Technical SEO

Technical SEO is best read as a sequence rather than a checklist, because each step depends on the one before it. A page nobody can crawl is never rendered. A page that renders empty is never indexed. A page that is not indexed cannot rank, however good it is.

That ordering is also the right order to debug in, and it is the thing most audits get wrong. A report that opens with image compression on a site whose product pages are blocked in robots.txt has started at the end.

Google crawls a page and then renders it, and those are two passes rather than one. The render can lag the crawl, and it runs on a budget. Content that only exists after client-side execution may be seen late, seen partially, or missed on a given pass.

For a marketing site with server-rendered content and some interactive extras, this rarely matters. For a site built as a single-page application it is not a detail, it is the whole question of whether the site is indexable, and it is worth establishing in the first hour of any engagement rather than the fourth month.

The check is cheap and agencies skip it anyway: fetch the page with JavaScript disabled, or use the URL Inspection tool in Search Console and read the rendered HTML it reports. If the product description is missing from that, it is missing from the index.

The fix is usually a framework decision rather than an SEO one, which is exactly why it has to be raised early. Moving a site to server-side rendering in month four is a rebuild; specifying it in month zero is a line in a brief.

Google has indexed mobile-first since 2019 and finished the transition for all sites by 2023. The practical meaning is blunt: the mobile rendering of a page is the version that gets indexed and ranked, and the desktop version is not consulted.

So anything present on desktop and absent on mobile is absent, full stop. Content collapsed behind a tab is generally fine, since it exists in the markup. Content not rendered on small screens at all is not.

The versions that catch teams out are the quiet ones. A responsive layout that drops a block of copy below a breakpoint. Structured data emitted only in the desktop template. Images that become decorative placeholders on mobile, taking their alt text with them.

The check is to compare the two renderings of the same URL rather than to assume that responsive means identical. Responsive means the layout adapts, which is not the same claim.

Technical SEO has a shape unusual in agency work: it is the highest-leverage category and the one where a single mistake is catastrophic rather than merely disappointing. A misplaced noindex, a staging robots.txt shipped to production, a canonical pointing at the wrong template, and a site can leave the index within days.

It will also be your change, in a log, with a date on it. That is worth being clear-eyed about rather than anxious about, and it argues for a specific way of working: staged rollouts, a check after deployment rather than only before, and Search Console coverage watched for the fortnight after any structural change.

The second risk is commercial rather than technical. Most of these fixes need developer time the client controls, and an agency that scoped implementation it cannot perform ends up accountable for work it was never allowed to do. Scope the audit and the implementation as separate things, with the second conditional on access.

And prioritise the audit. A two hundred row spreadsheet sorted by crawler severity is not a plan, it is a data export. Three items marked as the ones that matter, with the rest listed as backlog, is the version a client developer will actually start.