft releases its security updates. Your IT provider rolls them out to the laptops a week or two later, once they've had time to check nothing breaks. It's a sensible routine. Meanwhile, the firewall that runs your VPN has a flaw that criminals started exploiting last Thursday, and it isn't on anyone's monthly list.

That is the weak spot of patching on a schedule: it treats every update the same. Most vulnerabilities are never used in a real attack. A small number are used within days, and those are the ones that lead to ransomware and stolen data.

This article explains the three terms you'll see in every security advisory (CVE, CVSS and KEV), why a "critical" score doesn't always mean "urgent", and how a small business can decide what to patch first.

Too many updates to treat them all the same

The number of published vulnerabilities keeps rising. One independent count found a record 48,185 new CVEs in 2025, about 20% more than the year before. No IT team, let alone a small one, can give each of them the same attention.

Attackers, meanwhile, have become faster. A study of vulnerabilities exploited in 2023 put the average time between a flaw becoming known and its first exploitation at five days, down from 32 days in 2021 and 2022. The same study found that most of those vulnerabilities were exploited before a patch was even available.

Defenders are slower. A large annual breach study found that exploiting a vulnerability was the way in for 20% of breaches, and that organisations took a median of 32 days to fully fix the vulnerabilities it tracked on firewalls and VPNs. Attackers work in days, defenders in weeks: that is the gap a schedule alone can't close.

CVE and CVSS in plain terms

A CVE (Common Vulnerabilities and Exposures) is simply a reference number for a known flaw, such as CVE-2025-14847. It lets your IT provider, the vendor and the CCB talk about the same problem without confusion.

CVSS (Common Vulnerability Scoring System) gives that flaw a score from 0 to 10. For the current versions, the US National Vulnerability Database uses these bands: Low 0.1 to 3.9, Medium 4.0 to 6.9, High 7.0 to 8.9 and Critical 9.0 to 10.0.

The catch is in what the score measures. FIRST, the organisation that maintains CVSS, states plainly that CVSS measures severity, not risk. The base score describes how bad a flaw would be if exploited. It doesn't know whether anyone is exploiting it, whether the software is on one PC or on your firewall, or whether that device is reachable from the internet.

So a "critical" flaw in software on one laptop in a locked office may matter less this week than a "high" one in the VPN everybody uses from home.

Three questions that measure real-world risk

Is it being exploited right now?

This is the most important question, and there is a free answer. The US Cybersecurity and Infrastructure Security Agency (CISA) keeps the Known Exploited Vulnerabilities (KEV) catalog, a list of flaws with reliable evidence of active exploitation and a clear fix.

The list is short on purpose: a little over 1,600 entries in mid-2026, built up since 2021, compared with tens of thousands of new CVEs each year. Its deadlines are only binding on US federal agencies, but CISA recommends that every organisation prioritises the vulnerabilities on it. Each entry also says whether the flaw is known to be used in ransomware campaigns, which is the closest thing to a "drop everything" label.

Closer to home, the CCB publishes advisories on serious vulnerabilities, and many carry an explicit "Patch Immediately" in the title. Belgian organisations can also register for Safeonweb@work, the CCB's platform that includes alerts about threats and vulnerabilities affecting your organisation.

How likely is it to be exploited soon?

Most flaws aren't on any "exploited" list yet. For those, the Exploit Prediction Scoring System (EPSS), also maintained by FIRST, estimates the probability that a vulnerability will be exploited in the next 30 days. It's free and updated daily. You don't need to read it yourself, but your IT provider or scanning tool should be using it, or something like it, alongside CVSS.

Where does the vulnerable software sit?

The same flaw is more dangerous on a device anyone on the internet can reach. Edge devices (firewalls, VPN gateways, remote access tools) made up 22% of exploitation cases in that same breach study, up from 3% the year before. These devices are often outside the normal Windows update routine, which is exactly why they get forgotten.

A simple priority model

You don't need a complex scoring system. Three tiers cover most small businesses.

Priority What falls in it Sensible target
Urgent Known exploited (on KEV or a CCB "Patch Immediately" warning) and on an internet-facing system: firewall, VPN, remote access, website Within a day or two, out of hours if needed
High Known exploited on internal devices, or critical with a high likelihood of exploitation Within a week
Routine Everything else The normal monthly cycle

These targets are our working recommendation, not a legal standard. What matters is that you agree them in advance, so nobody has to debate on the day whether a VPN flaw can wait until the end of the month.

What to do:

  • Ask your IT provider how they decide what goes outside the monthly cycle, and whether they check KEV and CCB warnings.
  • Make sure someone owns the firewall and other edge devices. They need updates just as much as laptops.
  • When a patch isn't available yet, follow the vendor's or CCB's temporary mitigation, such as switching off the affected feature.

The schedule still matters

Risk-based patching is a fast lane on top of your routine, not a replacement for it. The monthly cycle keeps everything else current, and it is still where most fixes land. Keep it, and keep a small test group that gets updates first so a broken update doesn't stop the whole office.

Don't forget third-party software. PDF readers, browsers, video call apps and compression tools all have their own update cycles, and Windows Update doesn't cover them. A laptop with every Microsoft patch and a two-year-old PDF reader is still exposed.

You can't patch what you can't see

Prioritising needs a list: which devices you have, what software runs on them and which versions. If you use Microsoft Defender for Business, included in Microsoft 365 Business Premium, its vulnerability management dashboard shows an exposure score, a software inventory and security recommendations ranked by how much they reduce your risk. That only covers devices onboarded to Defender, which is one more reason to run the device count check from antivirus vs EDR.

A managed provider can run that scanning and patching continuously, but whoever does it, the inventory has to come first.

What that looks like in practice: when we started with Go-Flow, the first scan found 73 vulnerabilities from the KEV list still open and critical flaws waiting 41 days on average. With continuous scanning and risk-based patching, critical vulnerabilities are now fixed in 3 days. Read the Go-Flow case study.

What NIS2 expects

If your company falls under NIS2, Article 21 lists vulnerability handling among the minimum security measures. The NIS2 guide for Belgian businesses explains who is in scope. Even if you're not, your customers increasingly ask how quickly you patch, and "monthly, everything" is a weaker answer than "critical exploited flaws within days".

Where to start

  1. List what faces the internet: firewall, VPN, remote access tools, website and any server reachable from outside. These go first.
  2. Get a software inventory of your laptops and servers, for example from Defender's vulnerability dashboard.
  3. Agree the three priority tiers and their targets with whoever patches your systems, and write them down.
  4. Subscribe to warnings: register for Safeonweb@work and make sure someone checks CCB advisories and the KEV catalog, or that your provider does.
  5. Add third-party apps to the monthly cycle, not just Windows and Office.

The takeaway

Patching everything once a month is a good baseline, but attackers don't wait for your schedule. Ask three questions of every serious flaw: is it being exploited, how likely is it to be, and can the internet reach it. Start with your firewall and VPN this week: that's where a five-day head start matters most.

← All articles

Not sure where you stand?

Find your gaps before an attacker does

A free, no-obligation assessment maps your risks and gives you a prioritised plan, whether or not you work with us.