When the Line Goes Down
By Matt Collis
A Thursday afternoon
Late on the afternoon of Thursday 12 November, Outlook stopped working for a large part of the world. Within about an hour outage reports were coming in from 74 countries, Teams and Copilot followed, and for many firms email, calendars, chat and the CRM that sits on top of them all went at once. Microsoft has said the cause was not a cyberattack, and its preliminary incident report points to three separate problems lining up, starting with an automated failover that moved Teams traffic onto gateway servers shared with Exchange Online until those servers were overwhelmed. Services were not fully restored until the morning of Saturday 14 November.
The stories that came out of the financial sector were telling. One securities firm in Taiwan was reported to have gone back to fax to confirm trades, and UK advisers described losing email and their CRM at the same moment because both ran on the same stack. None of this was a sophisticated failure, and that is exactly the point.
Closer to home
I have some sympathy, because I had my own version of this earlier in the autumn. The fibre connection to my home office failed, and the fault turned out to be a part worth about £1. Replacing it took two weeks, and in the end it took a direct email to the chief executive of Openreach to get the repair done, because the standard fault route was going nowhere. For a fortnight a £1 component put the connection back in the dark ages, but the work did not go with it. I had Starlink installed and running within the first 24 hours, which restored connectivity while the fibre fault worked its way through the system, and I am keeping it in place as a permanent backup.
The customer service side was just as instructive. Sky's team tried their best, but they could only relay what came back from Openreach, which customers cannot deal with directly, and every 48-hour update told me when the next update would arrive rather than when the line would be fixed. Each one was technically a communication, but none of them helped me plan, and that is exactly the experience a client has when a firm reports on process rather than giving them something they can act on.
The work carried on and deadlines were met, but the experience was more instructive than any tabletop exercise I have run. The things that mattered were not in a policy document. They were whether a second connection could carry a full day of video calls, whether every file I needed was reachable from somewhere other than the home office, and how quickly I could tell clients that nothing had changed for them. It was also a lesson in escalation, because the fix only came once the problem reached someone with the authority to act on it. Having run operations at a FTSE-listed wealth manager, I have written plenty of business continuity plans, and two weeks without fibre was a useful reminder of the gap between the plan on paper and the plan on a Tuesday morning.
It was not the first time a single cut has taught me that lesson. Earlier in my career, an office I worked in had two separate fibre lines, deliberately routed to leave the building in opposite directions so that no single incident could take out both. Both went down on the same day, when a contractor working on the Underground cut through them, because out of sight the two routes met again in an old tube tunnel, and that was exactly where the cut was made. On paper we had diverse connectivity, but underground we had a single point of failure.
What the outage exposed
My two weeks and Microsoft's two days were different kinds of failure, and the difference matters. Mine cut the route to the work, but every application, file and client record was still there once Starlink was up. Microsoft's left connections working perfectly well and took away the applications themselves, so no backup line could help. When the apps a firm depends on go down at the same time as its access to data and files, it is a double whammy, with no way in and nothing to reach, and a continuity plan that only covers one of those layers covers half the problem.
Beyond that distinction, the two events share the same lessons, and they apply to firms of any size.
- Concentration sits in the plumbing. Most firms know they depend on their platform, custodian or DFM. Fewer have mapped how much of their day depends on a single productivity suite, where email, calendar, chat, document storage, identity and often the CRM all fail together. Even arrangements that look diverse on paper, like two fibre lines leaving in opposite directions, can share a single point of failure further down the chain.
- The communication channel is usually the first casualty. A continuity plan that tells staff to email the incident team does not work when email is the incident. Firms need a way to reach staff, clients and counterparties that does not share a supplier with the thing that has failed.
- Recovery is slower than the outage. Microsoft's own report described servers returning to service, then failing again once traffic came back, and a £1 part on my own line took two weeks and an email to the top of Openreach to replace. Plans built around a few hours of disruption need to be tested against something longer.
- People are part of the infrastructure. Hybrid working spreads the risk across home connections, personal devices and mobile coverage, which is a strength if it has been planned for and a weakness if it has not. In a hybrid firm, a single home broadband line can matter as much as the office network.
- Managing expectations is part of the fix. When something fails, clients need to know what is affected, what they can still do, and a realistic estimate of when normal service will return, with an honest update if that estimate moves. An update that only promises another update does not meet the Consumer Duty consumer support outcome, and a poorly communicated outage is how complaints start, often long after the service itself has recovered.
Where the regulation sits
The FCA's operational resilience rules in SYSC 15A have been fully in force since March 2025, requiring in-scope firms to identify important business services, set impact tolerances and show they can stay within them under severe but plausible scenarios. On 11 November, the day before the outage, the FCA published further observations from firms' operational resilience self-assessments, and it has previously identified failures at unregulated third parties as a leading cause of operational disruption across the firms it regulates.
Many smaller advice firms, DFMs and wealth managers sit outside SYSC 15A, but they do not sit outside the problem. The general business continuity requirements in SYSC 4 still apply, and the Consumer Duty consumer support outcome does not pause because a supplier is down. A client who cannot reach their adviser in the week a pension decision needs making is experiencing a service failure, whatever the cause.
What to check this quarter
I would suggest five practical steps, none of which needs a large budget.
- Map the single points of failure. List the services the firm runs on, including the productivity suite, CRM, platform access, telephony and broadband, and mark which ones fail together. Treat connectivity and applications as separate layers, because a backup for one does nothing for the other, and make sure the client contacts and records needed to keep serving clients can be reached if the main suite is down.
- Set up a channel that does not share a supplier. Keep an up-to-date contact list for staff, key clients and counterparties outside the main email system, and agree in advance which channel the firm will use when the usual one is down. Draft the first client message now, so that on the day it says what is affected and what clients can still do, not just that the firm is looking into it.
- Give every key person a second route to work. That might be a mobile data backup, a satellite connection, a second broadband provider or an agreed alternative location, and it should be tested with a full working day rather than a quick connectivity check.
- Test for longer than feels comfortable. Run at least one scenario that lasts days rather than hours, and record what broke, because that record is the evidence a regulator or a client will want to see.
- Ask your suppliers the same questions. Platforms, DFMs and outsourced administrators all sit on someone else's infrastructure, and their continuity arrangements are part of yours. Know the escalation route for each critical supplier before you need it, because the standard fault queue is not always the fastest way to a fix.
The bottom line
Last week's outage did not reveal anything new, but it did remind a lot of firms that resilience is not proved in the board pack, it is proved on an ordinary working day when something basic stops working. My own two weeks without fibre taught me more about how I really operate than any exercise, and the firms that came through last Thursday well will be the ones that had already asked, for each part of the business, what happens if this is not there tomorrow.
This article is commentary and does not constitute advice on any firm's circumstances. Trescore Advisers is a consultancy and is not authorised to give investment advice.

