How client-side rendering works
The server sends a small HTML file that is mostly empty — a title, a few script tags and an empty container element. The browser then downloads the JavaScript bundle, runs it, requests the data it needs from an API, and writes the real content into that container. Everything the visitor reads is created on their own device, after the page has already arrived.
This is the default behaviour of single-page application frameworks unless the project has been configured to do something else. React, Vue and Angular builds put together with a plain bundler all work this way. The giveaway is the raw page source: if the source is nearly blank while the page on screen is full of text, the browser built it.
Why client-side rendering matters
Google executes JavaScript, so a client-rendered page is not invisible. But rendering is a separate, queued and resource-hungry step that happens after the HTML has been fetched. Anything that exists only once the script has run — headings, body copy, internal links, the canonical tag, the meta description — reaches the index later than it would from a server response, and only when rendering succeeds.
Speed is the second cost. The visitor waits for the HTML, then the script, then the data request, then the paint. On a mid-range Android phone over a mobile connection, which is the normal case for most traffic in Nepal, that chain is felt. It shows up in Largest Contentful Paint and in people leaving long before it shows up in rankings.
Where client-side rendering goes wrong
The failures are rarely as dramatic as “Google cannot see my site”. They are small and easy to miss. A blocked script, or an API that refuses requests coming from a data centre, leaves the crawler with the empty shell. Navigation built as click handlers instead of real anchor tags gives the crawler nothing to follow. Titles written by JavaScript can be captured in their placeholder state. Infinite scroll with no paginated URLs hides most of a catalogue behind a button nobody clicks.
Teams also test the wrong thing. Loading the page in your own browser proves very little, because your browser runs every script and waits as long as it takes. A crawler has neither of those luxuries.
What to do about it
Decide rendering per template, not per site. Pages that have to be found — service pages, product pages, articles, category listings — should arrive as finished HTML from the server. Pages behind a login, dashboards, calculators and configurators can stay client-rendered with no search cost at all, because nobody is searching for them.
Where a rewrite is not on the table, start by checking what the crawler actually receives. The URL Inspection tool in Search Console shows the rendered HTML Google holds for a page, and comparing that against what you see is the fastest honest test available. Then make sure links are anchor tags with a real href, that no script or API endpoint is blocked in robots.txt, and that the title, canonical and meta description sit in the server response rather than being added later. Larger moves, such as switching to server rendering or pre-building pages, belong in a planned technical SEO workstream rather than a rushed fix.
