The crawl that showed 0 errors and 0 truth
An agency ran Screaming Frog in default text-only mode across a client's new Vue site. Clean: 200s everywhere, titles present, no broken links. Yet organic was flat for two months on a redesign that should've lifted it.
The investigation flipped one setting: JavaScript rendering, on. Same crawl, different planet. Half the internal links were rendered client-side and invisible to the text crawler — and to any bot that didn't execute JS. The text mode saw the raw HTML; users and Google saw the rendered DOM. The two diverged completely.
Turns out default-mode crawls of a JS site measure the wrong document. They validate the shell, not the page. The 'clean' report described HTML no human ever experienced.
They re-crawled with rendering, found 1,400 internal links absent from raw HTML, and moved the primary nav to server-rendered anchors.
Crawl depth to money pages dropped from 5 to 2. Indexed pages +22% in seven weeks, organic +15%. A crawler is only as honest as its rendering mode. On a JS site, text-only mode is a comforting lie.
Hydrate Diaries
@HydrateDiaries
The crawl that showed 0 errors and 0 truth
Этот пост опубликован в Telegram-канале Hydrate Diaries. Подписаться можно по ссылке: @HydrateDiaries.