Does server rendering actually change how a site ranks, measured rather than assumed

Moved a content site from client rendering to server rendering and watched what happened. The honest answer is narrower than either side of the argument claims.

What genuinely changed

• Time to first meaningful paint, substantially. That is a real user metric and it feeds the page experience signals, so it matters indirectly.
• Crawl efficiency on a large site. Rendering costs the crawler more than fetching HTML, so on a site with thousands of pages, more of it gets crawled per visit. On a forty page site this is irrelevant.
• Link previews everywhere else. Social cards, chat unfurls and anything that reads a page without running scripts. Not ranking, and often the actual reason people notice a difference in traffic.

What did not change

• Whether pages got indexed at all. They were being indexed before. Modern crawlers do run scripts; the argument that client-rendered pages are invisible has been out of date for years.
• Rankings for pages that were already indexed and already fast enough.

Where it does still decide things

• Anything behind an interaction. Content that only appears after a click or a scroll may never be seen, regardless of rendering mode.
• Slow client rendering, where a crawler gives up. The failure is a timeout, not a policy.
• Search engines other than the largest one, several of which run scripts poorly or not at all.

The honest summary: server rendering is a performance and preview decision that has second-order ranking effects. If a site is not ranking, rendering mode is rarely the reason and the content usually is.