03What “properly built” actually means in 2026.

Ask ten web studios what “properly built” means in 2026 and nine of them will answer with a framework. Next.js. Astro. Remix. SvelteKit. As if the question was about tooling.

It isn’t. It never was.

A properly built website in 2026 is a surprisingly boring list of things. Most of them were already true in 2016. The honest, unglamorous standards are about ownership, speed, accessibility, and longevity — and the framework is the least interesting part of any of them.

Here’s what each one actually means, in plain English.

1. Ownership: you hold the keys

A site is not properly built if you don’t own it outright. Ownership, in practice, is four things — and if any of them are missing, the studio has kept a leash on you.

• The domain.Registered to you, in your account, with your billing details. Not the studio’s registrar account with your name in the “Contact” field.
• The source code.In a repository you can access — your own GitHub, GitLab, or a zip file handed over on day one of launch. Not trapped inside a proprietary builder you can’t export from.
• The hosting.On an account you control, with a platform you can leave. Vercel, Netlify, Cloudflare, or your own VPS — all fine. A bespoke studio platform that only deploys your site for you — not fine.
• The analytics. Set up under your login. You should be able to log in and read your own traffic without asking.

This is unglamorous territory. It is also, by a wide margin, the thing most sites get wrong. A studio that keeps the domain or the repo is keeping leverage, not security. Leverage is for negotiations, not for clients.

If you can’t leave the studio without the site going dark, you don’t own it. You’re renting it from them.

2. Speed: measured, not promised

Every studio claims their sites are fast. Very few can show you a number.

A properly built site in 2026 passes Google’s Core Web Vitals on a real 4G connection. That’s not a vanity metric — it’s a ranking signal, a conversion factor, and the thing your visitor silently judges you on in the first 1.5 seconds of loading the page.

The three numbers that matter:

A Lighthouse performance score of 95+ on mobile is the rough benchmark. If your studio can’t show you those numbers at launch, the site is probablynot fast; it’s just not measured.

3. Accessibility: the legal and humane minimum

Accessibility used to be spoken about as a nice-to-have. In 2026 it’s neither optional nor discretionary. The Web Content Accessibility Guidelines (WCAG 2.2) are the working standard, and in the EU the European Accessibility Act is making them enforceable for many commercial sites from June 2025 onward. In the UK, public-sector sites are already required to conform.

A properly built site in 2026 meets WCAG 2.2 at Level AA, which in practice means:

• Every interactive element is reachable and usable with a keyboard alone.
• Text meets a 4.5:1 contrast ratio against its background.
• Every image has meaningful alt text, or is marked as decorative.
• Every form input has a label wired to it — not a placeholder pretending to be one.
• Focus states are visible, not stripped out to “look cleaner”.
• Motion and animation respect prefers-reduced-motion.
• The whole thing works with a screen reader, not just a mouse.

None of this is hard. All of it is routinely skipped, because it doesn’t photograph well on Dribbble. Properly built means shipping it anyway.

4. Longevity: the two-year test

The final standard is the simplest and the hardest. A properly built site in 2026 passes the two-year test: can you leave it alone for two years, and will it still work?

Two years is long enough for the following things to happen:

• Your CMS has released four major versions.
• A dozen of your plugins are deprecated, sold, or abandoned.
• Your hosting platform has changed its pricing twice.
• Browsers have deprecated three APIs you were quietly depending on.
• You have forgotten the password to at least one critical account.

A site that requires monthly attention just to keep workingis not properly built — it’s a commitment dressed up as an asset. The two-year test favours simplicity. Fewer plugins. Fewer third-party runtimes. Static output where possible. Boring, stable, well-documented tools.

This is why most of the sites I build come out the far side looking less technologically impressive than they could have. A properly built small-business site in 2026 should, honestly, look a bit like a site from 2019 with better typography. That’s a feature.

What “properly built” is not

Worth being blunt about the opposite column too.

1. It is not a framework choice.Next.js, Astro, plain HTML, Eleventy, a well-set-up Kirby instance — all can produce properly built sites. A fashionable framework with a bad build is still a bad build.
2. It is not a motion reel.Scroll-hijacked hero sections, GSAP-heavy animations, and cursor-following blobs are not craft. They’re decoration, and usually a tax on the four things above.
3. It is not a builder lock-in.Wix, Squarespace, and most “agency-built” WordPress themes share a common property: the asset is half-owned, the speed is mediocre, and you can’t leave cleanly. That’s a rental, not an asset.
4. It is not a CMS with 47 plugins. Every plugin is a future vulnerability, a future breaking update, and a future day of your life spent troubleshooting instead of running your business.

The unglamorous summary

Properly built, in 2026, comes down to a checklist that would have fitted on a napkin in 2016:

• You own the domain, the code, the hosting, the analytics.
• It’s fast, and you have the Lighthouse screenshot to prove it.
• It meets WCAG 2.2 AA — because it’s the law, the kindness, and the SEO all at once.
• It will still work if you ignore it for two years.

If a studio can’t answer “yes” to every one of those at handover, the site is not properly built. Framework unimportant. Pretty visuals unimportant. The four above, non-negotiable.

The craft is in the boring bits. It always was.