Remediation Backlog vs Exposure Window: Close the Gap
Backlogs are growing and attackers are getting faster. Learn why remediation stalls and how to shorten the exposure window on what matters most.
The Remediation Backlog Is Growing. The Exposure Window Is Shrinking.
Ask most security teams how many open vulnerabilities they have, and they can answer within seconds. Ask how long their most critical internet-facing weakness stayed exploitable last quarter, and the room goes quiet.
That second number is the one that matters now. The remediation backlog in most organisations keeps getting longer, while the time attackers need to turn a newly disclosed flaw into a working attack keeps getting shorter. In its Annual Review 2026, the NCSC told organisations to "prepare now for when a 'patch wave' arrives", expecting "an influx of updates to address vulnerabilities across all severities", with a number of them critical. A backlog that was manageable last year will not survive that.
This article looks at why remediation stalls, why the exposure window is a better measure than backlog size, and how to rebuild remediation so it can keep pace.
What Is the Exposure Window?
The exposure window is the time between a weakness becoming exploitable in your environment and the moment it is no longer exploitable, because it has been patched, mitigated, isolated or removed. It is measured per vulnerability, per asset, and it is what an attacker actually experiences.
Backlog size tells you how much work is waiting. The exposure window tells you how long the door was open. An organisation can have a backlog of 20,000 findings and short exposure windows on everything that matters. Another can have 2,000 findings and leave a critical internet-facing flaw open for two months. The second organisation is in more danger.
Why the Window Is Shrinking
Three changes are compressing the time defenders have.
Discovery has sped up on both sides. The NCSC assesses that AI is helping attackers with vulnerability discovery and exploit development, and that defenders face a race to shorten patch timelines before working exploits appear. The same capability is finding older flaws in widely used software, which is why the NCSC expects a surge of updates rather than a steady flow.
Attacks are moving at machine speed. Automated tooling can scan the internet for a newly disclosed weakness within hours of publication. Commentators are now arguing that cyber defence has to move at machine speed because the attack side already does.
Exposure is concentrated in shared components. When the NCSC issued its alert on two critical PHP vulnerabilities in October 2026, it warned that the flaws "may allow attackers to execute commands on any web server running PHP", told organisations not to limit discovery to outdated versions, and sent proactive notifications to UK organisations through its Early Warning service. One flaw in a common component creates thousands of exposure windows at once, many of them inside appliances and third-party products that the security team cannot patch directly.
Why the Remediation Backlog Keeps Growing
Most backlogs do not grow because teams are careless. They grow because the process between "finding raised" and "fix deployed" has too many places to stall.
| Where remediation stalls | What it looks like in practice |
| --- | --- |
| Ownership | The finding sits with security, but the system belongs to an application team, a supplier or nobody at all |
| Change control | Fixes wait for a monthly change window designed for planned work, not urgent security updates |
| Testing confidence | Teams delay patching because they cannot quickly prove the update will not break production |
| Vendor dependency | The vulnerable component is bundled inside an appliance or product, and the fix depends on the supplier's release |
| Legacy constraints | The system is end-of-life, the update does not exist, and replacement is unfunded |
| Risk acceptance | Exceptions are approved once and never revisited, so accepted risk quietly becomes permanent risk |
None of these is solved by a better scanner. They are operational problems, and they need operational fixes.
Patching Is Not the Only Way to Close the Window
The NCSC's PHP alert is a useful model because it gives organisations a sequence of actions, not just an instruction to patch. In order of priority, it advised organisations to find every affected installation starting with internet-facing systems, update, apply temporary mitigations if updating was not immediately possible, consider taking the system offline if mitigations could not be applied quickly, check for compromise and backdoors, and track updates from third-party vendors whose products bundle the component.
That sequence matters because it separates closing the exposure window from completing the patch. A configuration change, a web application firewall rule, a network restriction or switching a service off can close the window in hours, while the full patch follows through normal testing. Teams that only track "patched or not patched" miss the faster options.
It also adds a step many remediation programmes skip. If a system was exposed while a critical flaw was being actively exploited, applying the patch does not prove the system is clean. Checking for compromise has to be part of closing the finding.
Prioritisation Narrows the Problem. Remediation Capacity Solves It.
Risk-based prioritisation is essential, and we have written about moving from CVEs to attack paths so teams fix what attackers can actually use. The reason is simple: organisations can't fix everything, and cannot always be sure which fixes are safe to postpone.
But prioritisation only decides the order of the queue. If the queue moves too slowly, a perfectly ranked backlog still leaves critical windows open. The organisations that keep pace treat remediation as a capability with its own design, capacity and metrics.
Pre-approve the emergency path
Agree in advance what qualifies as an emergency security change, who can approve it, and how quickly it can be deployed outside the normal change window. Deciding this during a critical alert costs days.
Map ownership before the alert arrives
Every internet-facing asset should have a named technical owner and a named business owner. When a critical alert lands, the first hour should be spent fixing, not finding out who is responsible.
Make deployment routine
Organisations that already patch frequently through automated pipelines and staged rollouts find urgent patching far less disruptive. Building security checks into delivery pipelines, as we covered in shift-left security for continuous delivery pipelines, shortens the path from fix available to fix deployed.
Put suppliers inside the timeline
The NCSC advises organisations to plan to deploy updates "quickly, more frequently, and at scale, including across their supply chains." That means contractual expectations on how quickly suppliers notify you of vulnerabilities, release fixes and confirm whether their products are affected.
Give risk acceptances an expiry date
Every exception should have an owner, a compensating control and a review date. An accepted risk with no expiry is simply an unmanaged one.
Measure the Window, Not the Pile
The UK government's own progress shows what focused remediation can achieve. Under the Government Cyber Action Plan, the NCSC reports that vulnerability management helped cut the median time taken to fix domain-related vulnerabilities from 50 days to 8 days.
That is the right kind of metric. For the board and the security leadership team, a small set of measures says far more than a total count of open findings:
- Median exposure window for critical vulnerabilities on internet-facing assets.
- Time to first mitigation, tracked separately from time to full patch.
- Age of the oldest critical finding that is exploitable from outside.
- Percentage of risk acceptances past their review date.
- Proportion of findings with a named owner within one working day.
These numbers show whether the organisation can close a window quickly when it matters, which is exactly what a patch wave will test. For board-level context on why this is now a governance issue, see our analysis of the NCSC Annual Review 2026 and AI ransomware risk for UK boards.
Prove the Fix Worked
A closed ticket is not the same as a closed exposure. Patches fail to apply, mitigations get rolled back and new instances are deployed from old images. Re-testing critical fixes, and running regular continuous security validation, confirms that the window really closed. Penetration testing that shows what an attacker can actually achieve then tells you whether the remaining backlog creates a realistic route to compromise.
Close the Windows That Matter First
The backlog will probably keep growing. More software, more disclosures and faster discovery make that likely. What organisations can control is how long the most dangerous weaknesses stay open, and how quickly they can respond when the next critical alert arrives.
Cyber Advisory Services helps security teams shorten exposure windows through vulnerability management programme design, C²TEM for continuous discovery and risk-based prioritisation of internet-facing weaknesses, penetration testing to validate fixes, and incident response support to check for compromise when a flaw was exploited before it was closed. Talk to our team to measure your real exposure windows and build a remediation process ready for the next patch wave.