THOUGHT:

All your eggs in one basket

Recent outages at Microsoft, Google and Meta have exposed how much organisations depend on a handful of technology suppliers. What happens when one of them fails?

Imagine arriving at work to discover that your email has stopped working, your calendar is unavailable and some of the tools you rely on every day are behaving unpredictably.

That’s what happened on Thursday 12 November 2026, when a global Microsoft 365 outage disrupted Outlook and affected features across other services, including Teams, SharePoint, Copilot and Defender XDR.

For some organisations, it was an inconvenience. For others, including parts of the NHS, it disrupted communications, patient referrals and appointments.

The outage raises an uncomfortable question about technology supplier dependency. How much of your organisation relies on a single supplier, and what happens when that supplier stops working?

Three outages, three different risks

Microsoft isn’t the only major technology supplier to experience problems this year. Google and Meta have also suffered outages, each illustrating a different risk.

When your everyday tools share a supplier

Many organisations buy email, calendars, video calls, file storage and security tools as one package. It’s simpler to manage, often cheaper and allows staff to work across connected systems.

But that convenience comes with a risk.

On 12 November, problems with Microsoft’s infrastructure disrupted several services at once. The company reported issues with systems responsible for handling secure connections and said there was no indication of malicious activity.

Not everything stopped working, however. According to Microsoft’s incident reports, Teams’ core messaging and meetings remained largely functional, although calendar and scheduling features that depended on Exchange Online were affected.

That partial functionality offers a useful lesson in how digital services can be designed to cope with problems. We’ll come back to it later.

When your suppliers share a supplier

On 20 October 2026, a power failure in a Google Cloud data centre zone in London disrupted services used by organisations across the UK.

Among those reported to have been affected were Lloyds Bank, HSBC, Royal Mail, Rightmove, DPD and Revolut. Some customers couldn’t access banking services, make payments or track deliveries.

Most had no direct relationship with Google Cloud. They were using services provided by companies that depended on Google’s infrastructure.

An organisation might use several technology suppliers and assume that offers some protection. But if those suppliers rely on the same underlying infrastructure, a single failure could affect them all.

When your customers depend on one platform

On 12 June 2026, Facebook, Instagram and Messenger experienced widespread disruption. Meta’s advertising tools were also affected, preventing some businesses from managing campaigns with problems lasting about two hours.

For many businesses, these platforms are where they advertise, communicate with customers, handle enquiries and generate sales.

If they’re your primary marketing channels, losing access to them can cut off your route to customers. And unlike your own website or customer database, you have little control over when they’re restored.

Tech suppliers outages abstract image

When a supplier comes under attack

Technical failures aren’t the only threat. A cyberattack on a shared supplier can have similar consequences, but disruption can last considerably longer.

On 3 June 2024, the Qilin ransomware group attacked Synnovis, a pathology company providing blood-testing services to NHS organisations across south east London.

The hospitals themselves hadn’t been hacked. But because they relied on Synnovis, the attack had an immediate impact on patient care.

NHS England reported that 10,152 outpatient appointments and 1,710 elective procedures were postponed. King’s College Hospital subsequently identified the cyberattack as a contributing factor in a patient’s death, although it couldn’t determine whether it had directly affected the outcome.

The consequences extended well beyond the initial disruption.

In April 2026, an investigation by The Record found that South London and Maudsley NHS Foundation Trust was still operating without fully restored pathology systems. As of January 2026, the trust estimated that entering 161,560 pathology reports into patient records had been delayed.

Almost two years after the attack, the effects were still being felt.

Whether the cause is a technical failure or a cyberattack, organisations need to know how they’ll operate when a service they depend on becomes unavailable.

The cost of using multiple suppliers

The obvious answer is to use more suppliers. But that’s not necessarily the right one.

Every additional supplier means another contract, more systems to secure and potentially another organisation holding your data. Maintaining multiple services increases costs and complexity.

Larger technology providers also invest heavily in security and recovery. Moving away from them doesn’t automatically make an organisation safer.

Having a backup isn’t the same as being resilient.

An organisation probably doesn’t need two email platforms. It might, however, need an independent way to contact staff and customers when email is unavailable.

Keeping a separate, usable copy of essential information could allow work to continue during an outage. A business that depends on social media might benefit more from building its own mailing list than maintaining accounts on additional platforms.

Sometimes the answer is a technical backup. Sometimes it’s a documented manual process that staff know how to follow.

What matters is whether an organisation can continue providing essential services when something fails.

Having a backup isn't the same as being resilient.

Build for failure, not just success

When we design digital products, we spend a great deal of time thinking about how people will use them. We map user journeys, test interactions and work out how different systems communicate.

But what happens when one of those systems stops responding?

A website might rely on an external payment service, an app on a third-party API, or a digital health platform on data exchanged with several other systems.

Each connection creates a dependency, and decisions made during design and development can determine how much disruption an outage causes.

There’s a principle in software design called graceful degradation. Rather than allowing an entire product to fail because one component stops working, it continues to provide whatever functionality it can.

The Microsoft outage illustrates the principle. Teams’ core features remained largely operational even though some functions that depended on Exchange Online were unavailable. We can’t know from that alone whether this was deliberately designed as graceful degradation, but it shows how a service can remain useful when some of its dependencies fail.

The same thinking applies to websites and apps.

If an online shop’s recommendation engine stops working, customers should still be able to browse and buy products. If a digital health app loses access to an external service, it might still display previously downloaded information while explaining which features are unavailable.

Of course, not every service can continue safely. In digital health, displaying outdated or incomplete clinical information could be worse than temporarily disabling a feature. Sometimes stopping is the safer option.

These decisions shape both the technical architecture and the user experience. People need to understand what’s happened, what they can still do and whether they need to take action.

Designing with dependencies in mind also makes products easier to adapt if a supplier changes its pricing, withdraws a service or needs to be replaced.

Five questions to ask about your own suppliers

You don’t need a complicated disaster recovery strategy to start thinking about supplier dependency. These five questions are a useful place to start.

  1. What would stop working if a critical supplier went down? Which services and business processes depend on it, and who is responsible for managing them?
  2. How would we contact staff, clients and customers if our usual systems were unavailable? Do we have an independent alternative?
  3. Do our suppliers share the same underlying infrastructure? Could one failure affect several services at once, and do we understand those dependencies?
  4. If one platform handles most of our marketing or enquiries, how else could people reach us? Can we contact customers outside that platform?
  5. Could we continue working without our essential systems? Do we have accessible backups, alternative processes and a plan that staff have actually tested?

The answers might reveal a need for new technology, or simply a change in process or a better understanding of existing systems.

Preparing for the inevitable

Microsoft began restoring services following its November outage, although disruption continued into the next day. The consequences of the Synnovis attack were still being felt more than 18 months later.

Both incidents exposed the risks of depending on services over which organisations have limited control.

We can’t prevent our suppliers from experiencing problems. But we can understand those dependencies, make informed decisions about them and design digital products that cope better when things go wrong.

It’s much easier to have those conversations while everything is still working.