A single point of failure is anything that, alone, can stop your whole business. Every company has several. Most owners can't name theirs, because single points of failure are invisible right up until they fail. This is the walk-through we do on new environments. You can do most of it yourself in an afternoon.
The walk-through
Go through each item and ask the same question: if this disappeared at 9 a.m. tomorrow, what stops?
- The internet line. One ISP, one cable into the building. Phones, card readers, cloud apps, all of it rides that cable. A backhoe two blocks away is your outage. Fix: a second connection with automatic failover, often $100 a month.
- The server in the closet. One box running files, accounting, or the line-of-business app. Ask when its disks were last checked and where its backups actually live. Bonus points if it shares a power strip with a space heater.
- Power. No UPS, or a UPS whose battery died in 2023 and beeps sadly when anyone looks at it.
- The person. The one who knows all the passwords, the billing system, the weird export routine. People are the most common single point of failure and the least audited. Fix is not cloning them; it's documentation and a shared password manager with emergency access.
- The domain and DNS. Your domain registration sits in someone's personal account with a password nobody knows, and it expires on a credit card that got cancelled. This one takes down email and the website together and takes days to unwind.
- The admin account. One global admin on Microsoft 365 or Google Workspace, no second admin, no recovery method that works. Locked out means locked out.
- The software vendor. The industry app that runs your operation: what happens if the vendor has an outage, or a price change, or an acquisition? You can't always fix this one, but you can know your export options before you need them.
- The building. If the office floods, what's actually in it that can't be elsewhere? This question is what pushed half of continuity planning into the cloud.
Rank by fix cost, not by fear
List everything you found, then sort by how cheap the fix is, because several are nearly free: a second admin account (free), documented passwords (an afternoon), domain in a company account with auto-renew on a company card (an hour), UPS battery (under $100). Do the free tier this week.
The remaining items get the downtime math treatment: cost of the outage times rough likelihood, against cost of the fix. That's the whole method. No framework, no binder, just a ranked list and a calendar.
Redo it yearly
Businesses grow new single points of failure quietly: a new app becomes critical, a redundant thing gets "temporarily" simplified, the second admin leaves the company. The walk-through is an annual habit, and it feeds directly into the continuity plan, which is where the findings go to become instructions.
Want this handled instead of homeworked? That's the job.
Email us →