Skip to main content
green gradient background, "The Future of Application Security Is Already Here." and a read the report button.
Does Your Policy Cover an Outage Nobody Attacked?Loss Modeling & Aggregation
5 min readFor Cyber Insurance Buyers & Brokers

Does Your Policy Cover an Outage Nobody Attacked?

Evaluating Your Cyber Insurance Coverage

This guide is for buyers and brokers assessing whether a cyber insurance policy covers outages caused by third-party provider failures without an attack. We'll use the 12-14 November Microsoft 365 outage as a case study. CyberCube estimates insured losses from this event will likely be under $100 million, with a conservative upper bound of $350 million. The outage wasn't due to an attack; Microsoft attributes it to a technical issue during a network capacity upgrade.

Low industry losses don't guarantee your policy would have paid out. They indicate the event was mild enough that most policies weren't tested. Use this guide to evaluate your coverage before a longer outage tests it for you.

Key Concepts

Non-malicious system failure. An outage caused by error, misconfiguration, or faulty change management, not an attack. Many policies only trigger on "security failure" or "unauthorized access" when something hostile occurs.

Contingent Business Interruption. Coverage for income loss when a third-party provider you rely on goes down, not your own systems. This coverage is crucial because you didn't operate the failed platform.

Waiting period. The time before Business Interruption Coverage starts counting loss. CyberCube notes that waiting periods of eight to 12 hours are common in contingent system failure cover.

Shared technology provider dependency. Many insureds rely on the same provider. For insurers, this drives Aggregation Exposure Analysis. For you, it means your outage is everyone's outage, and fallback options may be strained.

Policy Evaluation

CyberCube's loss reasoning provides a useful test. It identified several conditions that suppress insured loss, each mapping to a policy question you can answer today.

Loss-suppressing factor Policy question
Coverage responds only where policies extend business interruption to non-malicious system failure at an outside provider Does your Contingent Business Interruption grant include non-malicious failure, or only security events?
Reported impacts were inconvenience and productivity loss, not lost revenue Does your loss definition capture extra expense and productivity loss, or only net income?
Email isn't in the critical transaction path for most businesses Is yours? If email carries orders, approvals, or customer commitments, your exposure differs from the typical insured.
Event ran roughly 38 hours, with severe disruption of about 18 hours Does your waiting period run from the start of the outage or from when your operations are affected?

The 38-hour total ran beyond common eight- to 12-hour waiting periods. A policy with a 12-hour waiting period still had a long tail to cover. A policy with a 24-hour or 48-hour waiting period might have paid little or nothing on a 38-hour event with 18 hours of severe disruption.

Steps to Take

Follow these steps to assess your policy.

1. Read the trigger language, not the summary. Pull the Contingent Business Interruption insuring agreement. Look for the phrase that defines the covered event. "Security failure" and "system failure" are not interchangeable. Confirm system failure is covered for outside providers, not only your own network.

2. Check for named-provider limits. Some policies sublimit or exclude specific dependent providers, or cover only providers under written contract. Confirm that your cloud productivity suite, identity provider, and hosting platform are in scope.

3. Test the waiting period against a realistic outage. Take the 38-hour duration and the 18-hour severe-disruption window. Calculate when your waiting period would start and end under each reading. Ask your broker whether the clock starts at provider failure or at your documented impact.

4. Map your own critical transaction path. CyberCube's point that email wasn't critical for most businesses is a prompt, not a reassurance. List which business processes break if email, identity, or collaboration tools stop. If any revenue process depends on them, you're outside the typical loss profile.

5. Document fallback channels. Workarounds limited loss in this event. Record your alternates (a secondary messaging platform, phone trees, a status page) in your business continuity plan. Insurers and adjusters will ask what you did to mitigate. Tie this to the Respond and Recover NIST CSF Core Functions.

6. Review aggregation exposure on your side. If one provider carries email, files, identity, and security tooling, the failure is correlated. In this outage, Teams, SharePoint, Copilot, and security tools dependent on Exchange Online were degraded, though core functions largely continued. Dependencies cascade, so map them.

Common Pitfalls

  • Assuming "cyber" means any outage. Many policies were built around attacks. Non-malicious failure is an extension you must confirm in writing.
  • Treating the loss estimate as your loss. The industry figure reflects the average insured. Your exposure depends on your own transaction path.
  • Ignoring the waiting period. A long outage can still produce no recovery if the waiting period eats the loss window.
  • Counting only revenue. CyberCube noted impacts focused on productivity loss, with little to no mention of lost revenue. If your policy pays only lost income, productivity loss may fall outside it.
  • Skipping evidence of mitigation. Without documented fallback steps, expect friction at First Notice of Loss and during adjustment.

Quick Reference Table

Item What to confirm Where to look
Trigger Non-malicious system failure covered at outside providers Contingent Business Interruption insuring agreement
Waiting period Length, and when the clock starts Declarations page and definitions
Provider scope Which providers count as dependent Schedule or definition of dependent provider
Loss measure Net income, extra expense, productivity loss Loss definition and exclusions
Sublimits Separate limit for contingent cover Declarations page
Critical path Revenue processes tied to the provider Your BCM documentation
Fallbacks Documented alternates and activation steps Continuity and recovery plans
Notice First Notice of Loss timing requirements Notice and claims conditions

One last point. CyberCube's Jon Laux observed that "not all layers of the technology stack are equally critical to business operations." Rank your dependencies by how directly they sit in your revenue path, then check that your coverage follows the same ranking.

Promotional banner graphic asking if you are ready for PCI DSS 4.0 with a call-to-action to get the guide

You Might Also Like