The 80,000 products Google never knew existed

The 80,000 products Google never knew existed

A marketplace had 80,000 products. Google indexed roughly 9,000. The catalog was 'all there' — you could browse to any product.

Browse being the operative word. Their category pages used infinite scroll: products loaded as you scrolled, fetched via client-side API calls, with no paginated URLs underneath.

Googlebot doesn't scroll. It rendered the first 24 products per category, found no <a href> links to deeper products or next pages, and dead-ended. The other 71,000 products had no crawlable path — they existed only as responses to scroll-triggered fetches.

The trap of infinite scroll without a URL backbone: the products are reachable by humans and completely unreachable by a crawler that won't fire scroll events. Discoverability collapses to whatever loads on first paint.

Fix: kept infinite scroll for users, but added real paginated URLs (?page=2) with server-rendered product links as the crawlable layer beneath it. Component-style pagination, crawler-friendly bones.

Result: indexed products went from 9,000 to 68,000 over ten weeks, long-tail organic traffic +140%.