Why Backups Alone Won't Keep You Operating

Most solo operators back up their files and stop there. What usually ends a one-person operation is losing the way back into an account, for example when one phone holds both the login and the recovery number. I audited my own stack for that and found two gaps.

Most solo operators back up their files and call themselves safe. Backup is the easy half, and it's rarely what ends a one-person operation. What usually ends it is an account the owner can't get back into, or a single device that held both the login and the phone number meant to rescue it.

I run my businesses on a one-person stack: cloud storage, an AI system that does much of the work, a domain and a handful of vendors, all hanging off one identity. One weekend I audited it the way I'd audit a company instead of a laptop. "Are my files backed up?" was the easy question, so I asked a harder one.

Backup protects the files. Continuity protects access, identity and the ability to keep operating tomorrow morning. They're two separate audits, and almost nobody runs the second one on themselves. What I found is worth sharing, because the pattern applies well beyond my setup.

The data was the strong part

Every critical folder copies to two independent cloud services, stored fully on disk with no silent gaps, so if I lose one provider, the other carries the operation. I'd built that part right.

The gaps were all in getting back in. My recovery email lived inside the same provider it was meant to recover, so one account lock could take the account and its rescue path together. My two-factor login had no offline backup either, on a stack where the login device and the recovery phone are the same object in my pocket. One dropped phone would have taken out two recovery channels at once.

No backup, however good, fixes that kind of gap, because the data was never at risk. The risk was in re-entry, and re-entry is the half people skip, since it stays invisible until the day it's the only thing that matters.

The ordinary failure that sets off a chain

I now put it simply: a one-person operation depends entirely on one person and that person's devices. The biggest risk is rarely the dramatic disaster everyone imagines. It's an ordinary failure that sets off a chain, like a broken laptop that also happened to be the login device, or a lost phone that was also the recovery key.

At that point, "is my data safe?" stops being the useful question. The useful one is "if this device disappeared right now, how long would it take me to be fully working again?", followed by "what single point of failure could stretch that from minutes to never?"

The test of resilience is how quickly an operator recovers when an ordinary failure exposes a dependency nobody had noticed. Well-run institutions never assume anything is permanent. They expect people to leave, devices to fail, passwords to be lost and systems to break, and they design reliable routes back to normal working.

A solo operator gets none of those safeguards by default. Spare access, written-down steps, backup login methods and a recovery plan all have to be built on purpose. Without them the gaps stay invisible until the exact moment they become critical, and then the cost shows up as lost time, lost access and lost work.

The same check for a small firm starting with AI

The pattern gets sharper when a small firm starts using AI tools as a team. Each new tool is one more account, and each account has its own way back in. In a hypothetical 15-person firm, the first AI subscription is often set up by whoever was keenest, on their own email and their own phone. Six months later that person leaves, and the firm finds that the admin login, the billing and the saved work all sit behind a phone it no longer has.

The fix is the same afternoon's work as mine. For each AI tool the firm pays for, someone writes down who holds the admin login, which email the recovery goes to, and whether that email belongs to the firm or to a person. Then the firm makes sure at least two people can get back in, and that neither of them depends on the other's phone. The list is usually under ten lines, and it's far easier to write before anyone has left.

So I'd suggest running a company of one, or a company of fifteen, like an institution, and auditing how you get back in as carefully as you audit your backups. The fix usually takes an afternoon, while learning the lesson the other way can cost the whole operation.

If your team is setting up its first AI tools and wants to do it safely from the start, ask about the workshop here.

Pick a time that suits you.

Book a call