The page that returned 200 but had nothing inside

The page that returned 200 but had nothing inside

A B2B analytics tool lost 41% of its long-tail traffic in six weeks. No de-index notice, no manual action. Pages still loaded fine in a browser.

The symptom was odd: Google Search Console flagged hundreds of URLs as Soft 404 — but only the data-heavy pages, the ones with comparison tables and pricing grids.

The investigation went to the rendered HTML. In a browser those tables filled in beautifully. In the rendered-page tester they were empty. The client-side fetch for table data fired on an onMount that depended on a cookie that Googlebot never sets. No cookie, no fetch, no content. Just an empty shell with a 200 status.

Google saw a 'valid' page with nothing on it and quietly decided it wasn't worth keeping.

The fix: move the first data fetch server-side so the table ships in the initial HTML, with the client merely re-hydrating it. No cookie dependency before paint.

Result: -41% organic crawled back to +6% in 11 weeks, and Soft 404 reports dropped from 380 URLs to 9. The page looked the same to humans the whole time.