Technical SEO
& performance audit
in Pasadena, California

A page built in the browser and a page delivered to a crawler can differ substantially, and the difference is invisible until someone looks at the delivered version.

The mechanism is straightforward. Content assembled by script after the page loads has to be executed before it exists. Search engines do execute script, in a second pass, and that pass is neither immediate nor guaranteed to reproduce what a person with a fast connection sees.

So the first check is the served source. Fetch the page the way a crawler does, without executing anything, and look for the things that matter: the main copy, the headings, the internal links, the structured data. Anything missing from that response depends on a later pass.

Common findings. Product or listing content that only appears after a request finishes. Internal links attached to elements that only become links once script runs, which means the crawler never follows them. Content behind an interaction, where a tab or an accordion loads its contents on click rather than hiding them in the document. Structured data injected by a tag manager rather than emitted by the page.

The fix is not always rendering everything on the server. The question is narrower: which parts must be present in the first response. Main copy, headings, internal links and structured data belong there. Interface behaviour, personalisation and anything that depends on who is looking do not.

Links deserve a specific note. A crawler follows an anchor with an address in it. An element that navigates through script is not that, no matter how it behaves for a person. On sites where navigation is built this way, whole sections are unreachable, and the sitemap hides the problem by listing pages nothing links to.

What we test and report: main content present in the served response, internal links present as real anchors, structured data present and valid without script, and behaviour on a throttled connection. Then the same set after the change, so the difference is a measurement rather than an assurance.

For teams in Pasadena, California, technical SEO begins at ~$600 (audit plus fixes) and goes up to ~$1,500 for the complete setup. Expect three to five weeks for a focused audit and its fix plan, the range this page gives. Mobile parity is among the first checks. The phone version is the one Google indexes.

Google crawls and indexes with a smartphone user agent. Whatever the mobile page leaves out does not count for ranking, even if the desktop page shows it.

Responsive sites mostly pass. Both devices get the same HTML and CSS rearranges it. Content inside collapsed tabs counts as long as it sits in the document.

Parity breaks when the server or a script sends different markup by device. Common findings. A mobile template that trims long copy to a teaser. Breadcrumbs and related-item blocks removed to save space. Structured data present on desktop only. Images swapped for smaller files with no alt text. A separate m-dot site whose pages lack the title of their desktop pair.

Lazy loading is the other gap. Sections that load on a scroll event may never load for Googlebot. It renders with a tall viewport and does not scroll.

Navigation deserves its own look. A hamburger menu whose links load only on tap means the mobile page carries far fewer internal links than desktop. Google sees the thinner graph. Keep menu links in the HTML and let CSS hide them.

Directives must match too. A noindex or canonical set on the m-dot version and missing on desktop, or the reverse, splits the pair in a way that is hard to diagnose from one side.

Tooling changed. Search Console retired its Mobile Usability report and the Mobile-Friendly Test at the end of 2023. We run these checks with Lighthouse and a real mid-range phone. A full-screen popup that covers the content right after a visitor arrives from search is one more thing we flag.

We compare both versions on sample URLs from each template. Word count, headings, internal links, structured data, alt text, meta tags. Each difference goes into the report with the component that causes it.