The product specs Google could never see

The product specs Google could never see

An electronics retailer couldn't rank for spec-based queries — '55 inch OLED 120hz' — despite having every spec on the page. Competitors with thinner pages outranked them.

The specs were there. We could read them. So we rendered the page in a headless browser sized to a phone viewport and scrolled... nothing. Sized to a tall desktop viewport... still nothing below the fold loaded.

Their spec table used IntersectionObserver to lazy-mount when scrolled into view. Googlebot renders with a viewport but does not scroll. The observer never fired, the table never mounted, and the rendered DOM Google stored had a placeholder div where the specs belonged.

This is the quiet killer of scroll-triggered hydration: Googlebot's rendering viewport is tall, but it doesn't simulate user scroll. Anything gated behind a scroll event is invisible.

Fix: kept the lazy-load for images, but server-rendered the spec table into the initial HTML. No observer dependency for indexable text.

Result: ranked top-5 for 280 spec-combination queries within 8 weeks, organic revenue on those pages +52%.