What last week’s Microsoft outage taught IT leaders (and what to do next)
Last week’s Microsoft outage was a sharp reminder of an uncomfortable truth: even the most mature cloud platforms can have a bad day, and when they do, the blast radius is global.
For many organisations, the disruption wasn’t just “email is down.” It was meetings that couldn’t start, frontline staff who couldn’t access shared files, and customer-facing teams forced into manual workarounds. Whether you were directly impacted or just watching the incident unfold, it’s worth treating this as a learning moment—not a headline.
What happened (in plain terms)
Public reporting indicates the outage affected Microsoft 365 services, including Outlook and Teams, with flow-on impacts to day-to-day collaboration and productivity. When a suite as deeply embedded as Microsoft 365 stumbles, the impact compounds quickly because so many business processes assume it’s always available.
The key point: this wasn’t a niche product issue. It hit the “core operating system” of modern knowledge work.
Why outages like this feel worse than they used to
Cloud has shifted the risk profile.
In the on-prem world, a failure was often localised: a server, a site, a region. In the cloud world, a failure can be systemic—shared identity, shared control planes, shared dependencies. The upside is scale and resilience; the downside is that when something breaks at scale, it breaks for everyone.
Three factors make these incidents particularly painful:
- Service concentration: Email, meetings, chat, files, identity, and security controls often sit under one vendor umbrella.
- Operational coupling: Business processes are now built directly into collaboration tools (approvals, customer comms, incident response, rostering).
- Human behaviour: Teams lose “offline muscle memory.” When the tools go down, people don’t have rehearsed alternatives.
The real lesson: resilience is a design choice
Most organisations can’t prevent a vendor outage. But you can reduce how much it hurts.
Here are practical moves that matter.
1) Define “minimum viable operations”
If Microsoft 365 is unavailable for 2 hours, what must still happen?
Examples:
- Customer support must still respond to priority cases
- Stores/branches must still take payments and schedule work
- Executives must still receive critical updates
Write this down. Agree it with the business. Then design around it.
2) Build an “outage playbook” people will actually use
A playbook shouldn’t be a 40-page PDF no one can find.
Keep it simple:
- Who declares an outage internally
- Where updates are posted (and how staff access them)
- What comms templates to use for customers and partners
- What the approved workarounds are (SMS trees, alternative meeting links, phone bridges)
3) Don’t let identity be a single point of failure
Many outages become “total outages” because identity and access are tied to the same ecosystem.
Consider:
- Break-glass accounts with tested access paths
- Clear separation between admin access and day-to-day identity dependencies
- Regular validation that emergency access works when core services are degraded
4) Rehearse the failure—briefly, regularly
You don’t need a massive disaster recovery exercise every quarter.
Do a 30-minute drill twice a year:
- “Teams is down. How do we run the incident call?”
- “Outlook is down. How do we contact customers with urgent updates?”
The goal is not perfection. It’s familiarity.
5) Communicate like a product team
During outages, the best internal comms are:
- Fast (even if incomplete)
- Clear (what’s impacted, what’s not)
- Actionable (what staff should do right now)
- Time-boxed (next update at X time)
Silence creates chaos. Over-explaining creates confusion.
A balanced view: cloud is still the right bet
It’s tempting to treat an outage as evidence that “cloud can’t be trusted.” That’s the wrong takeaway.
The better takeaway is this: cloud reliability is high, but not absolute. Your job as an IT leader is to design for the reality that some outages will happen—then ensure the business can keep moving while they’re resolved.
The question to ask your team this week
If Microsoft 365 went down again tomorrow for half a day:
- What would stop immediately?
- What would continue?
- What would we do first?
If those answers aren’t obvious, you’ve got your next improvement backlog.