Ransomware or a Late Patch? The Surprising Toll of 43 Days
Archi4 min read
Contents
- The Number That Reframes the Debate
- Path A: Ransomware With a Long Fuse
- Path B: The Unpatched Edge Device
- Where the Two Paths Converge
- What the 43 Day Figure Does Not Tell You
- A Practical Comparison for Budget Holders
- The Cost of Choosing Wrong
- Three Signals to Watch During Dwell Time
- The Question to Ask This Quarter
The Number That Reframes the Debate#
Security teams argue about which incident type is worse: a ransomware attack that encrypts everything, or a quiet intrusion that sits undetected for weeks. The argument usually centers on recovery costs and downtime. But a single number from incident response data changes the comparison: 43 days. That is the median dwell time for ransomware operators before they detonate a payload, according to several IR firms that track such figures. In other words, the encryption event is often the last chapter, not the first.
This matters because the two most common incident paths, ransomware and unpatched software exploitation, are often treated as separate problems with separate budgets. They are not. The 43 day window is the overlap. It is also where the cheapest fixes live.
Path A: Ransomware With a Long Fuse#
Ransomware crews rarely break in and encrypt in the same hour. They phish a credential, move laterally, escalate privileges, and then wait. The waiting is deliberate. Attackers map backups, test restore procedures, and time the payload for a weekend or a holiday. By the time the help desk gets the first encryption alert, the intruders have often been reading email for a month.
That long fuse is a gift, but only if someone is watching for it. Most organizations detect ransomware at the encryption stage, not the intrusion stage. The 43 day figure is a measure of how long the gift goes unopened.
Path B: The Unpatched Edge Device#
The second path looks different on paper. An internet facing appliance, a VPN concentrator, a file transfer server, runs a version with a known flaw. A patch exists. It is not applied. An attacker exploits it and gets a foothold. From there, the incident can become ransomware, data theft, or a cryptominer, depending on who shows up.
The surprising part is not that unpatched devices get exploited. It is how often the patch was available for weeks or months before the intrusion. In many incident reports, the gap between patch release and exploitation is measured in weeks, not hours. The defender had the fix. The attacker had the time.
Where the Two Paths Converge#
Put the two timelines side by side. Ransomware gives you an average of 43 days of dwell time. Unpatched software gives you a window that starts the day a patch ships and ends the day an exploit lands. Both are windows. Both are detectable. Both are routinely missed.
The convergence explains a pattern that shows up in post incident reviews: the breach that looks like a ransomware story often began as a patch management failure. The initial access vector was an old flaw. The dwell time was long. The encryption was just the finale.
What the 43 Day Figure Does Not Tell You#
The median hides the tails. Some ransomware incidents move from access to encryption in under 24 hours. Those are the ones that dominate headlines because they are dramatic. But the median is the more useful planning number because it describes the typical case, and the typical case is slow. Security budgets that only prepare for the fast tail are overpaying for the wrong scenario.
A Practical Comparison for Budget Holders#
If you have a fixed security budget, the comparison is not ransomware versus patching. It is detection during dwell time versus detection at the moment of impact.
Option one: invest in faster patching. This closes the window on known flaws. It is measurable. You can count days from release to deployment. The limit is that patching never reaches zero, and some appliances cannot be patched quickly without downtime.
Option two: invest in detection of lateral movement and privilege escalation. This targets the 43 day window directly. It assumes the perimeter will fail and watches for the behavior that follows. The limit is that detection requires tuning and staff, and false positives erode trust.
Option three: do both, but stage them. Start with the handful of edge devices that face the internet. Patch those first, then build detection around the identity and backup systems that ransomware crews target during dwell time. This is the option most IR firms recommend after a breach, which is also the most expensive time to implement it.
The Cost of Choosing Wrong#
The cost asymmetry is stark. A ransomware event that reaches encryption typically involves days of downtime, restoration from backups, and often a payment decision. An intrusion caught during dwell time is usually a cleanup: reset credentials, rebuild a host, review logs. The difference in cost is often an order of magnitude or more. The 43 day figure is the size of the opportunity to stay in the cheaper category.
Three Signals to Watch During Dwell Time#
Ransomware operators leave traces during the waiting period. Three are consistent across incident reports.
First, credential use from unusual locations or at unusual hours. The attacker is learning the environment.
Second, backup and restore system access. Before encrypting, crews test whether backups can be deleted or encrypted. Any unusual access to backup infrastructure is a warning.
Third, privileged account creation. New admin accounts that nobody requested are a classic dwell time artifact.
None of these require a new product. They require someone to look.
The Question to Ask This Quarter#
Ask your team a simple question: if an intruder were in the network right now, how many days would pass before we noticed? The answer is your real dwell time, not the industry median. If it is close to 43, the encryption event is not the incident you are preparing for. The waiting is. The patch you deferred and the alert you did not tune are the same problem wearing two hats. The cheapest fix is the one that shortens the window before the payload runs.
More articles