Skip to main content

What the November Microsoft 365 Outage Should Change About Your Business

The November Microsoft 365 outage showed how one provider disruption can affect much more than email. From around 18:00 GMT on Thursday 12 November, services including Outlook, Teams and SharePoint were disrupted for more than 21 hours. The useful question for a small business is not whether to abandon cloud services, but what work could still happen if they became unavailable.

What happened during the outage?

Microsoft described the underlying condition as “a large number of servers being unexpectedly unavailable”. The disruption affected Outlook and Exchange Online, Teams, SharePoint, Copilot, Microsoft Defender screening and access to Intune dashboards. Some users also reported problems reaching cloud-based Windows machines.

Meaningful recovery began shortly before midday GMT on Friday 13 November. Microsoft said there was “no evidence of malicious activity” and that the incident was “not related to cyber or AI / agents”.

Reporting by The Guardian described a configuration update that caused servers to become overloaded and go offline. When engineers applied a fix, some servers became unresponsive again as traffic returned. Microsoft then limited traffic to affected servers in specific locations and rolled some back to an earlier software configuration.

Those details come from reporting about the incident; Microsoft’s full post-incident account is not yet public. Charles Lamanna, Microsoft’s executive vice-president for Copilot, agents and platform, said there was “encouraging recovery progress” and that teams were “reviewing what happened and how we can prevent similar issues going forward”.

UK impact was reported across public services, including police services, the Bank of England, NHS ambulances and schools. It was also the second major Microsoft cloud incident reported in 2026, following one in September.

The available information points to an availability incident: services could not be reached reliably. That is different from evidence that customer data was lost or stolen. A service outage is not, by itself, a data-loss incident.

Why a configuration problem can affect so much

A cloud service is not one application running on one machine. It depends on connected infrastructure and shared services. A change in one part of that system can have knock-on effects elsewhere, even when the change is not a cyber attack.

For a customer, the distinction may not feel significant at first. If Outlook, Teams and SharePoint are all unavailable, staff cannot message colleagues, open shared files or respond to email in the usual way. If other business applications rely on the same Microsoft account for sign-in, those could be inaccessible too, even if their own systems are operating.

That is a dependency worth checking. List the services your staff use, then ask two questions about each: where is its information stored, and how do people sign in? A separate supplier name does not necessarily mean a separate route in. An application that uses Microsoft sign-in may still depend on Microsoft’s identity services for access.

The Microsoft 365 outage is a reason to map those connections, not a reason to assume that cloud services are inherently unsafe. It is also a reminder that having a backup copy of data solves a different problem from keeping staff working while a provider is unavailable.

What to do while a provider is down

When staff report problems, first check Microsoft’s service health page. Microsoft published updates there during the incident. It is a better starting point than assuming an individual computer is at fault.

If a Microsoft 365 outage is confirmed, avoid making hurried local changes. Reinstalling applications, changing settings or resetting passwords will not restore a provider’s unavailable servers and could create extra issues that need undoing later. Follow updates from the service provider and your IT support contact, and tell staff where those updates will be shared.

Then keep work moving where it can. If a task does not need the unavailable system, staff may be able to continue it. If a process depends on email or an online application, use a documented fallback rather than creating informal workarounds that could lose information or confuse staff.

That response depends on having a route to communicate outside the affected service. If email and Teams are both unavailable, a message in Teams telling staff to check email is no use.

A continuity plan for the work that matters

A short plan is more useful than a long document no one can find. Decide in advance:

  • Who leads the response? Name one person to check provider updates, decide which fallback to use and tell staff when normal working can resume.
  • How will staff contact each other? Agree an out-of-band route that does not rely on Microsoft 365, and make sure staff know how to use it.
  • Which three processes matter most? Choose the tasks that would cause the biggest practical difficulty if paused, such as accepting orders, arranging deliveries or handling customer queries.
  • Where is the offline contact list? Keep a short list of key customer and supplier phone numbers somewhere accessible without Microsoft 365. Limit it to what staff need and manage it appropriately.
  • How will you record work? Decide how to capture orders, jobs or requests while systems are down, who is responsible for entering them later, and how to check that nothing was duplicated or missed.
  • How will you review the incident? Record the start and end times, the processes affected, the temporary steps used and any work that needs reconciling afterwards.

Use the plan to identify what is genuinely email-dependent and what only happens to be done by email out of habit. A team might be able to log a job on paper and enter it later; another process may need an online system to check stock or authorise a transaction. The distinction determines what the fallback must actually cover.

What organisations actually did during the outage

Reported examples from those two days show more than an invented scenario would. In France, the municipality of Grans told residents to telephone rather than email, according to The Guardian. In Japan, the shipping company Sagawa Express could not accept online reservations for some services at airports and railway stations, and the pharmaceutical company Taisho took orders by phone. Some employees there went back to sending faxes.

None of that is sophisticated, and that is the point. The fallback is a telephone, someone who knows the customer, and a written record kept until systems return. Organisations that carried on were not using a clever technical workaround; they had a route to each other and a way to note what had been agreed.

At the scale of a Hampshire business with ten or fifty staff, the same dependency is easier to miss. If your order or job system signs in through Microsoft, and staff reach each other only through Teams, an outage of that length leaves no route between people and no reliable way to capture work. A printed contact list, an agreed staff call tree and a numbered paper order log would not restore Microsoft 365. They would give the business a controlled way to handle urgent work and reconcile it afterwards.

Is a second productivity suite the answer?

Not automatically. Running two full productivity suites means paying for and maintaining both, while staff may split conversations and files between them. It can also leave people unsure which system contains the current version of a document or where a message was sent.

For a small firm, a more proportionate approach is to define fallbacks for the few processes that cannot simply wait: a separate staff contact route, a way to take essential requests and a written method for tracking work until services return. The right options depend on how the business works and what information it handles.

Keep backups in their proper role

The November Microsoft 365 outage was about service availability, not reported data loss. A backup does not make Outlook or Teams work while Microsoft’s servers are down. It remains important for a different reason: recovering information after accidental deletion, corruption or ransomware.

Keep those purposes distinct when reviewing your plan. Your continuity procedure covers how staff communicate and handle work during an outage. Your backup and recovery arrangements cover restoring data when it has been lost or damaged. Our guide explains how to test backups and recovery.

Can you claim anything back from Microsoft?

Possibly, and it is worth understanding how the mechanism works before you assume either that compensation is automatic or that there is no point asking.

Each Microsoft online service carries its own service level agreement, published in the consolidated Online Services SLA document. The commitment is measured per service, per month, not against your Microsoft 365 subscription as a whole. Microsoft publishes its own worldwide uptime figures in the service health and continuity documentation. Read them with the caveat that they are global averages across every customer, not a record of your tenant, and that a strong quarterly figure still leaves room for a bad day.

Where availability falls below the commitment for a specific service, the SLA provides for a service credit. Four things catch businesses out.

Nothing is issued automatically. An incident appearing on the Service Health dashboard means your tenant might have been affected. It does not prove it was, and it does not generate a credit. You have to submit a claim, and Microsoft then assesses it against the applicable SLA and the evidence you supply. Advisory items are generally not eligible at all.

The credit is calculated per affected service, not per subscription. If SharePoint was down but Exchange was fine, the claim covers SharePoint. Credits are prorated according to the service affected and the length of the outage, so it is not a single figure you can apply across a Microsoft 365 plan.

The maths is about user-minutes, not wall-clock hours. Exchange Online availability, for example, has historically been measured on how long the service was unavailable multiplied by how many licensed users were affected. A widespread incident does not mean every user was affected. In a 30-day month there are 43,200 minutes, so availability only falls short of the commitment if the qualifying downtime averaged out at roughly 43 minutes per licensed user across the month. Seven hours of outage that hit every user would clear that comfortably; seven hours where a third of staff were unaffected might not.

That is why the version of this story circulating on social media, in which a long outage means a fixed refund, is wrong. Windows Latest checked the terms and found the credit steps up as availability falls, but the figures it quotes come from an older volume licensing agreement, and Microsoft publishes a newer version of the SLA on its own page. Read the agreement that applies to your subscription rather than a figure from a forum post, and remember that any credit applies to the eligible service fees, not to your whole Microsoft 365 bill.

There is a deadline. For services other than Azure, a claim must be submitted within one month after the end of the billing month in which the incident occurred. Azure claims get two months. Miss the window and the conversation is over.

What you actually need to make a claim

Microsoft’s documented requirements are specific, and this is where most claims fall down. You need:

  • your tenant GUID;
  • the outage incident identifier from Service Health, which is a number prefixed by the affected service, such as EX for Exchange Online or SP for SharePoint Online;
  • the affected subscription and service, stated accurately rather than as “all of Microsoft 365”;
  • the start and end times of the outage and its duration;
  • the number and locations of affected users;
  • evidence of impact: screenshots, error messages, monitoring records or support correspondence;
  • a record of what you did to try to resolve it at the time, and any support ticket you raised;
  • a request sent from an email address on the affected tenant’s own domain. A personal address will not be accepted.

The route depends on how you bought the licence. If you bought directly from Microsoft, an administrator should:

  1. Sign in to the Microsoft 365 admin centre.
  2. Open Help & support, and turn off the Support Assistant toggle if it blocks the search box.
  3. Search for a request to review an SLA service credit, and choose Contact support, then email.
  4. Put “SLA service credit request” and the incident ID in the subject, then attach the evidence above.

The incident IDs from this outage were EX1550543 for Exchange Online and MO1550821 for the broader Microsoft 365 incident, so quote the ones that actually appear in your own Service Health history. If you buy through a reseller or Cloud Solution Provider, the claim goes through that partner: Microsoft accepts these requests from CSP direct and indirect providers, not from customers or indirect resellers directly. Once approved, the credit appears on the next invoice and in the reconciliation data.

One exclusion to know about: these business credits do not automatically cover Microsoft 365 Personal or Family subscriptions, which follow different terms. Staff using a personal account for work email have no claim under this process at all, which is another reason business email should be on a business tenant.

Two practical points are worth acting on now. First, gather the evidence while the incident is happening, because the Service Health history only retains resolved incidents for a limited period and your own notes are what prove impact later. Second, for a broad incident Microsoft publishes a preliminary post-incident review through Service Health within 48 hours of resolution and a final one within five business days, which is useful supporting material for a claim and for your own review.

Be realistic about the sums involved. For a business of ten or twenty staff, the credit is usually small next to the disruption it represents. The reason to do it anyway is that the evidence you gather to support a claim is the same evidence that tells you what actually broke, for how long and for whom.

Write the one-page plan this week

Set aside time to write down the response lead, the staff fallback channel, the three business processes you would protect first, and where an offline contact list is kept. Add a simple way to record incoming work and reconcile it when systems are restored. Then walk through the plan with the people who would use it: can they find the contact details, and do they know who makes the decisions?

A plan that has never been tried may have gaps. Test it with a short discussion exercise: assume Outlook and Teams are unavailable for a morning, then work out how staff would communicate, record urgent work and tell customers what to expect. You do not need to simulate a technical outage to find out whether the instructions make sense.

BITSmart can review Microsoft 365 dependencies, provide managed IT support, Microsoft 365 implementation, cloud services, cybersecurity and real-time system monitoring, and help you write a practical continuity plan. Book a Call to discuss which systems and processes your business would need to keep moving during an outage.

Tony Edwards, Chief Everything Officer at BITSmart
Tony Edwards

BITSmart’s Chief Everything Officer. Tony helps small businesses across Hampshire keep their IT secure and explained in plain English.

Is your Microsoft 365 set up securely?

Book a free 30-minute call and get an honest view of where you stand.

You might also like

Want an honest view of your IT security

Book a free 30-minute call and we’ll tell you where you stand.