Skip to main content

Outlook Went Down. You Rent Your Systems, but You Own the Plan.

Gabe Hilado
Founder and CEO, Zenpo Software Innovations

I'll admit it. Zenpo is a small company, and I had no backup plan for an email outage.

We had no deadline the day Outlook went down. If we had, we would have done phone calls. If the phones had been out too, that's a bigger problem.

So we got lucky. The fallback was whatever we would have improvised that day.

What makes this a little embarrassing is that I wrote about continuity of operations planning back in 2010. A plane had just been flown into an IRS building, and I couldn't stop wondering whether they could recover their systems. I wrote that government agencies call it Continuity of Operations Planning, or COOP, and that the IT side of it mostly comes down to recovering servers, applications and databases. Have backups, and rehearse the restore.

That advice assumed you owned the server. In 2010, a lot of organizations still did.

Renting vs owning your business systems​

Back then, owning was an option, even for a small company. You could buy a server, put Exchange on it, and run your own email. You could run your own file server and your own accounting package on a machine down the hall.

Owning meant maintenance. Patches, backups, failed drives, the restore you were supposed to rehearse and probably didn't. When something broke, it was yours to fix, and that was the burden and the control in one package.

Now you rent. Email, files, chat, phones, accounting, sign-in. A small company can rent every one of them, and no maintenance comes with it. Somebody else patches the servers. Somebody else replaces the drives. That's the reason to rent.

Renting also took away the option to fix it yourself. You can't restore Exchange Online. You can't fail it over to another data center. When the provider's servers go down, you wait for the provider, same as everyone else on that service.

Thursday was a good example of what that waiting looks like.

What happened in the November 12 Microsoft 365 outage?​

Reports started coming in around noon Eastern on Thursday, according to Tom's Guide's live coverage. Microsoft's first status post said it was investigating issues "accessing Exchange Online mailboxes through multiple connection methods in the Americas," as NBC Chicago reported. Later that day it said the issue "may affect Microsoft 365 service connectivity for users in any region." Downdetector reports for Outlook peaked at just over 130,000 in the US, according to NBC Chicago.

Microsoft said the load on its infrastructure "isn't driven by an increase in traffic, but by a number of servers being unexpectedly unavailable." By Friday, it was telling customers the incident "was introduced by an issue during an earlier network capacity upgrade" that "surfaced following a recent network change," and that it was "a technical issue not related to cyber or AI / agents," per ABC News.

It reached well past office inboxes. New York City's Emergency Management Department warned that people might see delays reaching city agencies. Al Jazeera reported disruptions to government systems in Singapore and Japan and to communications at the Supreme Court of the Philippines. ABC News described it as an outage that spanned two workdays. On Friday, Microsoft was still saying, "We do not have an ETA for full recovery yet."

None of those organizations could fix it. Neither could I. Microsoft has promised a full post-incident report, and the cause is Microsoft's to explain. What each customer did while waiting was its own business, and the only part any of them owned.

What do you still own when you rent?​

The plan. That's the piece of COOP that never moved to the cloud, because it was always about people. Who decides the service is down. How everyone finds out. What work keeps going and what waits.

For each service you rent, I'd write down three answers.

What stops when it's down? Be specific. Not "email," but which work stops: quotes going out, invoices, a customer waiting on a reply, the meeting invites for tomorrow.

How long can you work without it? This changes with the calendar. We had no deadline on Thursday, so a day without email was survivable. With a proposal due at 2 p.m., the same outage is a different problem. Accounting being down for an afternoon is one thing. On payroll day it's another.

What do you switch to? One answer, written down, that everyone knows before they need it. For us it would have been phone calls. That's the improvised version of the right idea.

Then one more question for every answer to the third one: does the fallback depend on the same thing that just broke? That's where my "if the phones had been out too" comes back. If your phones, chat, and email all come from the same provider, or all sit behind the same rented sign-in, one bad day at that provider can take out the primary and the backup together. The fallback has to fail for different reasons than the thing it backs up.

Know what keeps working offline, too. Some rented services leave a copy on your own machines. Outlook on the desktop in Cached Exchange Mode keeps a local copy of the mailbox, so you can still read and search mail that synced before the outage. You just can't send through Exchange until it's back. OneDrive files set to stay on the device are on disk. Find out which of your rented services do something like this, and which leave you with nothing but an error page.

Then decide what you tell people outside. Customers and partners will keep emailing, calling, or logging in, and they won't know why nothing comes back. Pick where you'll post a short note and who posts it, and check that the place you pick doesn't run on the same provider.

Don't forget the information the fallback needs. If the phone numbers for your team, your customers and your vendors live only in Outlook, your phone fallback is locked inside the system that's down. Keep a copy somewhere else.

The plan also has to work when the person who wrote it is on a plane. That's the same test I'd put to any handoff of operational ownership: can the people left behind run it without the author in the room?

How do you test a plan for a rented service?​

My 2010 point was that having backups means nothing until you've restored from one. The rented version is the same idea with a different drill. You can't rehearse fixing Microsoft's servers. You can rehearse the thirty minutes after you find out they're down.

Pick a morning. Tell everyone a service is "down." The person who calls it sends the first message through the fallback. See who doesn't get it, whose number is wrong, and who has no idea where the contact list is. Fix those, then pick a different service next time.

When the service comes back, check what got lost in the gap. Microsoft said it extended its email timeout window to 48 hours to preserve delayed messages. Delayed still isn't the same as read. Someone should go through what arrived late and call back the people who matter.

You rent the service now. Nobody rents you the plan. Write it down for the services your business can't go a day without, starting with the one that went down Thursday.

Was the November 12 Outlook outage a cyberattack?​

No. Microsoft said it was "a technical issue not related to cyber or AI / agents" and that there was no indication of malicious or cyber activity.

What should a small business do when a cloud service goes down?​

Follow the plan you wrote beforehand: one person declares the outage, everyone switches to a fallback that doesn't depend on the failed service, and someone checks what was missed once it's back.

Is renting cloud services riskier than owning your own servers?​

The risk moves. Renting removes the maintenance and the option to fix outages yourself, so the planning for an outage becomes the part you still control.