By 
There are traffic drops you can fix with one plugin setting, and there are drops where the graph simply goes flat. This is a case study of the second kind: gfcnieuws.com, a Dutch-language news publisher on WordPress, watched its Google Discover traffic fall from more than a thousand clicks a day to absolute zero when the August 2023 core update landed. The stream came back six days after mobile Core Web Vitals passed validation. Here is the drop, the diagnosis, the work, and the dated receipt in between.
The drop: from a four-figure daily stream to zero
The site had never drawn fewer than about a thousand Discover clicks a day. That is a stable, four-figure daily stream, the kind of traffic a publisher plans its schedule around. When the August 2023 core update rolled through, the graph fell to absolute zero and stayed there for two months.
What made the drop unusual was what it wasn’t. Everything else in Search Console looked clean. The publisher had already spent six months optimizing for Discover. The only thing still failing was Core Web Vitals.
That shaped the whole engagement: when everything else in Search Console is clean, the fixable part of the page experience is on your side of the table.
The diagnosis and the fix (October 19 to November 2023)
A Core Web Vitals audit on a Discover-dependent news site turns up a familiar set of suspects. Here is what the work actually consisted of:
- The LCP image was lazy-loading itself out of contention. The hero image was pulled out of lazy-loading and preloaded with high fetch priority, so the browser fetched it first instead of last.
- An icon font carrying weight nobody used. The 78 KB icon webfont went down to 1 KB of glyphs actually used.
- Ad slots with no reserved space. CSS min-height went onto every ad slot so banners stopped shoving the layout as they loaded, a classic CLS source on an ad-funded publisher.
- Background noise on the main thread and the server. The Taboola widget was deferred, an ad-inserter JavaScript error was cleared, and WP-cron moved to a system cron so scheduled work stopped firing on visitor pageloads.
None of this is exotic. It is measurement, prioritization, and the discipline to fix the boring things in the right order.
The receipt
Traffic recoveries attract loose talk, so here are the dates:
- November 1: mobile Core Web Vitals passed.
- November 7: the first returning Discover clicks showed up, six days after the mobile pass.
- November 13: desktop validation followed.
An honest note on cause and effect
Nobody outside Google can force Discover back: not an agency, not a plugin, not a consultant with a theory. What can be said honestly about this engagement, and about others like it in this portfolio, is the order of events: passing Core Web Vitals validation preceded the traffic return, by six days here. That is a correlation recorded across engagements, not a promise from Google, which is why every number in this post is one the client could re-run in Search Console.
What held afterward
The contract didn’t end at validation. Through 2024 the relationship continued, holding CLS green against the things that normally break it on an ad-funded site: a floating video player, a load-more button, and new ad units. DebugBear RUM went in as well, so shifts surface in real-user data before Google’s 28-day averages even register them.
The engagement ran as three contracts from October 2023 to 2025, which is its own verdict: the work held, and the publisher kept it held.
“Solved all the CWV issues I had and passed CWV.”
The publisher of gfcnieuws.com, public Upwork feedback, score 4.85.
Need this kind of work on your site?
If your Discover traffic vanished after a core update and Search Console shows nothing but failing Core Web Vitals, that is the exact shape of the engagement above: find the fixable half, make it green, and document the dates. Get in touch and we’ll look at your graph together. The Google Discover recovery portfolio page carries five more recovery files like this one, each with the same arc: the drop, the fix, the recovery.