Cloud Lock-In Is Your Problem Now: Three Settings to Change This Week

Archi3 min read

cloudprivacybackup

Contents
  1. The Lock-In Argument Has the Wrong Audience
  2. Setting 1: Turn on a second copy that does not live in the same account
  3. Setting 2: Move domains and email off the same vendor as your website
  4. Setting 3: Write down what you would need to rebuild
  5. Why This Beats Waiting for Better Policy

The Lock-In Argument Has the Wrong Audience#

Most writing about cloud lock-in is aimed at CTOs and procurement teams. It talks about egress fees, proprietary databases, and multi-year commitments. That framing is fine for people signing seven-figure contracts. It is useless for the rest of us, the people who store family photos in one provider, run a small business site on another, and let a third hold every email they have ever sent.

The practical question for ordinary users is not "which cloud is most open." It is "what happens to my stuff if this account closes, this provider raises prices, or this service shuts down." The answer, for most people, is: nothing good, because they have never tested it.

My view is simple. Treat every cloud service as a rental with a notice period you have not read. Then change three settings this week so the notice period matters less.

Setting 1: Turn on a second copy that does not live in the same account#

Most providers offer some form of export or download. Very few users ever run it. A backup that sits in the same account as the original is not a backup. It protects against a deleted file. It does not protect against a locked account, a billing dispute, or a provider that changes its terms.

The fix is unglamorous. Pick the one service holding data you cannot recreate (photos, documents, message history) and set up a copy somewhere else. It can be an external drive, a second provider, or a home server. The point is that the copy is reachable without logging into the first account. Run the export once. Time it. If it takes four hours, that is information you want before an emergency, not during one.

Setting 2: Move domains and email off the same vendor as your website#

Small businesses get hurt here more than they expect. The website host, the domain registrar, and the email provider are often the same company because it was easy at signup. When that company has an outage or a billing error, the business loses its site, its incoming mail, and its ability to redirect either one.

Separating the registrar from the host is a 30-minute job and it is reversible. It also makes future moves cheaper, because the domain is not held hostage by a hosting contract. The same logic applies to personal email: if your address is tied to the same account that holds your photos and your phone backup, one lockout takes everything.

Setting 3: Write down what you would need to rebuild#

This is the setting that is not a setting. It is a note, kept outside the cloud, listing the accounts that matter and what each one holds. Not passwords, just the map: which provider holds which data, which email address is the recovery address, which payment card is on file.

People who have done this describe the same reaction. The list is longer than they expected and a few entries are services they had forgotten. That is the point. You cannot reduce dependence on something you have not counted.

Why This Beats Waiting for Better Policy#

There is a reasonable hope that regulation will force easier data portability. Some of it already exists in privacy law, and more is likely. But portability rules tend to cover the formal right to export, not the practical cost of doing it at short notice. A provider can comply with an export requirement and still make the process slow enough that most users never finish.

Ordinary users should not wait for that fight to resolve. The three changes above cost an afternoon and reduce exposure to the most common failure: an account problem that becomes a data problem because there was only ever one copy.

The industry will keep debating lock-in at the contract level. For everyone else, lock-in is measured in whether the export actually opens on a machine you control. Test that this week. If it fails, you have learned something useful while it is still cheap to fix.

More articles