Email outage in a business: what to do? What the Microsoft incident of 31 August 2026 teaches a small business

On 31 August 2026, late in the day, messages stopped going out. Not all of them, not everywhere, and with no readable explanation on screen. A quote sat waiting in the outbox. A search through a client's history returned nothing. An attachment refused to download. In a twenty-person company, nobody thinks first of an email outage at the provider: the first reflex is always the same, you suspect the connection, the workstation, or yesterday's update.

It was neither the connection nor the workstation. Microsoft opened two incidents and, according to the published milestones, the episode ran into days, not hours. For an owner the lesson is not technical. It is organisational: how many of your company's decisions travel through a single channel, and what happens exactly the day that channel stops?

An email outage is not an IT incident. It is an interruption of your organisation, and it is prepared for as such.

Email outage in a business: what to do when nothing is going out?

When the company mailbox stops responding, do three things in order: check whether the outage is general or local, warn your teams and your clients through an independent channel, then write down what will have to be sent again. The rest waits. An outage at the provider cannot be repaired from your desk.

That last sentence is the hardest to accept. The instinct, when a tool stops, is to act on the tool. But in a platform incident, every minute spent restarting or changing a setting is a minute lost, and it sometimes creates a second problem once the service is back. The only ground you control is how the day is organised.

What exactly happened between 31 August and 2 September 2026?

The incident began on 31 August 2026 at 18:30 BST, that is 17:30 UTC, according to the running coverage published by BleepingComputer. Microsoft opened two incident identifiers: EX1464935 for Exchange Online and MO1465074 for the multi-service incident. The milestones, in their original time zones, with the local equivalent for the start of the incident:

The running coverage from BleepingComputer lists services that go well beyond email: besides Exchange Online, Microsoft Teams, OneDrive for Business, SharePoint Online and the Microsoft 365 admin centre were affected, along with several technical building blocks reserved for administrators. The symptoms reported were delays and failures when sending and receiving, mailbox search failures, authentication errors, failed mailbox operations, attachments that could not be downloaded and connection problems with Outlook on the web.

How do you know whether the problem is yours or the provider's?

This is the first operational question, and it is settled quickly. Three checks are enough, with no particular skill.

• Check several machines. If the same symptom appears on different computers, different connections and for people working from home, the local hypothesis collapses.
• Check the Microsoft 365 admin centre, where incident notices live. Identifiers such as EX1464935 and MO1465074 can only be consulted there; they point to no public page and you will not find them in a search engine.
• Read public indicators with care. Downdetector aggregates user reports worldwide. The figure of more than 5,000 reports recorded on 31 August at 8pm Paris time by Clubic is a signal of simultaneity, not a measure of impact in one country.

The conclusion is not a technical diagnosis but a decision: either you keep looking on your own side, or you switch into organisation mode.

What does the business owner do within the hour?

An hour is short, which is why the order of play is decided in advance.

1. Warn people internally through an independent channel. The 31 August incident hit Exchange Online and Teams at once, so the fallback must not depend on the same provider: telephone, text message, or an external instant messenger.
2. Identify today's deadlines. A tender response, a confirmed appointment, a contractual document expected before a given date: those justify a phone call. The rest can slip by a day.
3. Warn the clients concerned, by telephone. A thirty second call explaining that your sends are delayed beats a silence read as negligence.
4. Keep the written list of missed sends, on paper or in a local document. It is the centrepiece of the recovery: without it, you reconstruct from memory.
5. Change nothing. No resending in a loop, no setting changed, no new installation. Sends retried ten times during an outage tend to go out ten times once the service is back.

How do you pick the day back up when the service returns?

The return to normal is the most underestimated moment. The service comes back, everyone rushes to their mailbox, and the accumulated load produces exactly what the outage had failed to cause: things slipping through.

Recovery follows a precise order. First your list of missed sends, because it holds your commitments, not other people's. Then the outbox, to check what actually went out and spot duplicates. Finally the inbox, handled by sender rather than by order of arrival, starting with those from whom you were expecting an answer.

Block out a dedicated slot rather than scraping ten minutes between two meetings. It is the principle set out in our method for piloting your inbox in 2026: a pile of messages is handled in decided batches, never in scraps. Our article Optimize Your Email Inbox: 10 Methods to Stop Wasting Time covers the routines for afterwards.

What does this episode say about your company's dependence?

It says something simple, which many owners discover in the moment: in a small business, email is not one communication tool among others. It is the register of commitments. Quotes, approvals and agreed deadlines travel through it, and often exist nowhere else. When it stops, what is interrupted is the traceability.

Two pieces of work follow, both organisational. The first: knowing in advance who warns whom and with which list of numbers. The second: reducing the volume that piles up in normal times, because a saturated mailbox turns a two day interruption into a lost week.

We should be direct here about what we do and about what we do not do. Neston is an Outlook add-in, so when Exchange Online is unavailable, Neston is unavailable too. We prefer to write it plainly: our assistant does not replace your mailbox, does not back it up, is neither a continuity plan nor a security product, and carries no availability guarantee. It works on the time and the context of daily processing, not on the outage.

Further reading

• Piloting your inbox in 2026: the complete method, the working framework of an owner
• Optimize Your Email Inbox: 10 Methods to Stop Wasting Time, routines that hold up over time
• Email productivity: the 2026 method to be more productive with your email, to structure daily processing
• Writing Professional Emails with AI: Complete Guide 2026, what drafting assistance changes
• Email Writing Score: Rate Your Emails 0-100, how message quality is measured
• The work email account when an employee leaves: what to do, the other situation where access to a mailbox becomes a management matter
• Automatic email & attachment filing for Outlook: complete 2026 guide, the methodical piece to frame inbox organization outside crisis mode

FAQ: email outage and getting back to work

Published 3 September 2026 · Reading time: 9 minutes · approx. 2,280 words