Technical SEO: The Five Things That Move Rankings

Technical SEO is everything that lets Google find your page, read all of it, and receive it quickly. The usual checklist runs to sixty items. Five of them decide whether a site rises. The rest is hygiene: worth doing, but it does not build rankings on its own.

The five

Crawlability. Rendering. One canonical address per thing. Delivery speed. And a mobile version that is the same page rather than a trimmed one. Every technical problem we have seen genuinely shift a ranking sits under one of those. The rest of the list, alt attributes and title lengths and heading order, matters at a different altitude. A page with messy headings that is fast and crawlable ranks. A structurally perfect page the crawler cannot reach does not.

Crawlability: the thing you assume is open

Crawl problems rarely announce themselves as errors. A Disallow left over from staging. A noindex hardcoded into a theme. Or a page that is only reachable through the site's own search box, with nothing linking to it anywhere. That last one is the most common and no tool warns you about it. Quick test: can you reach the page that matters to you from the homepage in a few clicks? If your answer is "not from the menu, but it is in the sitemap", Google treats that page as peripheral.

Rendering: the page that exists only in your browser

If the main text is assembled by JavaScript after load, a gap opens between what you see and what Google sees. Google does execute JavaScript, but not instantly and not unconditionally; the JavaScript SEO basics page explains the delay. For a shop that builds its catalogue with script, the delay is measured in weeks.

Cheapest diagnosis available: load the page with JavaScript disabled. If nothing is left, you have a rendering problem, not a content problem.

One address per thing

A product reachable through three category paths has three URLs. Add the www and non-www versions, trailing slash variants, and every URL generated by a filter parameter, and a modest catalogue produces thousands of addresses for a few hundred products. Canonical tags exist for this, and Google's guide to consolidating duplicate URLs covers the edge cases.

Note the difference from what happens between two separate articles. When two posts compete for one query, a canonical is not the fix; keyword cannibalization and how to resolve it is a different problem with a different answer.

Speed, the only part your customers also notice

Of the page experience metrics, the one that has mattered most in our projects is LCP, the time until the largest element on screen appears. The official LCP definition lives on web.dev, and we have written up what the page experience metrics actually measure separately.

What we did to our own site, and you can confirm it in the source of this page: every CSS animation and transition was removed, fonts were subset to the characters actually used, critical CSS was inlined, and the analytics tag does not load at all until the visitor interacts with the page. None of that is an SEO trick. It is sending less.

Here is the sentence an agency selling SEO retainers has no reason to write: if your server response time is bad at peak, no amount of browser-side optimisation compensates for it. The hosting side of speed has the numbers. If you would rather hand that part over, our speed work is exactly this.

One more thing that belongs under rendering. Google indexes the mobile version. Anything you hide on small screens effectively does not exist for search. We have found this on sites where a designer removed product descriptions to make the mobile layout cleaner, and nobody connected that to the traffic that stopped arriving.

Security belongs in this conversation

A site that has been compromised usually acquires pages it never created, and Google crawls them. Cleaning the site without cleaning the index leaves the job half done. That is why hardening and clean-up sits next to SEO here rather than in a separate department.

Where to start

The order we use on our own projects: confirm the important pages are indexed at all, confirm Google sees what you see, consolidate duplicate addresses, then work on speed. The free tools cover the first three steps.

And the boundary, stated plainly: technical SEO raises the ceiling. It does not put anything under it. A flawless, fast page that answers nobody's question ranks for nothing, whatever you do to the configuration. That is why our SEO work begins by deciding which pages are worth the technical effort in the first place.

Questions we get asked

Is technical SEO a one-off or ongoing?

The foundation is done once. After that it needs rechecking every time the theme changes, a plugin is added, or the URL structure moves. A theme change is the single most reliable way to break all of it at once.

How much does an SEO tool's score matter?

A perfect score ranks nowhere. That number is a list of things worth checking, not a prediction.

Do I need structured data?

For certain page types, products and FAQs among them, it improves how the result looks and that affects click-through. It does not raise position directly.

Next step: run your most important URL through URL Inspection and read the rendered HTML. If your main text is not in there, put the rest of the checklist aside for now.

Last reviewed: August 2026