Managed IT / Guide
Patch Management Outsourcing: What a Good SLA Actually Covers
Nov 18, 20269 min read
Patch management is one of the few security controls that works against threats nobody's found yet, and one of the easiest things for an SME to let slide. This guide covers the update, mitigate or accept decision NCSC itself recommends, and what a patch management outsourcing SLA needs to spell out so responsibility doesn't quietly disappear.

Somewhere on your network right now, a handful of machines are probably running software with a fix already published for a flaw someone else has already found. That's not a hypothetical: patch management — deciding which updates get applied, when, and to which systems — is one of the few security controls that works even against an attacker nobody has heard of yet, because it closes a door before anyone tries it. It's also one of the easiest things for a small business to let slide, because nothing visibly breaks the week you skip it. This guide covers what patch management actually involves as an ongoing process, the update-mitigate-accept framework worth knowing before you promise a system owner "we'll fix it," and — since most Dutch SMEs end up handing at least part of this to a provider — what a patch management outsourcing arrangement needs to spell out so responsibility doesn't quietly fall through the gap between you and whoever's doing the work.
What patch management actually involves
Patch management isn't a standalone task; it's the "treat" step inside a wider, cyclical process the Dutch National Cyber Security Centre (NCSC) calls vulnerability management. NCSC's own breakdown runs in five phases: assess (inventory your assets and scan them for known weaknesses), prioritise (score each finding by exposure, how critical the affected system is, the current threat landscape, and potential impact), treat (update, mitigate, or accept), evaluate (rescan and check whether the fix actually reduced risk), and improve (fix the underlying process, not just the one finding). The cycle doesn't stop once a patch is installed — it starts over with a fresh assessment, which is exactly why "we patched everything last quarter" isn't the same thing as having patch management in place.
For a small business, the point that matters most in that breakdown is the first one: you cannot patch what you haven't inventoried. NCSC's own guidance is blunt about this — without an accurate list of what's running, where, and on which version, there's no way to know where the priorities actually lie, and that applies as much to a forgotten network printer running outdated firmware as to the server everyone remembers.
Update, mitigate, or accept: the actual decision
When a known weakness turns up, there are only three real options, and NCSC frames them deliberately as a hierarchy rather than a menu. Updating is the default and, in NCSC's own words, always the best choice when it's available — but "available" doesn't mean "immediate." Before an update goes out, the system owner should weigh what it might break: bumping a database version can lose data or stop an integration working, so that call belongs with the people who depend on the system, not with whoever happens to be running the patch cycle that week. Testing an update before rolling it out, wherever that's practical, is NCSC's own standing advice, precisely because it surfaces compatibility problems while they're still cheap to fix.
Mitigation is the fallback when updating isn't an option yet — the vulnerable system sits in a time-sensitive process, say, and taking it offline isn't realistic this week. Mitigating means limiting what the weakness can actually do: isolating the system on its own network segment, blocking the networks an attack would have to come from, or temporarily disabling the affected feature and adding extra logging around it while a real fix is arranged. NCSC is explicit that mitigation is temporary — agree a deadline for it with whoever owns the risk, and when that date arrives, check again whether updating has become possible instead of quietly letting the mitigation become the permanent arrangement.
Accepting the risk is what's left once updating and mitigating are both off the table — commonly because a system is near the end of its life and a replacement is already underway, so patching it properly isn't worth the effort for whatever time it has left. Acceptance only works within an organisation's own risk appetite, and it carries the same requirement as mitigation: a review date, not an open-ended shrug. Our guide to replacing legacy software covers that end-of-life scenario in more detail, including the point where "accept, for now" stops being defensible.
One line from NCSC's own advice is worth repeating on its own: automatic updates for ordinary business software — a browser, a PDF reader, the handful of utilities everyone has installed — rarely break anything, and they save more staff hours than almost any other single setting. Save manual review and testing for the systems where a broken update is actually expensive.
Outsourcing patch management: what belongs in the SLA
Most SMEs that have looked at this seriously end up outsourcing at least part of it, and NCSC's own guidance treats that as entirely normal — it explicitly says organisations can buy tools or bring in a party to make vulnerability management "easier, more effective, or more efficient." What it doesn't say is that outsourcing makes the responsibility disappear. The same NCSC material is direct about this, phase by phase: when a third party carries out part of your vulnerability management, it's "important that the process is followed properly," and it's on you to confirm the provider can actually account for every point in that process, not assume it because you're paying them.
In practice that means putting a short list of things in writing before signing anything, rather than discovering the gaps when something goes wrong:
- Who updates what. NCSC's own phrasing is the right question to put to a provider directly: does updating devices and software packages fall under your scope, or theirs? Get that answered in the SLA, not in a sales conversation.
- Response timelines by priority, not one blanket turnaround. A critical, internet-facing weakness and a low-priority internal one don't belong on the same clock.
- Who sets priority, and on what information. If the provider scores and ranks vulnerabilities, they need your context — which systems are actually critical to your business, not just technically exposed — kept current, not handed over once at kickoff and never updated again.
- Your risk appetite, in writing. A provider can only judge whether a mitigation or acceptance is reasonable if they know what you're prepared to tolerate; otherwise every judgment call defaults to theirs.
- Reporting you can actually read. Ask for what was tested, what wasn't, and why — not just a list of applied patches. A gap you don't know exists is the one that gets exploited.
One detail is easy to miss: asset inventory and testing discipline don't go away just because a provider is doing the work. If you're outsourcing to free up capacity rather than because you've lost visibility into your own environment, keep — or ask for — a current list of what's running where. It's the one thing that lets you sanity-check a provider's reporting instead of taking it on faith.
The mistakes that show up every time
No current asset inventory. This is the single most common gap NCSC's own guidance calls out, and it's also the one that quietly undermines everything downstream — you can't prioritise or treat what you don't know exists. A single unmanaged network printer with outdated firmware has been the entry point into more than one otherwise well-patched network.
Treating "we have automatic updates on" as the whole answer. Automatic updates cover the software vendors ship updates for on a schedule you don't control. They don't cover the custom application nobody's touched the dependencies on, or the device that was never added to the list in the first place.
No agreed deadline on a mitigation. A temporary workaround with no review date is how a two-week stopgap quietly becomes a two-year one, without anyone ever deciding it should.
Assuming an outsourced provider owns a step nobody actually wrote down. The gap between "we assumed they handled that" and "that was never in the SLA" is where responsibility actually goes missing — not through anyone's bad faith, just through nobody writing it down.
FAQ
Is patch management the same as vulnerability management? No — patching (updating) is one of three options inside the "treat" phase of the broader vulnerability management cycle, alongside mitigating and accepting. Vulnerability management also includes the inventory, scanning, scoring, and evaluation steps around it; patch management on its own is just the part where a fix actually gets applied.
How often should we be patching? NCSC's own guidance doesn't set one fixed interval for every organisation — it ties patch and scan frequency to what your processes actually need and can act on, not to a calendar rule that merely sounds thorough. What matters more than a number is that critical, internet-facing systems get checked and treated meaningfully faster than something sitting on an isolated internal network.
Does outsourcing patch management mean we're no longer responsible if something goes wrong? No. NCSC's guidance is explicit that responsibility for following the process properly stays with your organisation even when a third party carries out the work — which is exactly why the SLA needs to say, in writing, who updates what and on what timeline, rather than leaving it assumed.
What if a system simply can't be patched anymore because the vendor has stopped supporting it? Then updating is off the table, and you're choosing between mitigating (isolating or restricting it) and accepting the risk with a review date — or replacing it. Our guides to Windows 10 end of support and replacing legacy software cover that situation for an operating system and a custom application respectively.
Should every system be on the same patch cycle? No — and treating them that way is itself one of the more common mistakes. Priority should follow exposure (is it reachable from the internet), criticality (what breaks if it's compromised), and the current threat landscape for that specific weakness, which is exactly the scoring NCSC's own prioritisation phase is built around.
Getting this off the whiteboard and into an SLA
Patch management done properly isn't glamorous, and it rarely gets credit on the days nothing happens — which is precisely the problem with leaving it informal. Our Managed IT & Support team builds and maintains the asset inventory this whole process depends on, runs the update-test-deploy cycle on a schedule that fits your business rather than a generic default, and writes the SLA terms above into the contract instead of leaving them assumed. Because an unpatched system is a security gap before it's anything else, our Cybersecurity & Identity team folds patch status into the same review as the rest of your exposure, rather than treating it as a separate IT housekeeping task. If outsourcing your broader IT support is still an open question rather than a settled one, our guide to IT help desk outsourcing covers that decision at the level above this one.
- Patch Management
- Vulnerability Management
- Managed IT
- SME
Related Articles
Managed IT
What Does Managed IT Cost for SMBs? Prices Per User (2026)
How much does managed IT support really cost for an SMB in the Netherlands? This guide breaks down price-per-user tiers, what each package includes, one-off and per-server costs, and how to compare quotes without surprises in 2026.
Managed IT
How to Choose the Right IT Partner for Your SMB (Checklist)
Choosing an IT partner is a multi-year decision that touches every part of your business. This practical checklist covers the questions to ask, the red flags to avoid, how to read an SLA, and how to compare managed service providers fairly.
Managed IT
What Does IT Workplace Management Cost in 2026?
How much should IT management per workplace actually cost? This guide breaks down 2026 price bands per tier, what a workplace really includes, and how to build the true monthly cost, not just the management fee.