Google can't see your content — and your site looks fine

Your site loads, works, and shows everything. To a crawler it may look almost empty, and nothing on screen tells you.

The problem

You ship a site. It looks right, it works, the client signs off. Weeks later they ask why none of the copy shows up in search.

It isn't a penalty and it usually isn't the writing. The crawler didn't receive the page you're looking at.

Why this happens

The server sends an HTML file. On older sites that file already held the content.

On sites built with React, Vue, Angular or similar, the server sends a near-empty shell and the content is assembled afterwards, in the browser, by JavaScript.

To a person that's invisible — milliseconds. To a crawler it isn't. Googlebot can execute JavaScript, but it doesn't always, doesn't always wait, and doesn't do it on every page. What gets stored can be the shell.

The result is a site that exists for visitors and doesn't exist for search.

How to tell

The fast manual check: open the page, right-click, View Page Source. That's the served HTML, before JavaScript. Search it for a paragraph you can see on screen. If it isn't there, the crawler may not have it either.

That works, but it's one page at a time and it won't tell you what else is missing.

That's what SiteMerit does automatically. We open your page in a real browser, run the JavaScript the way a visitor would, and compare what the server delivered against what appeared afterwards. The report lists the text blocks, links, metadata and structured data that only exist after JavaScript runs.

What to do

If the content only appears after JavaScript, three paths, from most complete to simplest:

• Server-side rendering. The server assembles the finished page and ships complete HTML. The most solid fix, and the biggest change.
• Static generation. For content that rarely changes, build the pages once and serve them ready. Simple and effective for blogs, docs and marketing pages.
• Get the essentials into the served HTML. If a rewrite isn't realistic, at minimum the title, meta description, canonical and main copy must be in the file the server delivers.

Measure before you commit. Rewriting a whole architecture over a problem that may affect two templates is expensive for the wrong reason.