It never becomes a project again.

Why did our site break after an update, and how do we stop it happening again?

Because something was updated without anyone checking what it would do. Nearly every break after an update is preventable: read what's changing, back up first, test on a copy, look at the live site afterwards, and fix before applying the next one. The real question is who owns that process. If the answer is "we're not sure", that's the risk.

You logged in on a Tuesday and the quote form was gone, or the homepage was a white screen, or the menu had turned into a list. Nothing had been touched. Except that something had: an update ran, and nobody was watching.

What actually breaks

A site is a stack of pieces that each change on their own schedule. One piece updates and assumes the others have too. A theme assumes an older version of the core software. Two plugins want the same thing done two different ways. The hosting company moves everyone to a new version of the server language overnight. Or someone pressed “update all” with no backup and no look afterwards.

Automatic updates are the biggest single cause, because they optimise for one thing: speed. They don’t check whether the homepage still loads, whether the form still sends, whether checkout still works. A subtle failure can sit there for days until a customer mentions it, and by then you’ve lost the enquiries that didn’t get through.

What stopping it looks like

None of this is complicated. It just has to be someone’s job.

Read first. Every update says what it changes. Someone reads it, and reads what other people found when they installed it.

Back up immediately before. Not last night. A snapshot taken seconds before the change, so undoing it is one click.

Test on a copy. The update goes onto a staging copy of the site first. If the form breaks there, nobody notices but me.

Look at the live site afterwards. A human loads the pages that matter and uses the form.

One at a time. If something’s wrong, fix it before the next update, so you always know which change caused what.

The question to ask whoever looks after your site

“Who owns updates?” Three answers are common. “The host does them automatically”: nobody is checking. “We do them when we remember”: irregular, untested, unbacked. “We’re not sure”: the highest risk of all. The answer you want is a person, a schedule, a backup every time, a test every time, and someone to email when it goes wrong anyway.

What a broken site costs

Emergency developer time at emergency rates. The enquiries that arrived while the form was down and went to someone else. The customer who saw a broken page and decided you were probably broken too. None of it shows up on an invoice, which is why it’s easy to under-insure against.

How I do it

Every site I look after gets exactly the process above, every time, for a fixed monthly amount. Updates are never automatic. There’s a backup before each one and a person after each one. And when something slips through anyway, you email me, not a ticket queue.

further reading: The difference between hosting and care

Ask me about this, or anything else.

talk to the real Warren
Secret Link