Microsoft outage: does your SME have a plan B?

Microsoft outage: does your SME have a plan B?

Updated on 19 November 2026.

In short: from Thursday 12 November at 17:20 to Friday 13 November at 17:10 (Belgian time), Microsoft 365 email was unavailable for most users, worldwide. It was not a cyberattack, and it will not be the last outage: you cannot prevent it, but you can decide now how your business will keep working next time. Half a day of preparation covers the essentials, and it costs almost nothing.

What happened on 12 November

On Thursday 12 November, late in the afternoon, Outlook stopped connecting. Not just for you: for most Microsoft 365 customers, in 74 countries and territories according to Downdetector. Email (Exchange Online) was hit hardest. Teams, Copilot, SharePoint, OneDrive and Defender were disrupted for the features that rely on it.

On 15 November Microsoft published a preliminary post-incident report. It puts the start of the impact on 12 November at 16:20 UTC (17:20 in Belgium) and the end on 13 November at 16:10 UTC (17:10). Just under 24 hours, then, and a full working day lost for many European businesses.

In plain terms, three things added up:

  • a dormant software defect, installed in July on the servers that act as the "front door" to email. It had gone unnoticed for more than three months, because Microsoft's load tests ran on less powerful hardware than production;
  • a rise in load: the night before, Teams had automatically failed over to those same servers, a routine operation;
  • reduced capacity: nearly 14% of those servers were already out of service because of another problem the day before.

When the first servers dropped out, traffic was redirected to the others, which dropped out in turn. In less than two hours, more than 80% of the servers had stopped responding. The partitioning between regions, meant to contain this kind of cascade, did not work. Repair was then slow: every restarted server was immediately swamped by waiting traffic. Europe returned to normal on Friday around 12:35, Belgian time.

Microsoft has announced a final report with the full list of fixes. What is already clear: there was no cyberattack, and nothing you could have done would have prevented the outage.

Not an isolated case

It is not the first time this year. From 31 August to 3 September 2026, another Microsoft 365 incident (reference EX1464935) had already disrupted email for several days, with features coming back piece by piece. No cloud provider is immune: Google, Amazon or a Belgian hosting company have outages too. The real question is therefore not "is Microsoft reliable?", but "what does my business do when its main tool stops?".

What it cost Belgian SMEs

On 13 November the Flemish federation Unizo (self-employed and SMEs) described "enormous disruption": business owners who could no longer reach their customers or sign contracts. According to its figures, one SME in two could not continue operating without Microsoft's digital services, and only about one business in ten has an emergency plan for this kind of incident. Its CEO, Bart Buysse, sums it up: you cannot prevent an outage at Microsoft, but you can stop your business from being completely paralysed.

What still worked, and what was lost

This is the good news of this outage, and it matters for what comes next: a good part of the work was still possible, provided you knew what to use.

Tool During the outage What to know
Outlook (email) No connection for most users, sending and receiving heavily delayed Messages already downloaded remained readable in Outlook installed on the PC, and replies could be prepared.
Incoming mail Delayed, not lost Microsoft extended the delivery window so that queued messages would not bounce back to the sender. They arrived afterwards, sometimes all at once.
Teams Chats and meetings worked for most users during most of the outage Teams was able to leave the failing infrastructure, email was not.
OneDrive, SharePoint, Word, Excel Restored earlier, with disruption Files synchronised to the PC could still be opened and edited.
Phone, website, invoicing software Not affected (unless they depend on Microsoft email) These are your natural back-up channels.

In other words: a business with Outlook installed on its PCs, its files synchronised and its customers' numbers in a phone lost a day of email, not a day of work. One that worked only in the browser, with files "in the cloud" and contacts "in Outlook", was left empty-handed.

The trap afterwards: fake "recovery" emails

Scammers did not wait for the outage to end. From 13 November, phishing campaigns exploited the confusion, and they were still circulating on the 17th:

  • a fake "Service Health" notice announcing that messages held during the outage would be deleted within 24 hours, with a "Restore messages" button. The email quoted Microsoft's real incident number to look credible (Mimecast);
  • a fake "Outlook is disconnected after the outage" message, asking you to "reconnect" your account with a code. Behind it, an imitation of the Microsoft sign-in page, trying either to steal the password or to get the scammer's device authorised on your account (Kaspersky).

The rule is simple and applies to every outage: Microsoft never asks you to click in an email to "recover" your messages after an incident. The real state of the service is checked on the official page or with your IT person, never from a link received by email. We covered these techniques in our article on AI in the hands of scammers, and the most expensive scenario in the one on fake IBAN changes.

What Microsoft owes you (not much)

Microsoft commits by contract (the Service Level Agreement, or SLA) to 99.9% monthly availability for email. Below that, it grants a credit on the invoice, not a refund, and only on the email part of the subscription:

Monthly availability Credit
below 99.9% 25%
below 99% 50%
below 95% 100%

November has 43,200 minutes. A user without email for the 23 h 50 of the outage therefore had availability of about 96.7%: at best a 50% credit on one month's email share. To give an idea, with a simple assumption: if that share were worth €5 per person per month, an SME of 10 people would recover €25 at most. The contract excludes everything else: lost revenue, working hours, unhappy customers.

Three practical conditions:

  • the credit is not automatic: you have to claim it, with the incident number (EX1550543), the start and end times for your users, their number and the errors encountered;
  • the deadline is short: the claim must arrive before the end of the month following the incident, i.e. 31 December 2026 for this outage;
  • if you buy Microsoft 365 through a reseller, the reseller files the claim. Its internal deadlines may be shorter.

The SLA measures downtime on Outlook on the web. If your users could still reach their email in the browser, Microsoft does not count it as downtime. The credit remains symbolic: it is only worth the effort for teams of several dozen people. Your real safeguard is your plan B.

The plan B in six points

This plan is aimed at an SME of 5 to 30 people. It requires no heavy investment, only decisions and a little preparation.

1. Know where to look and who decides

When Outlook stops responding, the first question is: "is the problem on our side or at Microsoft?". The answer is on the public Microsoft 365 service status page (status.cloud.microsoft), on the official X account @MSFT365Status and, for the administrator, in the admin center ("Service health"). Decide in advance who checks, who warns the team and who decides to switch to plan B. In a small business, it is often the owner.

2. Be able to work offline

  • Outlook installed on the PCs rather than the web version alone. It keeps a local copy of mail: 30 days by default in the new Outlook (adjustable up to 180), 12 months in classic Outlook.
  • Essential folders synchronised to the PC, not just viewable online: current contracts, price list, schedule, quote templates.
  • Key contacts outside Microsoft: the numbers of your 20 most important customers and suppliers in the company phone, or on a printed list.

3. A back-up channel ready to use

If your address is @your-company.be and your email is hosted by Microsoft, all mail to that domain goes through Microsoft, outage or not. The back-up channel must therefore be elsewhere:

  • a team instant-messaging group (Signal or WhatsApp), created in advance;
  • a back-up mailbox at another provider, known to your key customers (for example a Belgian or European address, or the one from your internet provider);
  • the phone, quite simply. If your telephony also runs through Teams, plan a forward to a mobile. Our article on business telephony explains how.

A back-up address is useless if nobody knows what to do with it: write down who uses it and who checks it, and keep the day's appointments and useful contact details accessible outside Microsoft.

4. A message ready for your customers

Write it today, calmly: three sentences for the website, voicemail and, if you have them, social media. "Our email is disrupted by a worldwide Microsoft outage. For anything urgent, call us on … or write to … . Messages sent to us are not lost and will be handled as soon as service is restored." This message reassures customers and stops those in a hurry from going elsewhere.

5. A security instruction for the whole team

"During and after an outage, we do not click on any email that talks about it." One sentence, said to the whole team, that would have stopped both campaigns described above.

6. Do not confuse back-up and plan B

A Microsoft 365 back-up (by Microsoft or a third-party tool) protects your data against deletion, ransomware or a handling error. It does not give you your email back during an outage: your messages are safe, but you still cannot send any. There are "email continuity" services that take over during an outage, but they mainly target larger companies. For an SME of 5 to 30 people, points 1 to 5 deliver the essentials at a fraction of the cost. For back-up itself, the 3-2-1 rule remains the reference: see our guide to protecting your data.

Should you leave Microsoft 365?

Not because of this outage. One day of interruption per year is about 99.7% availability: imperfect, but hard to beat for an SME running its own mail server, with its own outages, its updates and nobody to fix it on a Friday evening. Other cloud providers have outages too.

The question of dependence remains legitimate, but it arises for other reasons: licence costs, data location, US law. We covered it in leaving the big tech platforms as an SME and in the example of the Dutch tax authority. In both cases, the plan B described here remains useful: it also protects against an internet outage, a blocked account or a stolen PC.

And NIS2?

Most SMEs of 5 to 30 people are not directly subject to NIS2, the European cybersecurity rule transposed in Belgium in 2024. Those that are, or that work for a company that is, must provide for business continuity: back-up management, disaster recovery and crisis management are among the mandatory measures. The plan B above is a first building block, easy to show to a customer or an auditor.

An illustrative case

Invented case, but built from common situations. An accounting firm of eight people in Namur uses Microsoft 365 Business Standard. On Thursday 12 November at 17:30, Outlook no longer connects. Nobody knows whether the problem is in the office; they restart the router, then the PCs, and lose an hour. On Friday morning the outage continues. The staff who worked in the browser have nothing left. Two customers, getting no answer, call personal mobiles. An employee receives a "Restore messages" email and clicks; fortunately, she stops at the sign-in page.

With the plan B, the same day goes differently: at 17:35 the office manager checks Microsoft's status page and confirms a worldwide outage. She sends a message to the team's Signal group, puts the prepared message on the website and the voicemail, and repeats the instruction "we click on nothing". On Friday everyone works in installed Outlook and on synchronised folders. Urgent matters go by phone. On Monday the administrator files a credit claim, on principle.

In practice: three options

Option Cost Time Risk
Do nothing €0 0 At the next outage, a lost day, customers without answers and a team exposed to fake "recovery" emails.
Do it yourself €0 to a few euros a month (back-up mailbox) Half a day: appoint someone responsible, write the customer message, create the team group, print the key contacts Low for the organisation. The technical side still needs checking: Outlook installed and properly configured, file synchronisation, alerts.
Get support A few hours of service One day, spread out The lowest: workstation configuration checked, service health alerts sent to an address outside Microsoft, protection against code phishing, a written plan tested once with the team.

The first step costs nothing: appoint today the person who, at the next outage, checks the service status and warns the team.

For your IT person

What follows is for the person who administers your Microsoft 365. For the rest of the team, the six points of the plan B are enough.

  • Service health: in the Microsoft 365 admin center, Service health, configure email notifications to at least one address outside Microsoft 365 (otherwise the alert lands in the mailbox that is down). Keep the public page status.cloud.microsoft at hand.
  • Outlook offline: new Outlook, Settings, General, Offline, "Days of email to save" (7, 30, 90 or 180; 30 by default). Classic Outlook: check that Cached Exchange Mode is on (12 months by default).
  • OneDrive: for critical folders, "Always keep on this device" rather than Files On-Demand alone.
  • Device code phishing: the post-outage campaigns used sign-in by code (device code flow). With Entra ID P1 (included in Business Premium), a Conditional Access policy can block this authentication flow for users who do not need it. In all cases: multi-factor authentication everywhere.
  • SLA credit claim: admin center, Support, new service request, quoting EX1550543 (and MO1550821), with times, number of affected users and observed errors. Deadline: 31 December 2026. Through a reseller (CSP): the reseller files it.
  • Teams telephony: if calls go through Teams Phone, plan a call forward to a mobile that can be activated without Teams.
  • Inventory: know who is global administrator, with which authentication methods, and where the emergency access details are written down. Our articles on the pitfalls of a small PC fleet and on the Microsoft account give the method.

Test yourself

During the worldwide Microsoft 365 outage of 12 and 13 November 2026, were the emails sent to affected businesses lost?

Show the answer

Answer: No, they were delayed and then delivered, as Microsoft extended the delivery window

Incoming mail was queued. Microsoft extended the delivery window so that messages would not bounce back to the sender: they arrived after service was restored.

After a Microsoft 365 outage, you receive a Service Health email threatening to delete your held messages within 24 hours unless you click Restore messages. What do you do?

Show the answer

Answer: I do not click and I check the service status on the official page or with my IT person

This is a real phishing campaign, observed from 13 November 2026. Microsoft never asks you to click in an email to recover messages after an incident.

What does Microsoft pay a business deprived of Microsoft 365 email during an outage of almost 24 hours?

Show the answer

Answer: On request, an invoice credit on the month's email share, from 25 to 50% depending on duration

Microsoft's SLA provides a credit, not a refund: 25% below 99.9% monthly availability, 50% below 99%, 100% below 95%. It must be claimed, and business losses are excluded.

Sources

Abrams, L. (2026, November 16). Microsoft says months-old bug caused global Microsoft 365 outage. BleepingComputer. https://bleepingcomputer.com/news/microsoft/microsoft-says-months-old-bug-caused-global-microsoft-365-outage

Belga. (2026, November 13). Panne chez Microsoft : la panne a des conséquences pour de nombreuses PME belges [in French]. La DH. https://www.dhnet.be/dernieres-depeches/2026/11/13/panne-chez-microsoft-la-panne-a-des-consequences-pour-de-nombreuses-pme-belges-2WHX2FY5T5HF7PIG5TJLHVZ2SA/

Born, G. (2026, September 3). EX1464935: Microsoft kämpft noch mit M365-Ausfall [in German]. Born's Tech and Windows World. https://borncity.com/blog/?p=329122

Dedenok, R. (2026, November 17). A Microsoft outage as a pretext for phishing. Kaspersky Daily. https://www.kaspersky.com/blog/microsoft-365-outage-phishing/56724/

Directive (EU) 2022/2555 of the European Parliament and of the Council of 14 December 2022 on measures for a high common level of cybersecurity across the Union (NIS 2 Directive). (2022). Official Journal of the European Union. https://eur-lex.europa.eu/eli/dir/2022/2555/oj

Microsoft. (n.d.). How to work offline in Outlook for Windows. Microsoft Support. https://support.microsoft.com/en-us/Outlook/getstarted/how-to-work-offline-in-outlook-for-windows

Microsoft. (2026a, November 15). Preliminary Post Incident Report for Microsoft 365 (EX1550543/MO1550821). Published by Next. https://next.ink/wp-content/uploads/2026/11/PostIncidentReport.pdf

Microsoft. (2026b, November). Service Level Agreement for Microsoft Online Services (WW). Microsoft Licensing. https://www.microsoft.com/licensing/docs/view/Service-Level-Agreements-SLA-for-Online-Services

Mimecast Threat Research and Intelligence. (2026, November 14). Attackers exploit the Microsoft 365 outage with fake “held messages” lures. Mimecast. https://www.mimecast.com/blog/attackers-exploit-the-microsoft-365-outage-with-fake-held-messages-lures/

← Back to articles