The Boring Mechanics of a Website That Lasts

It is November 2026, the sky is the colour of a switched-off television, and I am doing the annual once-over of this website. No new fonts. No redesign. Just a mug of tea, a notebook, and a list of jobs so dull they would make a spreadsheet blush. This is the website equivalent of checking the tyre pressure and topping up the screen wash.

The strange thing is that I have come to rather enjoy it. A one-person site is a small machine. It does not ask for much, but it does ask for attention. The less dramatic the attention, the more room there is for the writing, the side projects and the half-decent ideas that the site exists to share in the first place.

Start with the domain, because everything else is decoration

I check the .co.uk registration first. Is auto-renew on? Is the card attached to it still in date? Have I got the registrar login details somewhere sensible, rather than in a browser tab from three laptops ago? I also glance at the SSL certificate and the hosting renewal date. A few years back a friend lost a domain because a payment card expired quietly in the night. I have not forgotten it.

This takes ten minutes. It is also the difference between a website and a dead link with a nice memory attached. Nominet, the UK registry, has decent guidance on keeping domain details current, and I would rather read it now than panic about it later.

Backups are the least interesting insurance

Automated backups are lovely. Untested backups are a rumour. So once a year I download a full export, put it somewhere off the hosting account, and actually try to open a couple of files. Posts, images, the lot. I am not running a bank here, but I would like to be able to rebuild the place if a server had a bad day.

My system is not sophisticated. A zip file, an external drive, and a note in the calendar. The note is the important bit. The drive does not remind me to use it.

Broken links are editorial dead ends

I run a link checker and then work through the red bits. Some links have rotted because a service shut down. Some point to an old page of mine that I renamed without thinking. A broken link is not just a technical hiccup; it is a reader hitting a wall while you are mid-sentence. I fix what I can, redirect what I cannot, and remove the rest.

The same pass is a good excuse to read a few old posts. Sometimes I spot a sentence I would now write differently. Sometimes I spot a project I had forgotten. Maintenance and editing are closer cousins than they look.

Privacy basics without the panic

I am not a lawyer, and this is not legal advice. But I do run a small UK site, which means I should know what data it collects and why. I check the contact form, the newsletter sign-up, and the analytics. If I am using cookies that are not strictly necessary, I make sure the consent side is honest and easy to understand. If I am not, I try not to add a banner just for the sake of looking official.

I also reread my privacy page and update the date if anything has changed. The Information Commissioner's Office has plain-English guidance for small organisations, and it is worth a skim once a year. The aim is not to build a fortress. It is to be straightforward with people who trust the site enough to get in touch.

Accessibility is a maintenance job, not a launch job

It is easy to think about accessibility when a site is new and then forget it forever. I go through the basics: heading order, image alt text, link text that makes sense out of context, colour contrast, and whether the forms can be used with a keyboard. I check a few pages on my phone as well, because that is where a lot of readers will meet the site first.

The point is not to pass an audit and frame the certificate. The point is not to make readers work harder than necessary. A small site can do this well. It mostly requires remembering to look.

The analytics that actually matter

I do not need a dashboard that makes me feel like a minor television executive. I need to know which pages people find useful, which search terms bring them in, and which pages are throwing 404 errors. Search Console is good for the last two. A simple, privacy-friendly analytics tool is enough for the rest.

I look for patterns, not applause. If a page gets no traffic but matters to me, it stays. If a page gets traffic and is embarrassingly out of date, I update it. If something is broken, I fix it. That is the whole strategy.

The final ten minutes

I write down what I changed, set a reminder for next November, and shut the laptop. The site is not a monument. It is more like a bicycle: it stays useful because someone checks the brakes now and then. The checks are boring. The point of them is not.

When the dull bits are done, the good bits get a cleaner run. That is the trade I am happy to make.