---
title: Ransomware isn't one problem. It's five.
description: Ransomware isn't one problem — it's five. A walkthrough of the five-layer framework (email, identity, containment, detection, backup), what 2026 breach data shows, and where ARJ covers.
---

[Making IT easy without to much clicking](https://www.arjdistribution.se/en/arjblog)

# [Ransomware isn't one problem. It's five.](https://www.arjdistribution.se/en/arjblog/ransomware-isnt-one-problem-its-five)

 Written by [Roger Jansson](https://www.arjdistribution.se/en/arjblog/author/roger-jansson) | Nov 21, 2026, 12:19:43 PM

*Most conversations about ransomware collapse into a single question — "do you have backup?" It's the right instinct, but the threat is layered, and understanding all five layers is what decides whether an organisation has a bad day or a bad year.*

Ransomware isn't really a special category of cyberattack. It's an ordinary intrusion — a stolen password, a tricked employee, an unpatched system — that happens to end with the attacker locking up files instead of only stealing them. That framing matters, because it means the defences that work against ransomware are, for the most part, the same defences that work against everything else. There is no silver bullet. There are five distinct jobs, and most organisations do one or two of them well.

This piece walks through the five: what each one does, what the 2026 breach data says about how much it matters, and where the gaps tend to show up. It also says plainly which layers our own portfolio covers and which it does not.

## 1. Stopping the break-in

Every attack starts somewhere. Verizon's 2026 Data Breach Investigations Report, built from confirmed breaches, puts exploitation of vulnerabilities as the most common way in at 31 per cent of breaches, up from 20 per cent the year before. Credential abuse follows at 13 per cent. Mandiant's M-Trends 2026, drawn from incident-response work, reaches a similar picture: exploits were the top initial infection vector at 32 per cent.

Phishing and stolen credentials have not gone away, but the weight has shifted towards unpatched, internet-facing systems. The human element is still present in 62 per cent of breaches in the Verizon data, so awareness training remains relevant, but patching has moved up the list. Hornetsecurity's own 2027 Ransomware Report, a survey of 400 IT and cybersecurity professionals, found that 84 per cent of organisations now provide awareness training, up from 74 per cent, while only 9 per cent believe their employees could confidently identify an AI-generated phishing or impersonation attempt. It is a self-reported vendor survey, so it says more about attitudes than about the threat itself.

31%of breaches began with an exploited vulnerability (up from 20%)

13%began with credential abuse

62%involved a human element

The controls here are email filtering that checks links and attachments, patch management for applications and systems, and security awareness training. Web filtering of general browsing also belongs in this layer. Link protection in email and in Microsoft Teams does not cover general browsing — it is worth separating the two. **Hornetsecurity Teams Protection scans URLs in Teams messages in chats and channels inside your tenant, including URLs pointing to files hosted on public servers, but does not scan (a) URLs to files shared via OneDrive or SharePoint, or (b) chats or channels started by an external user.** Compared with the later layers they tend to be inexpensive per user, which is worth noting given how many intrusions start here.

## 2. Locking the account, even if the door opens

Prevention fails sometimes. This layer assumes the front door gets opened occasionally and asks a narrower question: if a criminal has a stolen password, can they use it? And if malicious software lands on a laptop, does it get to run?

Multi-factor authentication is the most-cited control. **Microsoft Research (2023) reports that MFA reduces the risk of compromise by 99.22 per cent across the whole population. For accounts where attackers already have leaked credentials the figure is 98.56 per cent.** The catch is coverage. MFA enabled for 90 per cent of accounts leaves the other 10 per cent as the easiest way in, and many organisations do not know their real coverage without a tool that checks it. Conditional access policies, which decide who can sign in from where and on what device, belong here too, as does endpoint protection on the devices themselves.

**Worth knowing** Identity posture tools that continuously check Microsoft 365 for missing MFA, weak sign-in rules and configuration drift solve a problem that is otherwise invisible: rollout usually happens gradually, and nobody re-audits it.

## 3. Containing the spread

Criminals rarely stop at the first machine. It is a launchpad, not a destination. This is the layer many organisations underinvest in, partly because its value is hard to picture: it does not stop an attack, it limits how far one gets once it has started.

CrowdStrike's 2026 Global Threat Report puts the average "breakout time", the time between landing on one machine and moving to another, at **29 minutes** for 2025. Few security teams can reliably make a containment decision in under half an hour. That is the argument for treating this layer as architecture rather than reaction: segmentation, least-privilege access and privileged-access controls have to be in place before the incident starts, because there usually is not time to build them during one.

## 4. Catching it in the act

No prevention layer catches everything, which is why detection is its own discipline. Mandiant's M-Trends 2026 puts the global median dwell time, the time between compromise and detection, at **14 days** in 2025, up from 11 the year before. When organisations found the intrusion themselves, it took about 9 days. When an outside party had to tell them, the median was 25 days. Just over half of organisations detected malicious activity internally.

> In nearly half of cases, someone else — a partner, a customer, or the attacker's own ransom note — delivers the news first.

IBM's Cost of a Data Breach Report 2026 puts the average time to identify and contain a breach at 247 days, and found that organisations using security AI and automation extensively saved about USD 1.93 million and 65 days compared with those that did not. Detection tends to be the most expensive layer to close properly, usually running through endpoint detection and response (EDR/XDR) and a security operations function, whether in-house or bought as a service.

## 5. Making sure "no" is still an option

Ransomware's business model depends on the victim having no way to recover except by paying. Take that leverage away and the extortion largely collapses. Sophos' State of Ransomware 2026, based on 2,158 organisations in 17 countries that were hit in the past year, found that 56 per cent of attacks succeeded in encrypting data, and that 66 per cent of those victims restored from backups, up 12 percentage points from the year before.

66%of victims with encrypted data recovered from backups (Sophos 2026)

USD 1.7Maverage recovery cost per incident, up 11% (Sophos 2026)

USD 769kmedian ransom payment (Sophos 2026)

Backups are also a target. Veeam's 2023 Ransomware Trends report found that over 93 per cent of attacks went after backup repositories, and that 75 per cent of those attempts succeeded. That is an older figure, but the point still stands: it is not enough to have a backup. The copy has to be protected from changes and deletion by whatever compromised the rest of the environment, kept separate from the production tenant or servers where possible, and restored in a test before the day it is needed. A backup an attacker can also delete is only the appearance of one.

Preparation and confidence are not the same thing either. In the Hornetsecurity survey mentioned above, 91 per cent of respondents say they protect their backup storage against ransomware and 84 per cent have a disaster recovery plan, yet only 55 per cent are very or extremely confident in their ransomware protection strategy. Again, these are self-reported answers from a small sample, but they fit the point that having a plan is only part of being able to recover.

···

## What this adds up to

No single layer is sufficient, and that is not a shortcoming of any product; it is the nature of the problem. An organisation can do everything right at the front door and still be reached through a vulnerability nobody knew was exposed. It can catch an intrusion in nine days and still lose the backup that was supposed to make the timing irrelevant. The layers that feel least urgent — containment and detection — are the hardest to explain to a budget owner, and often the ones with the strongest numbers behind them.

## Where ARJ Distribution fits, and where it does not

For partners working out how the Microsoft 365 side of this stack fits together, here is how our portfolio maps to the five layers today, based on what the vendors state about their own products.

| Layer | Coverage | What is in the portfolio |
| --- | --- | --- |
| 1 · Stop the break-in | Covered | Hornetsecurity 365 Total Protection (email security with Advanced Threat Protection, which checks links and attachments in email delivered to the organisation), Teams Protection (links in Microsoft Teams, with the two limitations noted above), DMARC Manager and Security Awareness Service. Recast Release and Patch Management for application patching. We do not offer web filtering for general browsing. |
| 2 · Lock the account | Partial | Mostly the Microsoft 365 account side: Octiga (baselines for authentication and conditional access, MFA follow-up — the platform works regardless of licence, and features that require a specific Microsoft licence such as Entra ID P1 or P2 are hidden from customers who lack them) and EtherInsights (Secure Score, risky identities). Single sign-on through Recast Access Manager, an add-on to Application Workspace. No endpoint protection (EDR/AV). |
| 3 · Contain the spread | Not in the portfolio | No vendor for network segmentation, Privileged Access Management or zero-trust enforcement. Permission and posture tools give visibility into access in Microsoft 365 but do not enforce anything. |
| 4 · Catch it in the act | Partial, Microsoft 365 only | Octiga rates events in the Microsoft 365 audit log, alerts, and can block an account automatically after a critical event. EtherInsights shows risky identities and configuration drift. No EDR, XDR or SIEM, and no monitoring of endpoints or networks. |
| 5 · Be able to recover | Covered | Hornetsecurity 365 Total Backup (Microsoft 365) and VM Backup (servers), Xopero ONE, GitProtect.io and EtherInsights Tenant Backup for Microsoft 365 configuration. How well each product protects against deletion differs, so ask which one is meant. |

Layers three and four, and endpoint protection in layer two, are worth sourcing separately.

For a deeper walkthrough per layer and vendor, see our [security hub](https://www.arjdistribution.se/en/security?hsLang=en), which lays out the framework in detail and openly states what we cover and what we do not.

**Sources**  
 Verizon, *2026 Data Breach Investigations Report* — initial access vectors, human element, ransomware share.  
 Mandiant (Google Cloud), *M-Trends 2026* — initial infection vectors, global median dwell time, internal vs external detection.  
 Sophos, *The State of Ransomware 2026* — 2,158 IT and cybersecurity leaders in 17 countries.  
 IBM, *Cost of a Data Breach Report 2026* — time to identify and contain, security AI and automation savings.  
 CrowdStrike, *2026 Global Threat Report* — average eCrime breakout time.  
 Veeam, *Ransomware Trends Report 2023* — backup repositories targeted.  
 Hornetsecurity by Proofpoint, *Ransomware Report 2027* (survey of 400 IT and cybersecurity professionals, survey data 2026) — awareness training, AI confidence, backup protection and confidence. A self-reported vendor survey, used as a supplement only. [Get the report from Hornetsecurity](https://www.hornetsecurity.com/en/ransomware-report/?partnerid=0011v00002yaY0vAAE).  
 Microsoft Research (2023), *How effective is multifactor authentication at deterring cyberattacks?* — MFA effectiveness, 99.22 per cent across the whole population and 98.56 per cent for accounts with leaked credentials.  
 Product statements are based on the vendors' own documentation and may change; check current vendor information before relying on them.

 This piece reflects publicly available research as of November 2026 and is intended for general education, not as a substitute for a tailored risk assessment.

[View full post](https://www.arjdistribution.se/en/arjblog/ransomware-isnt-one-problem-its-five)

```json
{
  "@context" : "http://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Roger Jansson"
  },
  "dateModified" : "2026-11-21T15:03:31.057Z",
  "datePublished" : "2026-11-21T12:19:43Z",
  "headline" : "Ransomware isn't one problem. It's five.",
  "image" : {
    "@type" : "ImageObject",
    "height" : 627,
    "url" : "https://143252521.fs1.hubspotusercontent-eu1.net/hubfs/143252521/ransomware-hero-v4-en.png",
    "width" : 1200
  },
  "mainEntityOfPage" : "https://www.arjdistribution.se/en/arjblog/ransomware-isnt-one-problem-its-five",
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "height" : 60,
      "url" : "/hs/hsstatic/content_shared_assets/static-1.4092/img/default-amp-logo.png",
      "width" : 60
    },
    "name" : "ARJ Distribution makes IT easy with best of breed IT solutions"
  }
}
```